From f82ef212ceaf77a9dc3564ae15b664e606f0bf9e Mon Sep 17 00:00:00 2001 From: Immanuel Bayer Date: Sun, 10 Apr 2022 12:40:24 +0200 Subject: [PATCH] 0.8.0 --- .gitignore | 2 + .gitlab-ci.yml | 17 + .pylintrc | 3 +- CHANGELOG.md | 6 +- CONTRIBUTING.md | 10 +- Dockerfile | 2 +- Makefile | 14 +- README.md | 46 +- _templates/layout.html | 11 + conf.py | 64 +++ dev_notes.md | 5 +- docs/.gitignore | 1 - docs/Gemfile | 4 - docs/_config.yml | 66 --- docs/_includes/sidebar.html | 58 -- docs/_layouts/default.html | 110 ---- docs/_templates/copyright.html | 7 + docs/assets/images/company_logo.png | Bin 3170 -> 0 bytes docs/assets/images/company_logo.svg | 31 - docs/assets/images/favicon.ico | Bin 106291 -> 0 bytes docs/favicon.ico | Bin 106291 -> 0 bytes docs/feed.xml | 32 -- docs/internal-components.rst | 33 ++ docs/sidebar.json | 27 - docs/sitemap.xml | 25 - docs/tutorials.rst | 18 + index.rst | 15 + ipyannotator/__init__.py | 2 +- ipyannotator/_nbdev.py | 40 +- ipyannotator/annotator.py | 78 ++- ipyannotator/base.py | 136 +++-- ipyannotator/bbox_annotator.py | 67 ++- ipyannotator/bbox_canvas.py | 308 +++++++--- ipyannotator/bbox_video_annotator.py | 34 +- ipyannotator/capture_annotator.py | 202 +++---- ipyannotator/custom_input/buttons.py | 167 +++++- ipyannotator/custom_input/coordinates.py | 7 +- ipyannotator/custom_widgets/__init__.py | 0 ipyannotator/custom_widgets/grid_menu.py | 140 +++++ ipyannotator/datasets/factory_legacy.py | 2 +- ipyannotator/doc_utils.py | 78 +++ ipyannotator/explore_annotator.py | 40 +- ipyannotator/helpers.py | 12 +- ipyannotator/im2im_annotator.py | 342 ++++++----- ipyannotator/image_button.py | 130 ----- ipyannotator/ipytyping/__init__.py | 0 ipyannotator/ipytyping/annotations.py | 136 +++++ ipyannotator/mltypes.py | 23 +- ipyannotator/right_menu_widget.py | 49 +- ipyannotator/services/bbox_trajectory.py | 19 +- ipyannotator/storage.py | 66 ++- make.bat | 35 ++ nbs/00_base.ipynb | 283 ++++++++-- nbs/00a_annotator.ipynb | 138 +++-- nbs/00b_mltypes.ipynb | 105 +++- nbs/00c_annotation_types.ipynb | 371 ++++++++++++ nbs/00d_doc_utils.ipynb | 217 +++++++ nbs/01_bbox_canvas.ipynb | 496 ++++++++++++---- nbs/01_helpers.ipynb | 21 +- nbs/01a_datasets.ipynb | 2 +- nbs/01a_datasets_download.ipynb | 2 +- nbs/01a_datasets_factory.ipynb | 2 +- nbs/01b_dataset_video.ipynb | 6 +- nbs/01b_tutorial_image_classification.ipynb | 208 ++++--- nbs/01c_tutorial_bbox.ipynb | 122 ++-- nbs/01d_tutorial_video_annotator.ipynb | 148 +++-- nbs/02_navi_widget.ipynb | 2 +- nbs/02a_right_menu_widget.ipynb | 113 +++- nbs/02b_grid_menu.ipynb | 439 +++++++++++++++ nbs/03_storage.ipynb | 72 +-- nbs/04_bbox_annotator.ipynb | 160 ++++-- nbs/05_image_button.ipynb | 119 ++-- nbs/06_capture_annotator.ipynb | 363 +++++------- nbs/07_im2im_annotator.ipynb | 497 ++++++++++------ nbs/08_tutorial_road_damage.ipynb | 130 ++--- ...a_example.ipynb => 09_voila_example.ipynb} | 30 +- nbs/11_build_annotator_tutorial.ipynb | 69 +-- nbs/13_datasets_legacy.ipynb | 2 +- nbs/14_datasets_factory_legacy.ipynb | 11 +- nbs/15_coordinates_input.ipynb | 37 +- nbs/16_custom_buttons.ipynb | 19 +- nbs/17_annotator_explorer.ipynb | 64 ++- nbs/18_bbox_trajectory.ipynb | 21 +- nbs/19_bbox_video_annotator.ipynb | 44 +- nbs/20_image_classification_user_story.ipynb | 166 ++++-- poetry.lock | 531 +++++++++++++++++- pyproject.toml | 5 +- settings.ini | 2 +- voila.Dockerfile | 2 +- 89 files changed, 5466 insertions(+), 2273 deletions(-) create mode 100644 _templates/layout.html create mode 100644 conf.py delete mode 100644 docs/.gitignore delete mode 100644 docs/Gemfile delete mode 100644 docs/_config.yml delete mode 100644 docs/_includes/sidebar.html delete mode 100644 docs/_layouts/default.html create mode 100644 docs/_templates/copyright.html delete mode 100644 docs/assets/images/company_logo.png delete mode 100644 docs/assets/images/company_logo.svg delete mode 100644 docs/assets/images/favicon.ico delete mode 100644 docs/favicon.ico delete mode 100644 docs/feed.xml create mode 100644 docs/internal-components.rst delete mode 100644 docs/sidebar.json delete mode 100644 docs/sitemap.xml create mode 100644 docs/tutorials.rst create mode 100644 index.rst create mode 100644 ipyannotator/custom_widgets/__init__.py create mode 100644 ipyannotator/custom_widgets/grid_menu.py create mode 100644 ipyannotator/doc_utils.py delete mode 100644 ipyannotator/image_button.py create mode 100644 ipyannotator/ipytyping/__init__.py create mode 100644 ipyannotator/ipytyping/annotations.py create mode 100644 make.bat create mode 100644 nbs/00c_annotation_types.ipynb create mode 100644 nbs/00d_doc_utils.ipynb create mode 100644 nbs/02b_grid_menu.ipynb rename nbs/{09_viola_example.ipynb => 09_voila_example.ipynb} (63%) diff --git a/.gitignore b/.gitignore index 9e68efc..44ca8f6 100644 --- a/.gitignore +++ b/.gitignore @@ -1,3 +1,4 @@ +_build *.bak .gitattributes .last_checked @@ -137,3 +138,4 @@ checklink/cookies.txt #ipyannotator **/**/autogenerated*/ nbs/data +nbs/user_project diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml index d6d18b4..9f99e7d 100644 --- a/.gitlab-ci.yml +++ b/.gitlab-ci.yml @@ -98,3 +98,20 @@ build_whls_internal: poetry build -f wheel && poetry publish --repository PYPIPALAIMON -u gitlab-ci-token -p ${CI_JOB_TOKEN} " + +pages: + tags: + - docker + stage: release +# only: +# - master + script: + - > + docker run -i --rm + -v $(pwd)/public:/app/_build/html + $IMG_TAG + /bin/bash -c "poetry run make docs" + - rm -rf $(pwd)/public/.doctrees + artifacts: + paths: + - public diff --git a/.pylintrc b/.pylintrc index 5b34fd1..88ed6d3 100644 --- a/.pylintrc +++ b/.pylintrc @@ -93,7 +93,8 @@ disable=raw-checker-failed, no-name-in-module, line-too-long, missing-class-docstring, - wrong-import-position + wrong-import-position, + consider-using-f-string # Enable the message, report, category or checker with the given id(s). You can # either give multiple identifier separated by comma (,) or put this option diff --git a/CHANGELOG.md b/CHANGELOG.md index 2295748..61bb046 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -11,7 +11,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 - Allow user to use labels without using directories. - Refactor Im2Im and Capture annotators to render any widget on grid menu. -## [0.7.0] - 2022-02-21 +## [0.7.0] - 2022-02-19 ### Changed - Updated dependencies to fix Voila's conflict by [Ítalo Epifânio](https://github.com/itepifanio). @@ -20,8 +20,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 ### Fixed - BBoxAnnotator coordinate's input now changes according to the image size [Ítalo Epifânio](https://github.com/itepifanio). -- Right menu been rendered faster, improving VideoAnnotator navigation speed by [Ítalo Epifânio](https://github.com/itepifanio). -- Trajectory been drawn after delete object at VideoAnnotator by [Ítalo Epifânio](https://github.com/itepifanio). +- Faster right menu rendering, improving overall VideoAnnotator navigation speed by [Ítalo Epifânio](https://github.com/itepifanio). +- Don't draw trajectory for deleted objects in VideoAnnotator by [Ítalo Epifânio](https://github.com/itepifanio). ## [0.6.0] - 2022-01-31 diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 67657f3..b9e505d 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -33,14 +33,8 @@ nbdev_install_git_hooks ### Building docs -* With `nbdev` we can build docs from notebooks. For this, just run `nbdev_build_docs` and `nbdev` will build the documentation inside the `docs/` and update the `README.md`. - -* To complete reinstall `jekyll` config, run `nbdev_build_lib` and then `nbdev_build_docs`. - -* To run docs locally with `jekyll` you can run the command `make docs_serve` from the root of your repo to serve the documentation locally after calling `nbdev_build_docs` to generate the docs. - -*Note :* GitHub provides great documentation on the matter, please read [their documentation](https://docs.github.com/en/pages/setting-up-a-github-pages-site-with-jekyll/about-github-pages-and-jekyll) for more details on GitHub pages with `jekyll`. +* Ipyannotator uses sphinx and nbdev to build its documentation. To run locally you can run the command `make docs` from the from the root of your repository, this command will create the html files at the `_build` folder. ### Writing docs -As stated, `ipyannotator`uses `nbdev` and therefore, the notebooks pages will be converted into docs. For this reason, images should be inside the `docs/images` folders, so it can be assible to the documentation. After that, just load the image name when needed. +As stated, `ipyannotator` uses `nbdev` and therefore, the notebooks pages will be converted into docs. For this reason, images should be inside the `docs/images` folders, so it can be assible to the documentation. After that, just load the image name when needed. diff --git a/Dockerfile b/Dockerfile index 61aa6c8..dd11e79 100644 --- a/Dockerfile +++ b/Dockerfile @@ -47,7 +47,7 @@ RUN apt-get update && apt-get install -y --no-install-recommends \ ENV PYENV_ROOT=$HOME/.pyenv ENV PATH=$PYENV_ROOT/shims:$PYENV_ROOT/bin:$PATH -RUN git clone git://github.com/yyuu/pyenv.git .pyenv +RUN git clone https://github.com/pyenv/pyenv.git .pyenv RUN pyenv install 3.8.5 -f && pyenv global 3.8.5 diff --git a/Makefile b/Makefile index 4fb6349..5eed1ae 100644 --- a/Makefile +++ b/Makefile @@ -11,12 +11,14 @@ ipyannotator: $(SRC) sync: nbdev_update_lib -docs_serve: docs - cd docs && bundle exec jekyll serve +meta: + python ipyannotator/doc_utils.py -docs: $(SRC) - nbdev_build_docs - touch docs +docs: meta + sphinx-build . ./_build/html -a + +quick-docs: + sphinx-build . ./_build/html -a test: nbdev_test_nbs @@ -32,4 +34,4 @@ dist: clean python setup.py sdist bdist_wheel clean: - rm -rf dist \ No newline at end of file + rm -rf dist diff --git a/README.md b/README.md index 611eca6..b091449 100644 --- a/README.md +++ b/README.md @@ -1,32 +1,21 @@ -# ipyannotator - the infinitely hackable annotation framework +# Ipyannotator - the infinitely hackable annotation framework ![CI-Badge](https://github.com/palaimon/ipyannotator/workflows/CI/badge.svg) +Ipyannotator is a flexible annotation system. The library contains some pre-defined annotators that can be used out of the box, but it also can be extend and customized according to the users needs. -![jupytercon 2020](https://jupytercon.com/_nuxt/img/5035c8d.svg) - - -This is an pre-release version accompanying our [jupytercon 2020 talk](https://cfp.jupytercon.com/2020/schedule/presentation/237/ipyannotator-the-infinitely-hackable-annotation-framework/). -We hope this repository helps you to explore how annotation UI's can be quickly build -using only python code and leveraging many awesome libraries ([ipywidgets](https://github.com/jupyter-widgets/ipywidgets), [voila](https://github.com/voila-dashboards/voila), [ipycanvas](https://github.com/martinRenou/ipycanvas), etc.) from the [jupyter Eco-system](https://jupyter.org/). - - -At https://palaimon.io we have used the concepts underlying ipyannotator internally for various projects and -this is our attempt to contribute back to the OSS community some of the benefits we have had using OOS software. +We hope this repository helps you to explore how annotation UI's can be quickly built using only python code and leveraging many awesome libraries ([ipywidgets](https://github.com/jupyter-widgets/ipywidgets), [voila](https://github.com/voila-dashboards/voila), [ipycanvas](https://github.com/martinRenou/ipycanvas), etc.) from the [jupyter Eco-system](https://jupyter.org/). +At https://palaimon.io we have used the concepts underlying Ipyannotator internally for various projects and this is our attempt to contribute back to the OSS community some of the benefits we have had using OOS software. ## Please star, fork and open issues! - Please let us know if you find this repository useful. Your feedback will help us to turn this proof of concept into a comprehensive library. - ## Install - `pip install ipyannotator` - **dependencies (should be handled by pip)** ``` @@ -37,7 +26,6 @@ ipyevents = "^0.8.0" ipywidgets = "^7.5.1" ``` - ## Run ipyannotator as stand-alone web app using voila Using `poetry`: @@ -50,9 +38,8 @@ poetry run pip install voila ``` and run simple ipyannotator standalone example: ```shell -poetry run voila nbs/09_viola_example.ipynb --enable_nbextensions=True +poetry run voila nbs/09_voila_example.ipynb --enable_nbextensions=True ``` - Same with `pip`: @@ -62,11 +49,10 @@ Same with `pip`: pip install . pip install voila - voila nbs/09_viola_example.ipynb --enable_nbextensions=True + voila nbs/09_voila_example.ipynb --enable_nbextensions=True ``` - -# Documentation +## Documentation This library has been written in the [literate programming style](https://en.wikipedia.org/wiki/Literate_programming) popularized for jupyter notebooks by [nbdev](https://www.fast.ai/2019/12/02/nbdev/). Please explore the jupyter notebooks in `nbs/` to learn more about @@ -77,8 +63,8 @@ Also check out the following notebook for a more high level overview. - Tutorial demonstrating how ipyannotator can be seamlessly integrated in your     data science workflow. `nbs/08_tutorial_road_damage.ipynb` -- Slides + recoding of jupytercon 2020 talk explaining the high level concepts / vision - of ipyannotator. TODO add public link +- [Recoding of jupytercon 2020](https://www.youtube.com/watch?v=jFAp1s1O8Hg) talk explaining the high level concepts / vision + of ipyannotator. ## Jupyter lab trouble shooting @@ -106,16 +92,22 @@ For clean (re)install make sure to have all the lab extencions active: `jupyter labextension install @jupyter-voila/jupyterlab-preview` - -# How to contribute - +## How to contribute Check out `CONTRIBUTING.md` and since ipyannotator is build using nbdev reading the [nbdev tutorial](https://nbdev.fast.ai/tutorial.html) and related docs will be very helpful. +## Additional resources -## Copyright +![jupytercon 2020](https://jupytercon.com/_nuxt/img/5035c8d.svg) + +- [jupytercon 2020 talk](https://cfp.jupytercon.com/2020/schedule/presentation/237/ipyannotator-the-infinitely-hackable-annotation-framework/). + +##Acknowledgements +The authors acknowledge the financial support by the Federal Ministry for Digital and Transport of Germany under the program mFUND (project number 19F2160A). + +## Copyright Copyright 2020 onwards, Palaimon GmbH. Licensed under the Apache License, Version 2.0 (the "License"); you may not use this project's files except in compliance with the License. A copy of the License is provided in the LICENSE file in this repository. diff --git a/_templates/layout.html b/_templates/layout.html new file mode 100644 index 0000000..9c15de3 --- /dev/null +++ b/_templates/layout.html @@ -0,0 +1,11 @@ +{% extends "!layout.html" %} {% block rootrellink %} + +{{ super() }} {% endblock %} diff --git a/conf.py b/conf.py new file mode 100644 index 0000000..1a87751 --- /dev/null +++ b/conf.py @@ -0,0 +1,64 @@ +# Configuration file for the Sphinx documentation builder. +# +# This file only contains a selection of the most common options. For a full +# list see the documentation: +# https://www.sphinx-doc.org/en/master/usage/configuration.html + +# -- Path setup -------------------------------------------------------------- + +# If extensions (or modules to document with autodoc) are in another directory, +# add these directories to sys.path here. If the directory is relative to the +# documentation root, use os.path.abspath to make it absolute, like shown here. +# +# import os +# import sys +# sys.path.insert(0, os.path.abspath('.')) +from datetime import date + +# -- Project information ----------------------------------------------------- + +project = 'Ipyannotator' +copyright = f'{date.today().year}, Palaimon GmbH' +author = 'Palaimon GmbH' + +# -- General configuration --------------------------------------------------- + +# Add any Sphinx extension module names here, as strings. They can be +# extensions coming with Sphinx (named 'sphinx.ext.*') or your custom +# ones. +extensions = [ + 'myst_nb' +] + +# Add any paths that contain templates here, relative to this directory. +templates_path = ['docs/_templates'] + +# List of patterns, relative to source directory, that match files and +# directories to ignore when looking for source files. +# This pattern also affects html_static_path and html_extra_path. +exclude_patterns = ['_build', 'Thumbs.db', '.DS_Store', '.pytest_cache', '.venv/*'] + +# -- Options for HTML output ------------------------------------------------- + +# The theme to use for HTML and HTML Help pages. See the documentation for +# a list of builtin themes. +# +html_theme = 'pydata_sphinx_theme' + +# Add any paths that contain custom static files (such as style sheets) here, +# relative to this directory. They are copied after the builtin static files, +# so a file named "default.css" will overwrite the builtin "default.css". +html_static_path = ['_static'] + +html_theme_options = { + "icon_links": [ + { + "name": "GitHub", + "url": "https://github.com/palaimon/ipyannotator", + "icon": "fab fa-github-square", + } + ], + "show_nav_level": 2 +} + +execution_timeout = 1200 diff --git a/dev_notes.md b/dev_notes.md index 9b96964..7fa911b 100644 --- a/dev_notes.md +++ b/dev_notes.md @@ -1,5 +1,4 @@ -## env setup - +# Env setup To install lib and deps use: @@ -39,7 +38,7 @@ To do so, add the option: to the notebook metadata (`Edit > Edit Notebook Metadata`) -### How To build lib: +## How To build lib: ``` nbdev_build_lib diff --git a/docs/.gitignore b/docs/.gitignore deleted file mode 100644 index 57510a2..0000000 --- a/docs/.gitignore +++ /dev/null @@ -1 +0,0 @@ -_site/ diff --git a/docs/Gemfile b/docs/Gemfile deleted file mode 100644 index e9f579a..0000000 --- a/docs/Gemfile +++ /dev/null @@ -1,4 +0,0 @@ -source "https://rubygems.org" - -gem "jekyll", ">= 3.7" -gem "jekyll-remote-theme" diff --git a/docs/_config.yml b/docs/_config.yml deleted file mode 100644 index abaeda5..0000000 --- a/docs/_config.yml +++ /dev/null @@ -1,66 +0,0 @@ -repository: devops/ipyannotator -output: web -topnav_title: ipyannotator -site_title: ipyannotator -company_name: Palaimon GmbH -description: the infinitely hackable annotation framework -# Set to false to disable KaTeX math -use_math: true -# Add Google analytics id if you have one and want to use it here -google_analytics: -# See http://nbdev.fast.ai/search for help with adding Search -google_search: - -host: 127.0.0.1 -# the preview server used. Leave as is. -port: 4000 -# the port where the preview is rendered. - -exclude: - - .idea/ - - .gitignore - - vendor - -exclude: [vendor] - -highlighter: rouge -markdown: kramdown -kramdown: - input: GFM - auto_ids: true - hard_wrap: false - syntax_highlighter: rouge - -collections: - tooltips: - output: false - -defaults: - - - scope: - path: "" - type: "pages" - values: - layout: "page" - comments: true - search: true - sidebar: home_sidebar - topnav: topnav - - - scope: - path: "" - type: "tooltips" - values: - layout: "page" - comments: true - search: true - tooltip: true - -sidebars: -- home_sidebar - -plugins: - - jekyll-remote-theme - -remote_theme: fastai/nbdev-jekyll-theme -baseurl: /ipyannotator \ No newline at end of file diff --git a/docs/_includes/sidebar.html b/docs/_includes/sidebar.html deleted file mode 100644 index 5c811f9..0000000 --- a/docs/_includes/sidebar.html +++ /dev/null @@ -1,58 +0,0 @@ -{% assign sidebar = site.data.sidebars[page.sidebar].entries %} - - - - - diff --git a/docs/_layouts/default.html b/docs/_layouts/default.html deleted file mode 100644 index c940af5..0000000 --- a/docs/_layouts/default.html +++ /dev/null @@ -1,110 +0,0 @@ - - - - {% include head.html %} - - - - {% if page.datatable == true %} - - - - - - {% endif %} - - - -{% include topnav.html %} - -
-
- -
- {% assign content_col_size = "col-md-12" %} - {% unless page.hide_sidebar %} - -
- {% include sidebar.html %} -
- {% assign content_col_size = "col-md-9" %} - {% endunless %} - - -
- {{content}} -
- -
- -
- -
- - -{% if site.google_analytics %} -{% include google_analytics.html %} -{% endif %} - diff --git a/docs/_templates/copyright.html b/docs/_templates/copyright.html new file mode 100644 index 0000000..dcf7a2a --- /dev/null +++ b/docs/_templates/copyright.html @@ -0,0 +1,7 @@ +

+ {%- if hasdoc('copyright') %} + {% trans path=pathto('copyright'), copyright=copyright|e %}© Copyright {{ copyright }}.{% endtrans %}
+ {%- else %} + {% trans copyright=copyright|e %}© Copyright {{ copyright }}.{% endtrans %}
+ {%- endif %} +

\ No newline at end of file diff --git a/docs/assets/images/company_logo.png b/docs/assets/images/company_logo.png deleted file mode 100644 index 6cdbdf1abe07bd93bacd45f3c4c1451f6ab81939..0000000000000000000000000000000000000000 GIT binary patch literal 0 HcmV?d00001 literal 3170 zcma);c{Ei0AIGnU2$e?}%APf$NMsjAc4|<`lqifWF$^N1g|UReRAfmr6+(zoDcK8I zhDm6$XE!ruhK88mHJ;P6{LXWJzjN+=-|y#hzOT>w{k`Y@apRBMnQ!M4Xmcjrrc!~$bbyw^W-|a92B4<{^i+VJ2GGGFX#gVwU|=?;F;W342(op<4#d;h3Yw;} zac~X>pn*6z1jIAg$}Tr>%m%&zg1QN^bt5wfZYmq!$N=IXxT%{ZXtxQ1%3g1H{euNd zcJm+XpWv^+|ITiP`C|gxdLu1p%J$uG{42Xz{=M_Z@&6z8zo&sm2Cu)f8>YW@Zmxgn zruDCqz1R%DnfVVE>~Fs0zwi9F=0*ngk^cw&k7<7?LHtklU)C%Z3qel*4u15^V3?f) z!0{U=*A{M`t-O5v+qUlz_+4=4E+L5U?mePn;u4Zl(tBlO<>VC<_bKgHR#820P)+@i z#$ipZBicH;din;2Mn{cJOrd7S%q=XftZi)V?2n&-IXF6xj~+iMD0*64Qu?f{yyAIf)r*(auU@~YdHb&ReO-M+V^ed>hmWmo?Vmb2 zySfQIy~NLbU-}0IzmmQU4UdeDjZb`^oSL4Qots}EQ>cqHI%8@12Xkfh=h}K%WuOlD z5;)wg&7mtqKW-}k*eYrXHFiXNo6o#*)z(Y2B`>0+T2czNV_(#(tUQ4!f%4cJeNtvs_qPKN6Ka@orM# zL}usPbqaNzabS|BKN-1!w|icu`8;L)GuZaEfu56hGMVttvS zMe%dbxT#o7JNjC6*dlQx+^a+^lS4;vJ*10hI?eTVh^$p3ig?)EBvhr=1662W3jGqv z^DGmW`{?Ud0X2fjkLRI6k#RrIvvd)h&Q`@G(RVqMf_Io}mWRzPxbkQi+XmI%M8(J# z&A8pwmE_*_b9N*MePWEF)xw$RDDduh3rV`ER`P{0T2rNzTUbe4>*<&=B{s?cZ+aGW zH8GvzoAnj=bL9pd%JnFT_UrkrWqW)niKLRyh?C+oYuFN0=BVhp*$kkEbrY=_6?iU4 zdC7Mtyv~~(H!5*@jLf&lm2}vAKJnwSHAUu*hCp!!_zs>J2@ z_je-j_k04bIHCjxpJ_x>*M~}JF$~i6#?Rr-yuLg5w9_g(IN<$5LtOejoT)L_VH{Sa zuYf~E*25V5f-dfdQJ;~THzQO8#q%nCZc!~2dZtqk*%0W7g1v`&5o46!8APZ>{$@}j@Yq*e~w^f{q z?q%R;!#d9^b6&^UqX@cGbHnmmc7_Xeuk=1D$W%SfR%u3s6*omS&g${Mzk+xS+ilai zpt_XrDq>Q%PQ`eTL)r`{c1tTvr0(Jiv8nVe$u|-meK}<-p9ob^*8E<3*+G(^p=U#^ z$uEVyEA0w@o==vKG>|oUJu$U*165<#{9Q~KuH>mLRBpJGs-9;=5GmA4(sHG|Sx({) zGC$*Rp=0&3O?K38KL)PGGVD0?ylQosV1Pm?|@J^L{QMB9`Xa0S9)#djz%BIw77u0K_n)N6pKB>)mR)i zt3Dx(kx-V|$>P+m?staDt*F`1KF2UE&%lFq`DgvOcV^E#*y3CvwtRWNNrhMa9@X|6 zX?!+sO023qW@Za;YV9idgike^Uu^BFAs^Qdnl%l4ultNl^(Xf1a#ISUYv5ip;oVD~ zE=Nw~soV5^IMgN~a`Apu$LINL^bAn=*@~P+=3%OG3K z7wESH_rt8iNXR;Igrg4r)_I6NiwIwp?mzFxw2090QB`-#N8{^z+@0uyP$;tBKT|S) zOF?{ZuVS)?(&0jJf_Ppv=1ur}gM9>}4^=9MBq zEJx4|v#a`jyObl}v}sv}HK*ouZo-LLgx~dRu}s6t6Z5%L<20Pm&7x%om^c1T8635y zDC+=@-MxD@Xr&u&irn|ih<-TKCjlKCLu@79=>B=D&U~QVan|$9(MOo^A)(K-Gh20T z)4Z@I2}q;~Q?{qe_yxjZ?ugV&zgsUFpySaiTiuZ;Ii=p0Wxee!6Q8iW3oK4=t~uH_ z0@4C&wbxQqg=&?dvKFla$)4*m$`bS$d3de0r#?o{`B|IZ*``8SkI`=9bFuYM(pTjJ z*;w-m50jITXBV&Xiqw-6duv?L06Drm7o{y+Tr7_0THb-@SDk%EkEbdnRzIIq;7gX1C1``{9!fN zX?KA{GoF$`Lj-wmIVr$oC$yRJ6Zd*Qutui8Nns;q#YD8*%m(->A%fF|zVWd<8AdrP zt64m^<;D*xa%k%}7qxo90m__&$K7LQp_Or`J^VRrh?>I$eW33Y$!sTSQnL`%g*$bc zlO$qGC6;}l@;gMiD%q+<20N3v_K~nt29o_97a3Y1-nB=+o1Vx(Tg=1bUM6q18W`ga zD4s{y>`7THOy+zI#X2z<5!(>BNBIHI69h^pi64yv->zu&$t`v=a@KFY5`AQ`7%;&l z(>Pji>FKw%K^Mo67#iDvYQS>NFv(~bjw830Q&J7}rN`Ro4y6mKfS diff --git a/docs/assets/images/company_logo.svg b/docs/assets/images/company_logo.svg deleted file mode 100644 index 29f9c4c..0000000 --- a/docs/assets/images/company_logo.svg +++ /dev/null @@ -1,31 +0,0 @@ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - diff --git a/docs/assets/images/favicon.ico b/docs/assets/images/favicon.ico deleted file mode 100644 index ba88f90b9983fb7d204ec2963d1c507d8b9d3a11..0000000000000000000000000000000000000000 GIT binary patch literal 0 HcmV?d00001 literal 106291 zcmeHQ2Rzl?|NmYit3s*lQ7WaOC`wjCQ#;Zysz+&9Z6jB^NJ=HiC`vsE(GcOLp+Pi+ zQbeK@i88|ZpU-oj`_JR2$2Gg2&gb>IuFLm+zvsL^XME0QQWTBqLIKezdFqY?MOonA z>2&62C4L&EEJ;zix_^HjPf_#w(kK-b=I3+$Dav65{-*rjpG7FD-+me;g5TkOt9U5t z;RzbGZ}!Z|-2^2C@vXZ}oicGQ{1;yQ@aQFW=kPNn+@?;{o4@77Ctpi9#RB6*v>>gZ zygVy1QuDRLrz@M&ugRDyWO#T?)Kl6i(p_{AT_fR?qr&*{+lBRZ3hRpSE1Yu7ovY{} zzfk0`#%_<>JenhBu3c*#QS{X$QZrJsV#el_@&$MHpv6xkBloO~U9j|3ZJqOzwHu8# z8yTg2`|j*tvDLp|tFJsi7r#QtZG&o_og()YXd(&A6shqG-FS5QT@O+G<|BGg3Pbd{ zFo&PH7N6=e^rnX$P5S`oYRZ2=%Iqn-n zJS0}_+g>QC#)I)YDW4I}H!1i@NCcm$7Wa_2F7qwu-qzHCQD)U+;u1~Ot-HHTwN9#4 z{y2F$>!Nh_`1=ext`AE zjvp0TAthb%eN^1L^uXFsWj~V%O54wFHQUJu^DeVIp6-@S`(Qd&!kF^CS1m17T{WC5 z-_oqEzgfPRLyi7@x4egxVs2O?DuwOnK4;daKIG06HTy`DPZO4|8xyy=uyBCT zc*T7~=ra{RPD(GJKbT}zB^JKnq+jhl`|7=xClsax-mnbjetTDm%bO;6yijQcP0=7V z$Vht?SFlcsQy;%SmJjkZ7mnBFduDX=(hdJ|iEL?;V_Z3BU2eX-Ze8Gcnfi7^_+EGJ zD&LbXBI_pi)_ua`V*4e{rK-d(yJ}^IXP2+VYd$ZwG<12OCg=6Y+jrO=p}>kSmYWpc zM_%qNpgZ^3F_$M>1ETq7IPL4r$ShHx%~RVmTt|3*f#raKpYF_hch+W}Uv90fYWbdW1C@kS1Df}#pji>YhkAxD zRU3R#S*N0ek?G7mZKccf>>1S|V_mb9JY;wW$nv?@Nlx&XTs*Tn#9q>vvh6vb;K-v( z*G#o0%!@afR-H6fIR57Q+h50+m2Ui2=M;1Nq~c`rA@{iU`aJ0Wt!6k9wJE*8s2-Ad zBez;(SWb|{%$m=^8e$rsdrtLVFsrYS{IG=Cnfl(biz1^3jWSy68gO5OUp-Aw?MuY$ z7q%PlZCvl9XJnx7JsDr723*aU_Nte<%PoEMIk>XCFid%s(s(hCp*!#3>w3h~LT=no zOmXC2Wq3TgD1A}dmfJ5Gv?sjUH!3{4%8soR?IOHy@px_eIRVM6=f{$snF{YL8!$8V z#3;-8;z~gdk^(*GN}r{4op{d;qottH6$=VzI<#Pa&)mz#rId%niqR%kVpD{vDwz%E z&bT^SYaHu-%Sx)(*w>P4E-(6X`VDvceWf)AM(GddS{zaL;+~X$UX?``2`vprF{O_) z-z~AH1auWEKfd=_v!RZ5|7(r8?W6fFWrbA&&&wXVt0X-)h_n-y9^IFIno`MA36aPM z;VSZSndb6*Mx@}D{I|=?AI{4PO;+&?-y)w`_TEmBx;8%N>047Rp_7$TGmRG%B&}H| zUJ^4XAf6$&yO&e)XHoRhpN*W1LG6}4L1d2$T-9Zysh zt!F$l%Fun0$uE6yfHw8fv5!>6x)O`h1M^KU?A*K}U21mM?Tm$@)bn8@E?IuEa|m;) z>t#|>xg$h1CrFX9<`!UF3?H5Bk#?rU$M4eCqQo1bChpxdHM!ifZuvPSE_Ym4E}I%6 zXJ#!o`TkJ8yNM#yjy>0{5nrZ|@T*DAH@`^fR*kC_xRNh!y?d6z3Flp&%9{R^pu?)< zi9;Q=+;AI}e1iuq^2fS74+thE9+i0XJuKzUcc)W*;+^HBoXyoe zSKL_pWb^4s(fkZvefP(67iv8`9dKa3XrjDq{FFk+i%FDg7?-#CF{M5)$8&w&v@j-V znXi|r*=p-7n)a_7loEYdLvF ze5B>nPy^+!n}*(9ZTWJU{D>+kdl6$w{g!v|n6p}CYAUs+kL7gSA{^3JzOQ^)#S>c> zOE=9sB11WrPB!qEGp73d(bkzWNww2%5xzI$Pm9hr&EH_QQS^A7WW}dy zy$wiZu&g5WQXwfn%&YR&YQyD~J%W2Z_bIl{kFV%DQ`t&ENAIbQJKdL2fM%^<+UxjL z7alXsq;ZL*A3j*xUkO|ipt_XjuAD_)M#b{1^yymYins&)w0+J%mj}FN)kPQYj6Z$a zgllFkw_M4#_^$LKd0IyIHz%Ffomp^l3mFSTcau#rwxa-r4QJ$KT@U&{XR)<4htt=|Z<^m*YCdnC05<;~_C%=dKb z?KX|tzT2|Wtng}{-RcE8Rm!QaqD0LP&syj0JjcB@vLB<6yQb7ZOZ#2Cth<1nWVEY* z79Nae{LdL{C6`EOtSE_%>)wMdHD~0dqBW{R4)*H5>(z{p_YZsxNObnudB`Q^^`tcc zb58M3OwO84NpG9APt7fnn`6IW+t*KP)AA!FhNnynPLeOG8MWqd zZ<(b&yg4>|6+(5nZ2BeW9=^M1L*k75OrdVCX&<-hgh+q4KBFRF)CvYKKnFSnu{aCBgUB}|TYgJ>`FH%~Ui521GO zbK|fO!pO6URdd^s8%=SiP4J@Q>f4cb=MX8Lnx~gWv@&^BOh$t z+}CP%k)*!FR)NdM#=1JtyoTY?(ks6Ekw^O01HKtVY}j;MU|{Gh3rc(}*OBjmT2@tk z*LM>W4#y){*~MCVmZ*@fg@^ml#kaOucNw1O(CcfDw_ke=5SX;ic*6+XyvlC7smbeT zE4ow$&YViU?lF9_j_$LfVe`F`RwrHw3JYKEL9Ok-YNV4|Zy|;T*PR1XC1`u??_c@Q zW2M=N9&`}_#sZ$wXEoFF7b{onm&!Yv+@qj&l>x85{5?IH&rZ~c86}$r;ivG#85d|T z6Oz|VdauJ9aATB-T&mZoG_QVQ^t+PtJVpuq)mlBXZStZ*^t=;_Vv|g*g?skQ*DFwy7wkQqs?>2d+_88UKT-(>qfh3Al|L;OmfInkpmM{q+e8ReBQ-oNfECxmyN+P z{$3mH%@4#Y?k(~gBrQIBbWR$Nwx`Q#<;Su{y&1`QTW+kJ^W8wQ*SuUUcV3fdArSHd z6&QwG*&&OhzPdR4alT|z&qOVuV6IVJJ*Xu@XIurR<9APSo&e+g4AnHR@;e)arJgUm zjpVG3A>qOI2hdhYxulHSp|Rhse2B}SLWzwtcJpZrDhj4(0=nNnUl<@d$$QnyzFqIF z4(&IR#;@G%`k+Yt03Kbg)L}JUCh3}=d&!vifmiRcx5Ft)!*B3N2}PGrr{`Gsx+wLL zc)#Gp#(=9UdJA^DwW%M&VEreWcb!z>%*(DX1fRME4zqjgTDjmXmrUn1w#PhAGq){-Ugk!C!lXe;F`4 zPUMsA`sG8^PTPF!h9$y;u&09j>F0Z<4=k_yE}Hlu#?nywxriBeH`zP zeXq^OZaIF`Ta|mNsd=KjP{2i=^sPfe`|h==LxM*RT*pHEx?45Jjb2k{Zuvq;91qQ|o754mmVE;|?ug`tlk1O#+r&JUxMX(}7~ z%s*i-R8+?JA`mI7yL0DOl&K)E{$Rct4L@Yti6S@K;il9((|2nbsY|gG8i7wtccZ=g zkCAoKdw%m)mi<;w0j0;?4pEnV-i<|rRT;Uq@|t~V=Pu6Wc~CHB;{FpZ+?2FxoslfV zLXny}XDMZ2$|LB>D6X*a-`M|8H!tTZo^U%sCEs{E=lSI&yh1J+o1C*HvHX7I5zV@; zpNgZH#+M6}wEXf+VYQI2>b0YC5}d_pd(DfT%V{TSO@_T!F4j<__PvRGVG!H($Ro4tx9%^iMN`@%Q7 zH^yhDnpO`8|Lpu&_NDEjr%syA^aG~7yu-$Gd8oO}3N{nD%yaW$KyWbMEqVXDl6eZt zrXG5|;>(()cZ>Scrkwb|Ws_?yapr2?oZ^_;mGt=v?^Rt7PES!4|8O)#MQyD6`19Vh zoJalzc^h;TsdW7veerJITQ@j0gf`sq@Ilo{@|uEfS_};r6Gj;K^nsO?Me%EwDVz;F zShabHe)S)dUHZlMD4tRMK@*>b5=CtEsQ~fHE3X#^s;Dflt^Tvr;dYe%_Po_Etfi(< zWA~*>eDj^9x+=dg(ZxlR{-&R<`Q|4pE54lzvpZ?Jlq;^*A3 z6e-^+8LEEYjm+LjF>;F-_ePlIY%(j?p1Ee@TK&b0t4p;L%Z8a!kH)PtoNTftcHru) zt_F2@)VTNHIP*%$cUL}otQ;kH?S4(XfFkv3r)i*3tlgo~G9Twc?mZeq?j2ZvPsHer ztId{(P5WLx5fgc4S!~qj_>0B9)&`b$4J@e7)%3Wj`_02{8(hXyOMTAp(y(#7_>ge6 zxUkmoQ--xomS)w25c{J1Wj116EQaX>Y!#-g9oaDJYU^b6@LDT2>nSTFuN*2!^-90*8q$@oPYn*(hP%n1n9pC+qTW?jj+u|RjLPL7$ zXYs~~%Jmy>v*^@y7jMI^l0#htPjt=Ato-=ysQp;^iJ7(sc>7tSi}JKpePXhOy2{X| ztu(H^8ZC0rYh1{msujzg7)cAC?|Y~Gu@28&+0z?6C8kw}ME9OF9c!o}@Z-GR+^EHa z>QH9x-J9<75q4hi&RY<7x&K zOF17`t|_;>e|dCYyui9YOuACTqf#m^G;9l{Q9keTi7$D79sX%kBoj}K!kWm}iR&U0M6$8OyIM-^ zlX%~^0}J04Wplgfj2Dtu4SM!k{1Bz+E>RHeZP@GOd^sWdjXEPSuK4G9TxOD6N~!QX z(E%E->LQ0a@7vo=KzHDiYn}_%PcF&ZV4mMs`J-C8cyP9IP_^6Ec%ggcs@lswtITtd z=_6QnKn`DP=VLs~@uS`%Di}#bmFt`wv2aWO_mQsd30Bp5#g$4bn{cagq9N5)xZapC37k7Y_$eCe4b zTdQ&Hs4L@v<*{zgdF#boybF?iH039h=tm6-*s}MaT+Px)`6YFS3|B`?;rUxmjoufx zRCn9Jz})V&YCCorrhLBA(>d9DyQ?Yh>I)&6>8s`~jY|%yUg#dOg+4Q>rn}jfLA%R} zx9lvjRsC3gG{s+|2R$?)bN<5HN2&ycc0^kW((?w4vo1bfw!!jSjHTdquUGSL&p!}1 z&x6{aY9*-j@l9Tg&dooyHrMs#-ZR*XuD@@AgYr`g1DEM1O@^2TEu3#rnQ5`he_iOc z4>Ycel|1T`tYh=+@|S50wXV%khzoWW_T5_LyJ4aIR{5m?RFHwF@TeO130$dZ8W$IB zaq={ZJ0g=W!gzQ|)Uc0%=wrb>{4T;&q!n$#T)n(9%n z{D?oi^|VtCS`W_bKUP9OH|(&?vV$iF+ay!*ckh;{59+GND4>d}xgERttVX+b%`o3d z&Abp{qjA(rMlB&VC4H__>BXt?eJ69JUf42*G2y~tcae16r}WOG2zGgNd*A!xZ~R4k zU50?px;Mj6y3dq8r*%TTxSvd7JXCbrHGZu;EvmPBpIalAu3sXt^z{1FFk59S$-->D z@{$AJqORKpM9JOj-v6%)h0K)Q`BqG(cpl5DzY>0{PoJ;-URi7zw=uQ2)Q(X(fsrv` z@9j1IGF|ZInV>j3gL_36&$#glqdj=TkIt9PC{+@vTs5k0ad?-Az;PDl>a(-F%F_gF zd9mcLA)EMKk8*pR=rlcyJNZMhOk(m}9sOdOU7)Vjkecm5VYByQxpT4YQ{7T|Iu%vc zC#vL(4s{|jYS47~c|1APW=ApcXwTks$@oGQMM}M|`h#`0I#L_fNvwKK#doRPNYRG! zCZ5XAiPHVk-(MoiH$|1ZT>18hD@wPA=JM!?T#AX|^L7$l6^bRQo><-ux0gJ(B#2)v zoi_a8r8qBcD+hI*_1Oby>J#_^d}dY)?u;6M4}m-K!qqp=dpG=+js{ipe7$;ddX>Vx zw|#uw|H9g~^Oy{NV#cU)K2-HOBWql&FE-CmQzy-w7&_i;yAuAL4u8P58!+xL2@E;^ zBt(cHKoB4Z5CjMU1Ob8oL4Y7Y5FiK;1PB5I0fGQQfFM8+AP5iy2m%BFf&f8)AV3fx z2oMAa0t5kq06~BtKoB4Z5CjMU1Ob8oL4Y983Ix7XBa~c(P(UU^`%@9xlY-EZX9!)* zMd))CLUnbmxHgYd~}H<@D=8&Jz!0kN+DJ5y~s$^d6D+ zH3tED5<)!!*yF$DErhbaH0SQ*T24p68S`J1@SoG!)Nbt~{BJjCJH?F&|2qYn+O3g< z|Lq2Cr?@fUf2UwmyET&Vzulni6gMXP?-Xomw?-2Fw;Qyb;>Lvkoq|p6)=0wtc7wK4 z+?epcQ?RMs8cF!yZqRm$8x#I_3O2P{BMJZ84cbm|W5WMV!KQX=B;kL%LE9;AO!(g^ z*wk)~B>ZnTXgkG?3I96|3_)|>`A#C#{~SUy@L1S_*M`t+W-MBp1M!>Q zur2ssg=2zVl_GTd1#=yT7{aFFiJChO;QFq9%s4gJBOopn;uZvl`{+T8EJIMBpviz|*xh4?iRm)i0%&{F=_;ypeO#{{3sL`eS%GhVp4d~EIY5T9(0 z<;I)um~vvvQTTno*%JQ4e0z=KfOaG^7!?Ixt1VT1<5%SGI=z1PAj|9l=kZ%ai4xkrwe1MR_HKq=zIrneu+P+xFl=zlm zhGThdI$jf6B5Obns~wM+XVl-D!~fpngIG_$Wb!*&*twv zN?30U=R#O>2sMsCYxxiI0OSuk_ynOWd^QBVP+Jr131_g3pcmOZJrm@xgR@F1*#~~g zlGgAac&`?UkT=c?Re*Wao&b3|Ec(A_{rCRT68^*b3;SxyFWOh(Y_Jo{_U!7RKyP7%_0{dYM|H;0 z(Gva-!*eh6HS_HLEBXu~Cg_f4MX|pIfQ>Tt?XPGpXFs#K{D-;Z{P5T3fdn1@pRd^0 z7lKaE6U!>?a6I8`1~<{V=JJ04K2t~IdBh1gf^J~Wb@tW(7`JP0BLw-b3H?Bzx%>w` zKG=XcfuLpq@bq0X{kGq7>l{Y{7 zPJU~i{YM#}qmpnAcut~sSnssrCVTvcb0Fj}C-eh>=FESH)p&vJd7MNgRX9E%jXnO0 zV}B(lEV~l=fk1Qk?|^OePzQ#S2=uz(L%`1Z4>2bYPeSMi0?pz73A}c4S_}@@n?N3A zM+SsABa^rveJ8&)Z~uq$Ht66vjX(zIhUa!uYXHcvu!oTvHa|Gnoca$SJ9Cok&*}xl z0fGNPQ~Zbj&yB`on1jwQtUmmzchIBv!oJh&UzOeoiRKg&wox_!+zCx^n8X;UmUN=4t8g;t_^f&SEcJJIV*MUtR@60w!wY?gnl3Zv9sW7&klZ&j|24WoJ3HQ z7V1^A(|(W<_hMgSLO&3AiFGFs*ToKgkh=xq|2V<+%j!Wc-uoBCu-AUbKLdWmgnl3Z z>*uC>?Cqa0MuJ{&jt&UM;ZtnSg**W4j9lz`VeLjObne0bT7|*6V;zH9I)MIXeQ&Tbyu)dW`437PjfL({8A73%X9oO-AUS z2!Iba$i?j72xn;U0ebn1?ZBYJ^UFYp-B=KfFoZ&0|FhrZ6B7Z5!G#!}o_LP1GX~_a z90<9-Aa3_p5X|9Q1=yb_oW1r!pG@Krf>>lijU#Xy<2mCRdw7B!74+!vonJw)_CO9% z$QjAbcm=N3!?8}hUz?t`Bg zp?^o<1D;!O76d&hJ97(o06vwwu?@MseZnE1Fz_CHczJx-YajH%APS*7SZ773X$0U* z2mbM3BV~7f!F7=D)&4#+CabMuEOC?m@;N<($0gV)K=xo~Y{E65?}c2%oaVFEG$(%D z6_9h|N-jeEFi+XV8Lk5z+>$t)m-h`q;E&Z71TqQ8`CxZ~_$^v9cn`T~AV(pm`75^t zF715}kg*`QXLrBmjRAqMfq_OyH25_2!7tM4**YDxU%~W`L_mRxd8NS;F}FLLx|aiIuNB* z%v@jKg9Ek)@F#(J4mn=H7Ih!Tmw-S%_^G#5 z4yu-+_uu#yti>>QA%{+D#z1}kwCC?I*0$p{hxm84?f3&5>rlL}Lq4tcy<4d;S= zsU+81+j5?HH;@M)pX1c4?CV@x!^`G;7i7IY`0N2WWQnZT9^7Z%6Xby$EDzXWoiDqw z7R}*fORg2b@*&iT0{b=b>25#Xv+fM?0OU}DoD1V4TP~lsgxB@I19Ki^eESDX-(SLe z7Kb|K9po!G-$U#uosb>OPKAb6GOcLsN1Yoa)^C8sXg52$JPHa#4P#)|5 zAm<$T9e_N_iE*eM&xK$AMDQJfaR9NQkf#}Z>_7+D96f#gwVm}ltnspVjh}{n5}@uM z_??1n0`h(l;yeUk9fo}Y^74T!4!S?c%L4Yd(P2#eJ@^iSeg^b+(no>-L4Y7Y5FiK;1PB5Ifwo5g z@?Ssrz^vZ|`89thg8C-2a2&djFEgLzXq>Bp)9To@CqE#@nw3|9^=^i9Y5XoR^^MiF zfO;^$OP;_fIBy$^@2q#oq5r$%hH{AZ1O);k&hWd4)&uRWy>&bc>P|H=B#TK8G+ z|F-^vTnu$1;`0#-&t}#Qyo~dkK`jcXH{8}WPTG5>urvN4*HTs~LhiV>7UbTS!TGql z;oMwMYX@pRLT+slf1JlF7@_GnCmYmWgc<>zftu<+oQENI9n=_MjgO|@A?GFJD}%a; zoSuvDXZ?q|IH@0*HNPO=UQ>N;?EjGS(J-d%YQD6`c=}oYqj28f{^wfO?}k`_P66tj za3VFU{=pgtIZPpU zA?O8Q4}f(Z)?A9|tnwJSrjnKDvC46N>OO79W+S^uHVQfmqFbZSTZ?-)?Q zfxf`n$2)WuhfsAhaZ~qPP-=H>O zTM(T88mm=zE~_o>-uiohOxlqDP`{(C&aF@zsNwxVUjRB2P9Ufq+pvAoxTaZK)hUH} z-thiV5A`9ow{rr49t-%kAk{8rt$xKUl?m)*__gQW;;O71#_|?22j5o zYRxv6p#F&`J~uYD?n908k65?HNd$Wx)b487fA|ct;QjaP*Xe~ha|HLjv2hQ0Kz%4q zBsjA-whjOWP=5mW2{meeR(Aj}OefFM8+AP5iy2m%Cw{}clMNCrSY{B{UHo(L8EGQ+R`&l91DKQy5K)SxGwAqWrz z2m%Cwc0&O4SD*(6eLwj0`(l6o&ZN@)jaaOhSG?+AoOh zN#Ix8S?I6!*Rz2^`@zQAS%|d%*K&&BPw;OI0_=_-@XdUJ{cs_64_>fFu@mRenzpg? z-G6I8*wpS}e+-j&gygYJ4(yW# zH~;J%Vhh0!7vgFUq&^ z=hNKqxeY)6toM2EVy0ZyG+SeLgqVewdGeFW7$`+ymmM>-S-1 zRLAh3vGzkuT62B+Bk`QC?;8t#>>Y8AXxM)6qw&vVAA8o7-?0CN682%N&x5EAh`(#t zemKj3uUK;lZ*_m|7xmi@eirRb4ciZXZ66z1Z`f%E#7@?4KYVU)`h@*UAx@X|+}5IJoL#@|;FEkCpK&^pz=smrT)!OZ4} z7f;#TzCwJ*`7DHbV%d|`9(Lb>2l_Z({a5?%!d!rH53$4z<@%rX1=bqat4i@a<`gRb zj?jfyOrKFj`K-8H zu*s~x#mp-J@tL5vgfld&J#fx`+5+8iV{NYge7HaO$7SL;*M{-`aB((PF%-Cj@PF)X= z>8A8k^>Zt+!!LGP2)-@QC(xULUh`)Jb$ulN);)`0U0_?;ZJ*(_1o8ty8vqN)bzR>s z0+_&Dt8cfVC;hCIE&KbaOs2mNi?@(Z$s`^j;4#FcgRP$RKhQ}x*Y3?utH2)y@?^1| z)9@X9d~C2!NmUIycWc=-;Hz~M>)#=V9ptM3n<3=a*?OO8Z^o=3L=fmS2sBU!bofWm zQ-E4R9>4xyPr-vfG$2j`6zq&IvE1EKf<0vhmStN?bMf9j727OZO7*`FYc=EtYie(; z|19|aBy6XH{No_+!rlmR8z2v}L-2i&5y58xY{W@L2tm#lut~9FGluWOUJLdu7D4;L zem);#+Z2Ip1iqdP$r1Nw2$r#%A~@@SuUW(O9{Qk*<&>rf_G{?>(P#J`f8PXr?O+ds v*cWyP_93_)^5j6?IJh6Qt2ulGAIpIuFLm+zvsL^XME0QQWTBqLIKezdFqY?MOonA z>2&62C4L&EEJ;zix_^HjPf_#w(kK-b=I3+$Dav65{-*rjpG7FD-+me;g5TkOt9U5t z;RzbGZ}!Z|-2^2C@vXZ}oicGQ{1;yQ@aQFW=kPNn+@?;{o4@77Ctpi9#RB6*v>>gZ zygVy1QuDRLrz@M&ugRDyWO#T?)Kl6i(p_{AT_fR?qr&*{+lBRZ3hRpSE1Yu7ovY{} zzfk0`#%_<>JenhBu3c*#QS{X$QZrJsV#el_@&$MHpv6xkBloO~U9j|3ZJqOzwHu8# z8yTg2`|j*tvDLp|tFJsi7r#QtZG&o_og()YXd(&A6shqG-FS5QT@O+G<|BGg3Pbd{ zFo&PH7N6=e^rnX$P5S`oYRZ2=%Iqn-n zJS0}_+g>QC#)I)YDW4I}H!1i@NCcm$7Wa_2F7qwu-qzHCQD)U+;u1~Ot-HHTwN9#4 z{y2F$>!Nh_`1=ext`AE zjvp0TAthb%eN^1L^uXFsWj~V%O54wFHQUJu^DeVIp6-@S`(Qd&!kF^CS1m17T{WC5 z-_oqEzgfPRLyi7@x4egxVs2O?DuwOnK4;daKIG06HTy`DPZO4|8xyy=uyBCT zc*T7~=ra{RPD(GJKbT}zB^JKnq+jhl`|7=xClsax-mnbjetTDm%bO;6yijQcP0=7V z$Vht?SFlcsQy;%SmJjkZ7mnBFduDX=(hdJ|iEL?;V_Z3BU2eX-Ze8Gcnfi7^_+EGJ zD&LbXBI_pi)_ua`V*4e{rK-d(yJ}^IXP2+VYd$ZwG<12OCg=6Y+jrO=p}>kSmYWpc zM_%qNpgZ^3F_$M>1ETq7IPL4r$ShHx%~RVmTt|3*f#raKpYF_hch+W}Uv90fYWbdW1C@kS1Df}#pji>YhkAxD zRU3R#S*N0ek?G7mZKccf>>1S|V_mb9JY;wW$nv?@Nlx&XTs*Tn#9q>vvh6vb;K-v( z*G#o0%!@afR-H6fIR57Q+h50+m2Ui2=M;1Nq~c`rA@{iU`aJ0Wt!6k9wJE*8s2-Ad zBez;(SWb|{%$m=^8e$rsdrtLVFsrYS{IG=Cnfl(biz1^3jWSy68gO5OUp-Aw?MuY$ z7q%PlZCvl9XJnx7JsDr723*aU_Nte<%PoEMIk>XCFid%s(s(hCp*!#3>w3h~LT=no zOmXC2Wq3TgD1A}dmfJ5Gv?sjUH!3{4%8soR?IOHy@px_eIRVM6=f{$snF{YL8!$8V z#3;-8;z~gdk^(*GN}r{4op{d;qottH6$=VzI<#Pa&)mz#rId%niqR%kVpD{vDwz%E z&bT^SYaHu-%Sx)(*w>P4E-(6X`VDvceWf)AM(GddS{zaL;+~X$UX?``2`vprF{O_) z-z~AH1auWEKfd=_v!RZ5|7(r8?W6fFWrbA&&&wXVt0X-)h_n-y9^IFIno`MA36aPM z;VSZSndb6*Mx@}D{I|=?AI{4PO;+&?-y)w`_TEmBx;8%N>047Rp_7$TGmRG%B&}H| zUJ^4XAf6$&yO&e)XHoRhpN*W1LG6}4L1d2$T-9Zysh zt!F$l%Fun0$uE6yfHw8fv5!>6x)O`h1M^KU?A*K}U21mM?Tm$@)bn8@E?IuEa|m;) z>t#|>xg$h1CrFX9<`!UF3?H5Bk#?rU$M4eCqQo1bChpxdHM!ifZuvPSE_Ym4E}I%6 zXJ#!o`TkJ8yNM#yjy>0{5nrZ|@T*DAH@`^fR*kC_xRNh!y?d6z3Flp&%9{R^pu?)< zi9;Q=+;AI}e1iuq^2fS74+thE9+i0XJuKzUcc)W*;+^HBoXyoe zSKL_pWb^4s(fkZvefP(67iv8`9dKa3XrjDq{FFk+i%FDg7?-#CF{M5)$8&w&v@j-V znXi|r*=p-7n)a_7loEYdLvF ze5B>nPy^+!n}*(9ZTWJU{D>+kdl6$w{g!v|n6p}CYAUs+kL7gSA{^3JzOQ^)#S>c> zOE=9sB11WrPB!qEGp73d(bkzWNww2%5xzI$Pm9hr&EH_QQS^A7WW}dy zy$wiZu&g5WQXwfn%&YR&YQyD~J%W2Z_bIl{kFV%DQ`t&ENAIbQJKdL2fM%^<+UxjL z7alXsq;ZL*A3j*xUkO|ipt_XjuAD_)M#b{1^yymYins&)w0+J%mj}FN)kPQYj6Z$a zgllFkw_M4#_^$LKd0IyIHz%Ffomp^l3mFSTcau#rwxa-r4QJ$KT@U&{XR)<4htt=|Z<^m*YCdnC05<;~_C%=dKb z?KX|tzT2|Wtng}{-RcE8Rm!QaqD0LP&syj0JjcB@vLB<6yQb7ZOZ#2Cth<1nWVEY* z79Nae{LdL{C6`EOtSE_%>)wMdHD~0dqBW{R4)*H5>(z{p_YZsxNObnudB`Q^^`tcc zb58M3OwO84NpG9APt7fnn`6IW+t*KP)AA!FhNnynPLeOG8MWqd zZ<(b&yg4>|6+(5nZ2BeW9=^M1L*k75OrdVCX&<-hgh+q4KBFRF)CvYKKnFSnu{aCBgUB}|TYgJ>`FH%~Ui521GO zbK|fO!pO6URdd^s8%=SiP4J@Q>f4cb=MX8Lnx~gWv@&^BOh$t z+}CP%k)*!FR)NdM#=1JtyoTY?(ks6Ekw^O01HKtVY}j;MU|{Gh3rc(}*OBjmT2@tk z*LM>W4#y){*~MCVmZ*@fg@^ml#kaOucNw1O(CcfDw_ke=5SX;ic*6+XyvlC7smbeT zE4ow$&YViU?lF9_j_$LfVe`F`RwrHw3JYKEL9Ok-YNV4|Zy|;T*PR1XC1`u??_c@Q zW2M=N9&`}_#sZ$wXEoFF7b{onm&!Yv+@qj&l>x85{5?IH&rZ~c86}$r;ivG#85d|T z6Oz|VdauJ9aATB-T&mZoG_QVQ^t+PtJVpuq)mlBXZStZ*^t=;_Vv|g*g?skQ*DFwy7wkQqs?>2d+_88UKT-(>qfh3Al|L;OmfInkpmM{q+e8ReBQ-oNfECxmyN+P z{$3mH%@4#Y?k(~gBrQIBbWR$Nwx`Q#<;Su{y&1`QTW+kJ^W8wQ*SuUUcV3fdArSHd z6&QwG*&&OhzPdR4alT|z&qOVuV6IVJJ*Xu@XIurR<9APSo&e+g4AnHR@;e)arJgUm zjpVG3A>qOI2hdhYxulHSp|Rhse2B}SLWzwtcJpZrDhj4(0=nNnUl<@d$$QnyzFqIF z4(&IR#;@G%`k+Yt03Kbg)L}JUCh3}=d&!vifmiRcx5Ft)!*B3N2}PGrr{`Gsx+wLL zc)#Gp#(=9UdJA^DwW%M&VEreWcb!z>%*(DX1fRME4zqjgTDjmXmrUn1w#PhAGq){-Ugk!C!lXe;F`4 zPUMsA`sG8^PTPF!h9$y;u&09j>F0Z<4=k_yE}Hlu#?nywxriBeH`zP zeXq^OZaIF`Ta|mNsd=KjP{2i=^sPfe`|h==LxM*RT*pHEx?45Jjb2k{Zuvq;91qQ|o754mmVE;|?ug`tlk1O#+r&JUxMX(}7~ z%s*i-R8+?JA`mI7yL0DOl&K)E{$Rct4L@Yti6S@K;il9((|2nbsY|gG8i7wtccZ=g zkCAoKdw%m)mi<;w0j0;?4pEnV-i<|rRT;Uq@|t~V=Pu6Wc~CHB;{FpZ+?2FxoslfV zLXny}XDMZ2$|LB>D6X*a-`M|8H!tTZo^U%sCEs{E=lSI&yh1J+o1C*HvHX7I5zV@; zpNgZH#+M6}wEXf+VYQI2>b0YC5}d_pd(DfT%V{TSO@_T!F4j<__PvRGVG!H($Ro4tx9%^iMN`@%Q7 zH^yhDnpO`8|Lpu&_NDEjr%syA^aG~7yu-$Gd8oO}3N{nD%yaW$KyWbMEqVXDl6eZt zrXG5|;>(()cZ>Scrkwb|Ws_?yapr2?oZ^_;mGt=v?^Rt7PES!4|8O)#MQyD6`19Vh zoJalzc^h;TsdW7veerJITQ@j0gf`sq@Ilo{@|uEfS_};r6Gj;K^nsO?Me%EwDVz;F zShabHe)S)dUHZlMD4tRMK@*>b5=CtEsQ~fHE3X#^s;Dflt^Tvr;dYe%_Po_Etfi(< zWA~*>eDj^9x+=dg(ZxlR{-&R<`Q|4pE54lzvpZ?Jlq;^*A3 z6e-^+8LEEYjm+LjF>;F-_ePlIY%(j?p1Ee@TK&b0t4p;L%Z8a!kH)PtoNTftcHru) zt_F2@)VTNHIP*%$cUL}otQ;kH?S4(XfFkv3r)i*3tlgo~G9Twc?mZeq?j2ZvPsHer ztId{(P5WLx5fgc4S!~qj_>0B9)&`b$4J@e7)%3Wj`_02{8(hXyOMTAp(y(#7_>ge6 zxUkmoQ--xomS)w25c{J1Wj116EQaX>Y!#-g9oaDJYU^b6@LDT2>nSTFuN*2!^-90*8q$@oPYn*(hP%n1n9pC+qTW?jj+u|RjLPL7$ zXYs~~%Jmy>v*^@y7jMI^l0#htPjt=Ato-=ysQp;^iJ7(sc>7tSi}JKpePXhOy2{X| ztu(H^8ZC0rYh1{msujzg7)cAC?|Y~Gu@28&+0z?6C8kw}ME9OF9c!o}@Z-GR+^EHa z>QH9x-J9<75q4hi&RY<7x&K zOF17`t|_;>e|dCYyui9YOuACTqf#m^G;9l{Q9keTi7$D79sX%kBoj}K!kWm}iR&U0M6$8OyIM-^ zlX%~^0}J04Wplgfj2Dtu4SM!k{1Bz+E>RHeZP@GOd^sWdjXEPSuK4G9TxOD6N~!QX z(E%E->LQ0a@7vo=KzHDiYn}_%PcF&ZV4mMs`J-C8cyP9IP_^6Ec%ggcs@lswtITtd z=_6QnKn`DP=VLs~@uS`%Di}#bmFt`wv2aWO_mQsd30Bp5#g$4bn{cagq9N5)xZapC37k7Y_$eCe4b zTdQ&Hs4L@v<*{zgdF#boybF?iH039h=tm6-*s}MaT+Px)`6YFS3|B`?;rUxmjoufx zRCn9Jz})V&YCCorrhLBA(>d9DyQ?Yh>I)&6>8s`~jY|%yUg#dOg+4Q>rn}jfLA%R} zx9lvjRsC3gG{s+|2R$?)bN<5HN2&ycc0^kW((?w4vo1bfw!!jSjHTdquUGSL&p!}1 z&x6{aY9*-j@l9Tg&dooyHrMs#-ZR*XuD@@AgYr`g1DEM1O@^2TEu3#rnQ5`he_iOc z4>Ycel|1T`tYh=+@|S50wXV%khzoWW_T5_LyJ4aIR{5m?RFHwF@TeO130$dZ8W$IB zaq={ZJ0g=W!gzQ|)Uc0%=wrb>{4T;&q!n$#T)n(9%n z{D?oi^|VtCS`W_bKUP9OH|(&?vV$iF+ay!*ckh;{59+GND4>d}xgERttVX+b%`o3d z&Abp{qjA(rMlB&VC4H__>BXt?eJ69JUf42*G2y~tcae16r}WOG2zGgNd*A!xZ~R4k zU50?px;Mj6y3dq8r*%TTxSvd7JXCbrHGZu;EvmPBpIalAu3sXt^z{1FFk59S$-->D z@{$AJqORKpM9JOj-v6%)h0K)Q`BqG(cpl5DzY>0{PoJ;-URi7zw=uQ2)Q(X(fsrv` z@9j1IGF|ZInV>j3gL_36&$#glqdj=TkIt9PC{+@vTs5k0ad?-Az;PDl>a(-F%F_gF zd9mcLA)EMKk8*pR=rlcyJNZMhOk(m}9sOdOU7)Vjkecm5VYByQxpT4YQ{7T|Iu%vc zC#vL(4s{|jYS47~c|1APW=ApcXwTks$@oGQMM}M|`h#`0I#L_fNvwKK#doRPNYRG! zCZ5XAiPHVk-(MoiH$|1ZT>18hD@wPA=JM!?T#AX|^L7$l6^bRQo><-ux0gJ(B#2)v zoi_a8r8qBcD+hI*_1Oby>J#_^d}dY)?u;6M4}m-K!qqp=dpG=+js{ipe7$;ddX>Vx zw|#uw|H9g~^Oy{NV#cU)K2-HOBWql&FE-CmQzy-w7&_i;yAuAL4u8P58!+xL2@E;^ zBt(cHKoB4Z5CjMU1Ob8oL4Y7Y5FiK;1PB5I0fGQQfFM8+AP5iy2m%BFf&f8)AV3fx z2oMAa0t5kq06~BtKoB4Z5CjMU1Ob8oL4Y983Ix7XBa~c(P(UU^`%@9xlY-EZX9!)* zMd))CLUnbmxHgYd~}H<@D=8&Jz!0kN+DJ5y~s$^d6D+ zH3tED5<)!!*yF$DErhbaH0SQ*T24p68S`J1@SoG!)Nbt~{BJjCJH?F&|2qYn+O3g< z|Lq2Cr?@fUf2UwmyET&Vzulni6gMXP?-Xomw?-2Fw;Qyb;>Lvkoq|p6)=0wtc7wK4 z+?epcQ?RMs8cF!yZqRm$8x#I_3O2P{BMJZ84cbm|W5WMV!KQX=B;kL%LE9;AO!(g^ z*wk)~B>ZnTXgkG?3I96|3_)|>`A#C#{~SUy@L1S_*M`t+W-MBp1M!>Q zur2ssg=2zVl_GTd1#=yT7{aFFiJChO;QFq9%s4gJBOopn;uZvl`{+T8EJIMBpviz|*xh4?iRm)i0%&{F=_;ypeO#{{3sL`eS%GhVp4d~EIY5T9(0 z<;I)um~vvvQTTno*%JQ4e0z=KfOaG^7!?Ixt1VT1<5%SGI=z1PAj|9l=kZ%ai4xkrwe1MR_HKq=zIrneu+P+xFl=zlm zhGThdI$jf6B5Obns~wM+XVl-D!~fpngIG_$Wb!*&*twv zN?30U=R#O>2sMsCYxxiI0OSuk_ynOWd^QBVP+Jr131_g3pcmOZJrm@xgR@F1*#~~g zlGgAac&`?UkT=c?Re*Wao&b3|Ec(A_{rCRT68^*b3;SxyFWOh(Y_Jo{_U!7RKyP7%_0{dYM|H;0 z(Gva-!*eh6HS_HLEBXu~Cg_f4MX|pIfQ>Tt?XPGpXFs#K{D-;Z{P5T3fdn1@pRd^0 z7lKaE6U!>?a6I8`1~<{V=JJ04K2t~IdBh1gf^J~Wb@tW(7`JP0BLw-b3H?Bzx%>w` zKG=XcfuLpq@bq0X{kGq7>l{Y{7 zPJU~i{YM#}qmpnAcut~sSnssrCVTvcb0Fj}C-eh>=FESH)p&vJd7MNgRX9E%jXnO0 zV}B(lEV~l=fk1Qk?|^OePzQ#S2=uz(L%`1Z4>2bYPeSMi0?pz73A}c4S_}@@n?N3A zM+SsABa^rveJ8&)Z~uq$Ht66vjX(zIhUa!uYXHcvu!oTvHa|Gnoca$SJ9Cok&*}xl z0fGNPQ~Zbj&yB`on1jwQtUmmzchIBv!oJh&UzOeoiRKg&wox_!+zCx^n8X;UmUN=4t8g;t_^f&SEcJJIV*MUtR@60w!wY?gnl3Zv9sW7&klZ&j|24WoJ3HQ z7V1^A(|(W<_hMgSLO&3AiFGFs*ToKgkh=xq|2V<+%j!Wc-uoBCu-AUbKLdWmgnl3Z z>*uC>?Cqa0MuJ{&jt&UM;ZtnSg**W4j9lz`VeLjObne0bT7|*6V;zH9I)MIXeQ&Tbyu)dW`437PjfL({8A73%X9oO-AUS z2!Iba$i?j72xn;U0ebn1?ZBYJ^UFYp-B=KfFoZ&0|FhrZ6B7Z5!G#!}o_LP1GX~_a z90<9-Aa3_p5X|9Q1=yb_oW1r!pG@Krf>>lijU#Xy<2mCRdw7B!74+!vonJw)_CO9% z$QjAbcm=N3!?8}hUz?t`Bg zp?^o<1D;!O76d&hJ97(o06vwwu?@MseZnE1Fz_CHczJx-YajH%APS*7SZ773X$0U* z2mbM3BV~7f!F7=D)&4#+CabMuEOC?m@;N<($0gV)K=xo~Y{E65?}c2%oaVFEG$(%D z6_9h|N-jeEFi+XV8Lk5z+>$t)m-h`q;E&Z71TqQ8`CxZ~_$^v9cn`T~AV(pm`75^t zF715}kg*`QXLrBmjRAqMfq_OyH25_2!7tM4**YDxU%~W`L_mRxd8NS;F}FLLx|aiIuNB* z%v@jKg9Ek)@F#(J4mn=H7Ih!Tmw-S%_^G#5 z4yu-+_uu#yti>>QA%{+D#z1}kwCC?I*0$p{hxm84?f3&5>rlL}Lq4tcy<4d;S= zsU+81+j5?HH;@M)pX1c4?CV@x!^`G;7i7IY`0N2WWQnZT9^7Z%6Xby$EDzXWoiDqw z7R}*fORg2b@*&iT0{b=b>25#Xv+fM?0OU}DoD1V4TP~lsgxB@I19Ki^eESDX-(SLe z7Kb|K9po!G-$U#uosb>OPKAb6GOcLsN1Yoa)^C8sXg52$JPHa#4P#)|5 zAm<$T9e_N_iE*eM&xK$AMDQJfaR9NQkf#}Z>_7+D96f#gwVm}ltnspVjh}{n5}@uM z_??1n0`h(l;yeUk9fo}Y^74T!4!S?c%L4Yd(P2#eJ@^iSeg^b+(no>-L4Y7Y5FiK;1PB5Ifwo5g z@?Ssrz^vZ|`89thg8C-2a2&djFEgLzXq>Bp)9To@CqE#@nw3|9^=^i9Y5XoR^^MiF zfO;^$OP;_fIBy$^@2q#oq5r$%hH{AZ1O);k&hWd4)&uRWy>&bc>P|H=B#TK8G+ z|F-^vTnu$1;`0#-&t}#Qyo~dkK`jcXH{8}WPTG5>urvN4*HTs~LhiV>7UbTS!TGql z;oMwMYX@pRLT+slf1JlF7@_GnCmYmWgc<>zftu<+oQENI9n=_MjgO|@A?GFJD}%a; zoSuvDXZ?q|IH@0*HNPO=UQ>N;?EjGS(J-d%YQD6`c=}oYqj28f{^wfO?}k`_P66tj za3VFU{=pgtIZPpU zA?O8Q4}f(Z)?A9|tnwJSrjnKDvC46N>OO79W+S^uHVQfmqFbZSTZ?-)?Q zfxf`n$2)WuhfsAhaZ~qPP-=H>O zTM(T88mm=zE~_o>-uiohOxlqDP`{(C&aF@zsNwxVUjRB2P9Ufq+pvAoxTaZK)hUH} z-thiV5A`9ow{rr49t-%kAk{8rt$xKUl?m)*__gQW;;O71#_|?22j5o zYRxv6p#F&`J~uYD?n908k65?HNd$Wx)b487fA|ct;QjaP*Xe~ha|HLjv2hQ0Kz%4q zBsjA-whjOWP=5mW2{meeR(Aj}OefFM8+AP5iy2m%Cw{}clMNCrSY{B{UHo(L8EGQ+R`&l91DKQy5K)SxGwAqWrz z2m%Cwc0&O4SD*(6eLwj0`(l6o&ZN@)jaaOhSG?+AoOh zN#Ix8S?I6!*Rz2^`@zQAS%|d%*K&&BPw;OI0_=_-@XdUJ{cs_64_>fFu@mRenzpg? z-G6I8*wpS}e+-j&gygYJ4(yW# zH~;J%Vhh0!7vgFUq&^ z=hNKqxeY)6toM2EVy0ZyG+SeLgqVewdGeFW7$`+ymmM>-S-1 zRLAh3vGzkuT62B+Bk`QC?;8t#>>Y8AXxM)6qw&vVAA8o7-?0CN682%N&x5EAh`(#t zemKj3uUK;lZ*_m|7xmi@eirRb4ciZXZ66z1Z`f%E#7@?4KYVU)`h@*UAx@X|+}5IJoL#@|;FEkCpK&^pz=smrT)!OZ4} z7f;#TzCwJ*`7DHbV%d|`9(Lb>2l_Z({a5?%!d!rH53$4z<@%rX1=bqat4i@a<`gRb zj?jfyOrKFj`K-8H zu*s~x#mp-J@tL5vgfld&J#fx`+5+8iV{NYge7HaO$7SL;*M{-`aB((PF%-Cj@PF)X= z>8A8k^>Zt+!!LGP2)-@QC(xULUh`)Jb$ulN);)`0U0_?;ZJ*(_1o8ty8vqN)bzR>s z0+_&Dt8cfVC;hCIE&KbaOs2mNi?@(Z$s`^j;4#FcgRP$RKhQ}x*Y3?utH2)y@?^1| z)9@X9d~C2!NmUIycWc=-;Hz~M>)#=V9ptM3n<3=a*?OO8Z^o=3L=fmS2sBU!bofWm zQ-E4R9>4xyPr-vfG$2j`6zq&IvE1EKf<0vhmStN?bMf9j727OZO7*`FYc=EtYie(; z|19|aBy6XH{No_+!rlmR8z2v}L-2i&5y58xY{W@L2tm#lut~9FGluWOUJLdu7D4;L zem);#+Z2Ip1iqdP$r1Nw2$r#%A~@@SuUW(O9{Qk*<&>rf_G{?>(P#J`f8PXr?O+ds v*cWyP_93_)^5j6?IJh6Qt2ulGAIp - - - {{ site.title | xml_escape }} - {{ site.description | xml_escape }} - {{ site.url }}/ - - {{ site.time | date_to_rfc822 }} - {{ site.time | date_to_rfc822 }} - Jekyll v{{ jekyll.version }} - {% for post in site.posts limit:10 %} - - {{ post.title | xml_escape }} - {{ post.content | xml_escape }} - {{ post.date | date_to_rfc822 }} - {{ post.url | prepend: site.url }} - {{ post.url | prepend: site.url }} - {% for tag in post.tags %} - {{ tag | xml_escape }} - {% endfor %} - {% for tag in page.tags %} - {{ cat | xml_escape }} - {% endfor %} - - {% endfor %} - - diff --git a/docs/internal-components.rst b/docs/internal-components.rst new file mode 100644 index 0000000..67d7290 --- /dev/null +++ b/docs/internal-components.rst @@ -0,0 +1,33 @@ +Internal Components +======================================== + +Ipyannotator has been written in the +`literate programming style `_ +popularized for jupyter notebooks by `nbdev `_. +You can explore the following section to understand more about Ipyannotator's internal codebase. + +.. toctree:: + :maxdepth: 1 + + ../nbs/00a_annotator + ../nbs/00b_mltypes + ../nbs/01_helpers + ../nbs/01a_datasets + ../nbs/01a_datasets_download + ../nbs/01a_datasets_factory + ../nbs/01b_dataset_video + ../nbs/02_navi_widget + ../nbs/02a_right_menu_widget + ../nbs/03_storage + ../nbs/04_bbox_annotator + ../nbs/05_image_button + ../nbs/06_capture_annotator + ../nbs/07_im2im_annotator + ../nbs/12_debug_utils + ../nbs/13_datasets_legacy + ../nbs/14_datasets_factory_legacy + ../nbs/15_coordinates_input + ../nbs/16_custom_buttons + ../nbs/17_annotator_explorer + ../nbs/18_bbox_trajectory + ../nbs/19_bbox_video_annotator \ No newline at end of file diff --git a/docs/sidebar.json b/docs/sidebar.json deleted file mode 100644 index ab9529d..0000000 --- a/docs/sidebar.json +++ /dev/null @@ -1,27 +0,0 @@ -{ - "Getting started": { - "Overview": "/" - }, - "Tutorials": { - "Tutorial: Image classification": "tutorial_image_classification-dsl.html", - "Tutorial: Road damage": "tutorial_road_damage.html", - "Tutorial: BBox": "tutorial_bbox.html" - }, - "Documentatinon": { - "Image classes": "base.html", - "Annotator": "annotator.html", - "Draw Box in Canvas": "bbox_canvas.html", - "Helpers": "helpers.html", - "Helpers for Dataset Generation": "datasets.html", - "Dowload datasets": "datasets_download.html", - "Dataset factory": "datasets_factory.html", - "Navi Widget": "navi_widget.html", - "Storage": "storage.html", - "Bounding Box Annotator": "bbox_annotator.html", - "Image button": "image_button.html", - "Capture annotator": "capture_annotator-dsl.html", - "Image to image annotator": "im2im_annotator-dsl.html", - "Voila example": "viola_example.html" - } -} - diff --git a/docs/sitemap.xml b/docs/sitemap.xml deleted file mode 100644 index 9799959..0000000 --- a/docs/sitemap.xml +++ /dev/null @@ -1,25 +0,0 @@ - ---- -layout: none -search: exclude ---- - - - - {% for post in site.posts %} - {% unless post.search == "exclude" %} - - {{site.url}}{{post.url}} - - {% endunless %} - {% endfor %} - - - {% for page in site.pages %} - {% unless page.search == "exclude" %} - - {{site.url}}{{ page.url}} - - {% endunless %} - {% endfor %} - diff --git a/docs/tutorials.rst b/docs/tutorials.rst new file mode 100644 index 0000000..a67550f --- /dev/null +++ b/docs/tutorials.rst @@ -0,0 +1,18 @@ +Tutorials +======================================== + +Ipyannotator uses tutorials to demonstrate how the library can be used. + +All tutorials are statically generated from jupyter notebooks. +All notebooks can be found on our `Github's repository `_ + +.. toctree:: + :maxdepth: 1 + + ../nbs/01b_tutorial_image_classification + ../nbs/01c_tutorial_bbox + ../nbs/01d_tutorial_video_annotator + ../nbs/08_tutorial_road_damage + ../nbs/09_voila_example + ../nbs/11_build_annotator_tutorial + ../nbs/20_image_classification_user_story \ No newline at end of file diff --git a/index.rst b/index.rst new file mode 100644 index 0000000..4ad8aca --- /dev/null +++ b/index.rst @@ -0,0 +1,15 @@ +.. Ipyannotator documentation master file, created by + sphinx-quickstart on Wed Feb 23 00:18:12 2022. + You can adapt this file completely to your liking, but it should at least + contain the root `toctree` directive. + +.. include:: README.md + :parser: myst_parser.sphinx_ + +.. toctree:: + :caption: Getting started + :maxdepth: 1 + :hidden: + + docs/tutorials + docs/internal-components diff --git a/ipyannotator/__init__.py b/ipyannotator/__init__.py index 49e0fc1..777f190 100644 --- a/ipyannotator/__init__.py +++ b/ipyannotator/__init__.py @@ -1 +1 @@ -__version__ = "0.7.0" +__version__ = "0.8.0" diff --git a/ipyannotator/_nbdev.py b/ipyannotator/_nbdev.py index d4f0827..266e84e 100644 --- a/ipyannotator/_nbdev.py +++ b/ipyannotator/_nbdev.py @@ -2,19 +2,20 @@ __all__ = ["index", "modules", "custom_doc_links", "git_url"] -index = {"validate_project_path": "00_base.ipynb", - "Settings": "00_base.ipynb", - "generate_subset_anno_json": "00_base.ipynb", - "StateSettings": "00_base.ipynb", +index = {"StateSettings": "00_base.ipynb", "BaseState": "00_base.ipynb", + "AnnotatorStep": "00_base.ipynb", "AppWidgetState": "00_base.ipynb", + "Annotator": "00a_annotator.ipynb", + "Settings": "00_base.ipynb", + "generate_subset_anno_json": "00_base.ipynb", + "validate_project_path": "00_base.ipynb", "AnnotatorFactory": "00a_annotator.ipynb", "Bboxer": "00a_annotator.ipynb", "Im2Imer": "00a_annotator.ipynb", "Capturer": "00a_annotator.ipynb", "ImExplorer": "00a_annotator.ipynb", "VideoBboxer": "00a_annotator.ipynb", - "Annotator": "00a_annotator.ipynb", "Coordinate": "00b_mltypes.ipynb", "BboxCoordinate": "00b_mltypes.ipynb", "BboxVideoCoordinate": "00b_mltypes.ipynb", @@ -22,17 +23,30 @@ "Output": "00b_mltypes.ipynb", "InputImage": "00b_mltypes.ipynb", "OutputImageLabel": "00b_mltypes.ipynb", + "OutputLabel": "00b_mltypes.ipynb", "OutputImageBbox": "00b_mltypes.ipynb", "OutputVideoBbox": "00b_mltypes.ipynb", "OutputGridBox": "00b_mltypes.ipynb", "NoOutput": "00b_mltypes.ipynb", + "AnnotationStore": "00c_annotation_types.ipynb", + "LabelStore": "00c_annotation_types.ipynb", + "LabelStoreCaster": "00c_annotation_types.ipynb", + "is_building_docs": "00d_doc_utils.ipynb", + "nbglob": "00d_doc_utils.ipynb", + "upd_metadata": "00d_doc_utils.ipynb", + "hide": "00d_doc_utils.ipynb", + "collapse_cells": "00d_doc_utils.ipynb", "draw_bg": "01_bbox_canvas.ipynb", "draw_bounding_box": "01_bbox_canvas.ipynb", "BoundingBox": "01_bbox_canvas.ipynb", "get_image_size": "01_bbox_canvas.ipynb", - "draw_img": "01_bbox_canvas.ipynb", + "ImageCanvas": "01_bbox_canvas.ipynb", + "ImageCanvasPrototype": "01_bbox_canvas.ipynb", + "CanvasScaleMixin": "01_bbox_canvas.ipynb", + "ScaledImage": "01_bbox_canvas.ipynb", + "FitImage": "01_bbox_canvas.ipynb", + "ImageRenderer": "01_bbox_canvas.ipynb", "points2bbox_coords": "01_bbox_canvas.ipynb", - "coords_point2bbox": "01_bbox_canvas.ipynb", "coords_scaled": "01_bbox_canvas.ipynb", "BBoxLayer": "01_bbox_canvas.ipynb", "BBoxCanvasState": "01_bbox_canvas.ipynb", @@ -88,10 +102,13 @@ "BBoxVideoItem": "02a_right_menu_widget.ipynb", "BBoxList": "02a_right_menu_widget.ipynb", "BBoxVideoList": "02a_right_menu_widget.ipynb", + "Grid": "02b_grid_menu.ipynb", + "GridMenu": "02b_grid_menu.ipynb", "group_files_by_class": "03_storage.ipynb", "construct_annotation_path": "03_storage.ipynb", "setup_project_paths": "03_storage.ipynb", "get_image_list_from_folder": "03_storage.ipynb", + "strip_path": "03_storage.ipynb", "MapeableStorage": "03_storage.ipynb", "AnnotationStorage": "03_storage.ipynb", "JsonLabelStorage": "03_storage.ipynb", @@ -103,15 +120,14 @@ "BBoxAnnotatorGUI": "04_bbox_annotator.ipynb", "BBoxAnnotatorController": "04_bbox_annotator.ipynb", "BBoxAnnotator": "04_bbox_annotator.ipynb", + "ImageButtonSetting": "05_image_button.ipynb", "ImageButton": "05_image_button.ipynb", "CaptureState": "06_capture_annotator.ipynb", - "CaptureGrid": "06_capture_annotator.ipynb", "CaptureAnnotatorGUI": "06_capture_annotator.ipynb", "CaptureAnnotationStorage": "06_capture_annotator.ipynb", "CaptureAnnotatorController": "06_capture_annotator.ipynb", "CaptureAnnotator": "06_capture_annotator.ipynb", "Im2ImState": "07_im2im_annotator.ipynb", - "ImCanvas": "07_im2im_annotator.ipynb", "Im2ImAnnotatorGUI": "07_im2im_annotator.ipynb", "Im2ImAnnotatorController": "07_im2im_annotator.ipynb", "Im2ImAnnotator": "07_im2im_annotator.ipynb", @@ -138,6 +154,8 @@ modules = ["base.py", "annotator.py", "mltypes.py", + "ipytyping/annotations.py", + "doc_utils.py", "bbox_canvas.py", "helpers.py", "datasets/generators.py", @@ -145,16 +163,16 @@ "datasets/factory.py", "navi_widget.py", "right_menu_widget.py", + "custom_widgets/grid_menu.py", "storage.py", "bbox_annotator.py", - "image_button.py", + "custom_input/buttons.py", "capture_annotator.py", "im2im_annotator.py", "debug_utils.py", "datasets/generators_legacy.py", "datasets/factory_legacy.py", "custom_input/coordinates.py", - "custom_input/buttons.py", "explore_annotator.py", "services/bbox_trajectory.py", "bbox_video_annotator.py"] diff --git a/ipyannotator/annotator.py b/ipyannotator/annotator.py index cd17ef2..6f14549 100644 --- a/ipyannotator/annotator.py +++ b/ipyannotator/annotator.py @@ -6,14 +6,14 @@ import json from abc import ABC, abstractmethod from pathlib import Path -from typing import Tuple, Type +from typing import Tuple, Type, List from skimage import io -from tqdm import tqdm +from tqdm.notebook import tqdm -from .base import generate_subset_anno_json, Settings +from .base import generate_subset_anno_json, Settings, AnnotatorStep from .mltypes import (Input, Output, OutputVideoBbox, - InputImage, OutputImageLabel, + InputImage, OutputImageLabel, OutputLabel, OutputImageBbox, OutputGridBox, NoOutput) from .bbox_annotator import BBoxAnnotator from .bbox_video_annotator import BBoxVideoAnnotator @@ -25,14 +25,19 @@ # Internal Cell class AnnotatorFactory(ABC): - io: Tuple[Type[Input], Type[Output]] + io: Tuple[Type[Input], List[Type[Output]]] @abstractmethod def get_annotator(self): pass def __new__(cls, input_item, output_item): - subclass_map = {subclass.io: subclass for subclass in cls.__subclasses__()} + subclass_map = {} + + for subclass in cls.__subclasses__(): + for subclass_output in subclass.io[1]: + subclass_map[(subclass.io[0], subclass_output)] = subclass + try: subclass = subclass_map[(type(input_item), type(output_item))] instance = super(AnnotatorFactory, subclass).__new__(subclass) @@ -40,41 +45,37 @@ def __new__(cls, input_item, output_item): except KeyError: print(f"Pair {(input_item, output_item)} is not supported!") -# -# Define all supported annotators with correct Input/Output pairs for internal use below: -# - class Bboxer(AnnotatorFactory): - io = (InputImage, OutputImageBbox) + io = (InputImage, [OutputImageBbox]) def get_annotator(self): return BBoxAnnotator class Im2Imer(AnnotatorFactory): - io = (InputImage, OutputImageLabel) + io = (InputImage, [OutputImageLabel, OutputLabel]) def get_annotator(self): return Im2ImAnnotator class Capturer(AnnotatorFactory): - io = (InputImage, OutputGridBox) + io = (InputImage, [OutputGridBox]) def get_annotator(self): return CaptureAnnotator class ImExplorer(AnnotatorFactory): - io = (InputImage, NoOutput) + io = (InputImage, [NoOutput]) def get_annotator(self): return ExploreAnnotator class VideoBboxer(AnnotatorFactory): - io = (InputImage, OutputVideoBbox) + io = (InputImage, [OutputVideoBbox]) def get_annotator(self): return BBoxVideoAnnotator @@ -105,13 +106,18 @@ def explore(self, k=-1): annotator = AnnotatorFactory(self.input_item, self.output_item).get_annotator() - return annotator(project_path=self.settings.project_path, - input_item=self.input_item, - output_item=self.output_item, - annotation_file_path=anno_, - n_cols=self.settings.n_cols, - question="Classification ", - has_border=True) + self.output_item.drawing_enabled = False + annotator = annotator(project_path=self.settings.project_path, + input_item=self.input_item, + output_item=self.output_item, + annotation_file_path=anno_, + n_cols=self.settings.n_cols, + question="Classification ", + has_border=True) + + annotator.app_state.annotation_step = AnnotatorStep.EXPLORE + + return annotator def create(self): anno_ = construct_annotation_path(project_path=self.settings.project_path, @@ -120,13 +126,18 @@ def create(self): annotator = AnnotatorFactory(self.input_item, self.output_item).get_annotator() - return annotator(project_path=self.settings.project_path, - input_item=self.input_item, - output_item=self.output_item, - annotation_file_path=anno_, - n_cols=self.settings.n_cols, - question="Classification ", - has_border=True) + self.output_item.drawing_enabled = True + annotator = annotator(project_path=self.settings.project_path, + input_item=self.input_item, + output_item=self.output_item, + annotation_file_path=anno_, + n_cols=self.settings.n_cols, + question="Classification ", + has_border=True) + + annotator.app_state.annotation_step = AnnotatorStep.CREATE + + return annotator def improve(self): # open labels from create step @@ -140,7 +151,6 @@ def improve(self): if type(self.output_item) == OutputImageLabel: #Construct multiple Capturers for each class - # out = [] for class_name, class_anno in tqdm( group_files_by_class(loaded_image_annotations).items()): @@ -181,7 +191,6 @@ def improve(self): captured_path = Path(self.settings.project_path) / "captured" # Save annotated images on disk - # for im, bbx in tqdm(di.items()): # use captured_path instead image_dir, keeping the folder structure old_im_path = Path(im) @@ -212,5 +221,12 @@ def improve(self): else: raise Exception(f"Improve is not supported for {self.output_item}") + if isinstance(out, list): + def update_step(anno): + anno.app_state.annotation_step = AnnotatorStep.IMPROVE + return anno + out = [update_step(anno) for anno in out] + else: + out.app_state.annotation_step = AnnotatorStep.IMPROVE return out \ No newline at end of file diff --git a/ipyannotator/base.py b/ipyannotator/base.py index f10b649..67b9328 100644 --- a/ipyannotator/base.py +++ b/ipyannotator/base.py @@ -3,23 +3,89 @@ __all__ = [] # Internal Cell - import json import random from pubsub import pub from pathlib import Path +from enum import Enum, auto from typing import NamedTuple, Optional, Tuple, Any, Callable +from abc import ABC from pydantic import BaseModel, BaseSettings # Internal Cell -def validate_project_path(project_path): - project_path = Path(project_path) - assert project_path.exists(), "WARNING: Project path should point to " \ - "existing directory" - assert project_path.is_dir(), "WARNING: Project path should point to " \ - "existing directory" - return project_path +class StateSettings(BaseSettings): + class Config: + validate_assignment = True + + +class BaseState(StateSettings, BaseModel): + def __init__(self, uuid: str = None, *args, **kwargs): + super().__init__(*args, **kwargs) + self.set_quietly('_uuid', uuid) + self.set_quietly('event_map', {}) + + def set_quietly(self, key: str, value: Any): + """ + Assigns a value to a state's attribute. + + This function can be used to avoid that + the state dispatches a PyPubSub event. + It's very usefull to avoid event recursion, + ex: a component is listening for an event A + but it also changes the state that dispatch + the event A. Using set_quietly to set the + value at the component will avoid the recursion. + """ + object.__setattr__(self, key, value) + + @property + def root_topic(self) -> str: + if hasattr(self, '_uuid') and self._uuid: # type: ignore + return f'{self._uuid}.{type(self).__name__}' # type: ignore + + return type(self).__name__ + + def subscribe(self, change: Callable, attribute: str): + key = f'{self.root_topic}.{attribute}' + self.event_map[key] = change # type: ignore + pub.subscribe(change, key) + + def unsubscribe(self, attribute: str): + key = self.topic_attribute(attribute) + pub.unsubscribe(self.event_map[key], key) # type: ignore + del self.event_map[key] # type: ignore + + def topic_attribute(self, attribute: str): + return f'{self.root_topic}.{attribute}' + + def is_subscribed(self, attribute: str) -> bool: + return attribute in self.event_map # type: ignore + + def __setattr__(self, key: str, value: Any): + self.set_quietly(key, value) + + if key != '__class__': + pub.sendMessage(f'{self.root_topic}.{key}', **{key: value}) + +# Internal Cell +class AnnotatorStep(Enum): + EXPLORE = auto() + CREATE = auto() + IMPROVE = auto() + +# Internal Cell + +class AppWidgetState(BaseState): + annotation_step: AnnotatorStep = AnnotatorStep.CREATE + size: Tuple[int, int] = (640, 400) + max_im_number: int = 1 + index: int = 0 + +# Internal Cell +class Annotator(ABC): + def __init__(self, app_state: AppWidgetState): + self.app_state = app_state # Internal Cell @@ -68,50 +134,10 @@ def generate_subset_anno_json(project_path: Path, project_file, return subset_file # Internal Cell - -class StateSettings(BaseSettings): - class Config: - validate_assignment = True - - -class BaseState(StateSettings, BaseModel): - def __init__(self, uuid: str = None, *args, **kwargs): - super().__init__(*args, **kwargs) - self.set_quietly('_uuid', uuid) - - def set_quietly(self, key: str, value: Any): - """ - Assigns a value to a state's attribute. - - This function can be used to avoid that - the state dispatches a PyPubSub event. - It's very usefull to avoid event recursion, - ex: a component is listening for an event A - but it also changes the state that dispatch - the event A. Using set_quietly to set the - value at the component will avoid the recursion. - """ - object.__setattr__(self, key, value) - - @property - def root_topic(self) -> str: - if hasattr(self, '_uuid') and self._uuid: # type: ignore - return f'{self._uuid}.{type(self).__name__}' # type: ignore - - return type(self).__name__ - - def subscribe(self, change: Callable, attribute: str): - pub.subscribe(change, f'{self.root_topic}.{attribute}') - - def __setattr__(self, key: str, value: Any): - self.set_quietly(key, value) - - if key != '__class__': - pub.sendMessage(f'{self.root_topic}.{key}', **{key: value}) - -# Internal Cell - -class AppWidgetState(BaseState): - size: Tuple[int, int] = (640, 400) - max_im_number: int = 1 - index: int = 0 \ No newline at end of file +def validate_project_path(project_path): + project_path = Path(project_path) + assert project_path.exists(), "WARNING: Project path should point to " \ + "existing directory" + assert project_path.is_dir(), "WARNING: Project path should point to " \ + "existing directory" + return project_path \ No newline at end of file diff --git a/ipyannotator/bbox_annotator.py b/ipyannotator/bbox_annotator.py index 583f694..c519acb 100644 --- a/ipyannotator/bbox_annotator.py +++ b/ipyannotator/bbox_annotator.py @@ -15,7 +15,8 @@ from ipywidgets import AppLayout, Button, HBox, VBox, Layout from .mltypes import BboxCoordinate -from .base import BaseState, AppWidgetState +from .base import BaseState, AppWidgetState, Annotator +from .mltypes import InputImage, OutputImageBbox from .bbox_canvas import BBoxCanvas, BBoxCanvasState from .navi_widget import Navi from .right_menu_widget import BBoxList, BBoxVideoItem @@ -29,6 +30,7 @@ class BBoxState(BaseState): image: Optional[Path] classes: List[str] labels: List[List[str]] = [] + drawing_enabled: bool = True # Internal Cell @@ -47,15 +49,9 @@ def __init__( self._app_state = app_state self._bbox_state = bbox_state self._bbox_canvas_state = bbox_canvas_state + self.on_btn_select_clicked = on_btn_select_clicked - self._bbox_list = BBoxList( - max_coord_input_values=None, - on_coords_changed=self.on_coords_change, - on_label_changed=self.on_label_change, - on_btn_delete_clicked=self.on_btn_delete_clicked, - on_btn_select_clicked=on_btn_select_clicked, - classes=bbox_state.classes - ) + self._init_bbox_list(self._bbox_state.drawing_enabled) if self._bbox_canvas_state.bbox_coords: self._bbox_list.render_btn_list( @@ -64,6 +60,7 @@ def __init__( ) app_state.subscribe(self._refresh_children, 'index') + bbox_state.subscribe(self._init_bbox_list, 'drawing_enabled') bbox_canvas_state.subscribe(self._sync_labels, 'bbox_coords') self._bbox_canvas_state.subscribe(self._update_max_coord_input, 'image_scale') self._update_max_coord_input(self._bbox_canvas_state.image_scale) @@ -73,6 +70,19 @@ def __init__( display='block' ) + def _init_bbox_list(self, drawing_enabled: bool): + self._bbox_list = BBoxList( + max_coord_input_values=None, + on_coords_changed=self.on_coords_change, + on_label_changed=self.on_label_change, + on_btn_delete_clicked=self.on_btn_delete_clicked, + on_btn_select_clicked=self.on_btn_select_clicked, + classes=self._bbox_state.classes, + readonly=not drawing_enabled + ) + + self._refresh_children(0) + def __getitem__(self, index: int) -> BBoxVideoItem: return self.children[index] @@ -134,18 +144,21 @@ def _update_max_coord_input(self, image_scale: float): self._bbox_list.max_coord_input_values = BboxCoordinate(*coords) # Internal Cell - class BBoxAnnotatorGUI(AppLayout): def __init__( self, app_state: AppWidgetState, bbox_state: BBoxState, - on_save_btn_clicked: Callable = None + fit_canvas: bool, + on_save_btn_clicked: Callable = None, + has_border: bool = False ): self._app_state = app_state self._bbox_state = bbox_state self._on_save_btn_clicked = on_save_btn_clicked self._label_history: List[List[str]] = [] + self.fit_canvas = fit_canvas + self.has_border = has_border self._navi = Navi() @@ -169,7 +182,7 @@ def __init__( ) ) - self._image_box = BBoxCanvas(*self._app_state.size) + self._init_canvas(self._bbox_state.drawing_enabled) self.right_menu = BBoxCoordinates( app_state=self._app_state, @@ -200,6 +213,7 @@ def __init__( self._redo_btn.on_click(self._redo_clicked) bbox_state.subscribe(self._set_image_path, 'image') + bbox_state.subscribe(self._init_canvas, 'drawing_enabled') bbox_state.subscribe(self._set_coords, 'coords') app_state.subscribe(self._set_max_im_number, 'max_im_number') @@ -212,6 +226,14 @@ def __init__( pane_widths=(2, 8, 0), pane_heights=(1, 4, 1)) + def _init_canvas(self, drawing_enabled: bool): + self._image_box = BBoxCanvas( + *self._app_state.size, + drawing_enabled=drawing_enabled, + fit_canvas=self.fit_canvas, + has_border=self.has_border + ) + def _highlight_bbox(self, btn: ActionButton): self._image_box.highlight = btn.value @@ -258,7 +280,6 @@ def on_client_ready(self, callback): self._image_box.observe_client_ready(callback) # Internal Cell - class BBoxAnnotatorController: def __init__( self, @@ -280,7 +301,7 @@ def __init__( if render_previous_coords: self._update_coords(self._last_index) - def save_current_annotations(self, coords: dict): + def save_current_annotations(self, coords: List[BboxCoordinate]): self._bbox_state.set_quietly('coords', coords) self._save_annotations(self._app_state.index) @@ -319,7 +340,7 @@ def handle_client_ready(self): # Cell -class BBoxAnnotator: +class BBoxAnnotator(Annotator): """ Represents bounding box annotator. @@ -332,24 +353,28 @@ class BBoxAnnotator: def __init__( self, project_path: Path, - input_item, - output_item, + input_item: InputImage, + output_item: OutputImageBbox, annotation_file_path: Path, + has_border: bool = False, *args, **kwargs ): - self.app_state = AppWidgetState( + app_state = AppWidgetState( uuid=str(id(self)), **{ 'size': (input_item.width, input_item.height), } ) + super().__init__(app_state) + self._input_item = input_item self._output_item = output_item self.bbox_state = BBoxState( uuid=str(id(self)), - classes=output_item.classes + classes=output_item.classes, + drawing_enabled=self._output_item.drawing_enabled ) self.storage = JsonCaptureStorage( @@ -367,7 +392,9 @@ def __init__( self.view = BBoxAnnotatorGUI( app_state=self.app_state, bbox_state=self.bbox_state, - on_save_btn_clicked=self.controller.save_current_annotations + fit_canvas=self._input_item.fit_canvas, + on_save_btn_clicked=self.controller.save_current_annotations, + has_border=has_border ) self.view.on_client_ready(self.controller.handle_client_ready) diff --git a/ipyannotator/bbox_canvas.py b/ipyannotator/bbox_canvas.py index cde84be..8f59632 100644 --- a/ipyannotator/bbox_canvas.py +++ b/ipyannotator/bbox_canvas.py @@ -1,8 +1,10 @@ # AUTOGENERATED! DO NOT EDIT! File to edit: nbs/01_bbox_canvas.ipynb (unless otherwise specified). -__all__ = ['draw_img', 'points2bbox_coords', 'coords_point2bbox', 'coords_scaled', 'BBoxCanvas', 'BBoxVideoCanvas'] +__all__ = ['points2bbox_coords', 'coords_scaled', 'BBoxCanvas', 'BBoxVideoCanvas'] # Internal Cell +import io +import attr from math import log from pubsub import pub from attr import asdict @@ -11,21 +13,78 @@ from enum import IntEnum from typing import Dict, Optional, List, Any, Tuple +from abc import ABC, abstractmethod from pydantic import root_validator from .base import BaseState +from .doc_utils import is_building_docs from .mltypes import BboxCoordinate, BboxVideoCoordinate -from ipycanvas import MultiCanvas, Canvas, hold_canvas +from ipycanvas import MultiCanvas as IMultiCanvas, Canvas, hold_canvas from ipywidgets import Image, Label, Layout, HBox, VBox, Output +from PIL import Image as PILImage # Internal Cell +if not is_building_docs(): + class MultiCanvas(IMultiCanvas): + pass +else: + class MultiCanvas(Image): # type: ignore + def __init__(self, *args, **kwargs): + super().__init__(**kwargs) + image = PILImage.new('RGB', (100, 100), (255, 255, 255)) + b = io.BytesIO() + image.save(b, format='PNG') + self.value = b.getvalue() + + def __getitem__(self, key): + return self + + def draw_image(self, image, x=0, y=0, width=None, height=None): + self.value = image.value + self.width = width + self.height = height + + def __getattr__(self, name): + ignored = [ + 'flush', + 'fill_rect', + 'stroke_rect', + 'stroke_rects', + 'on_mouse_move', + 'on_mouse_down', + 'on_mouse_up', + 'clear', + 'on_client_ready', + 'stroke_styled_line_segments' + ] + + if name in ignored: + def wrapper(*args, **kwargs): + return self._ignored(*args, **kwargs) + return wrapper + return object.__getattr__(self, name) + + @property + def caching(self): + return False + + @caching.setter + def caching(self, value): + pass + + @property + def size(self): + return (self.width, self.height) + + def _ignored(self, *args, **kwargs): + pass +# Internal Cell def draw_bg(canvas, color='rgb(236,240,241)'): with hold_canvas(canvas): canvas.fill_style = color canvas.fill_rect(0, 0, canvas.size[0], canvas.size[1]) # Internal Cell - def draw_bounding_box(canvas, coord: BboxCoordinate, color='white', line_width=1, border_ratio=2, clear=False, stroke_color='black'): with hold_canvas(canvas): @@ -54,7 +113,6 @@ def draw_bounding_box(canvas, coord: BboxCoordinate, color='white', line_width=1 coord.width - 2 * gap, coord.height - 2 * gap) # Internal Cell - class BoundingBox: def __init__(self): self.color = 'white' @@ -118,48 +176,118 @@ def get_image_size(path): pil_im = pilImage.open(path) return pil_im.width, pil_im.height -# Cell - -def draw_img(canvas, file, clear=False, has_border=False) -> Tuple[int, int, float]: - """ - draws resized image on canvas and returns scale used - """ - with hold_canvas(canvas): - if clear: - canvas.clear() - - sprite1 = Image.from_file(file) +# Internal Cell +@attr.define +class ImageCanvas: + image_widget: Image + x: int + y: int + width: int + height: int + scale: float - width_canvas, height_canvas = canvas.width, canvas.height - width_img, height_img = get_image_size(file) +# Internal Cell +class ImageCanvasPrototype(ABC): + @abstractmethod + def prepare_canvas(self, canvas: Canvas, file: str) -> ImageCanvas: + pass +# Internal Cell +class CanvasScaleMixin: + def _calc_scale( + self, + width_canvas: int, + height_canvas: int, + width_img: float, + height_img: float + ) -> float: ratio_canvas = float(width_canvas) / height_canvas ratio_img = float(width_img) / height_img if ratio_img > ratio_canvas: # wider then canvas, scale to canvas width - scale = width_canvas / width_img - else: - # taller then canvas, scale to canvas hight - scale = height_canvas / height_img + return width_canvas / width_img + + # taller then canvas, scale to canvas height + return height_canvas / height_img + +# Internal Cell +class ScaledImage(ImageCanvasPrototype, CanvasScaleMixin): + def prepare_canvas(self, canvas: Canvas, file: str) -> ImageCanvas: + image = Image.from_file(file) + width_img, height_img = get_image_size(file) + + scale = self._calc_scale( + int(canvas.width), + int(canvas.height), + width_img, + height_img + ) image_width = width_img * min(1, scale) image_height = height_img * min(1, scale) - image_x = 0 - image_y = 0 - - if has_border: - canvas.stroke_rect(x=0, y=0, width=image_width, height=image_height) - image_width -= 2 - image_height -= 2 - image_x, image_y = 1, 1 - - canvas.draw_image(sprite1, - image_x, - image_y, - width=image_width, - height=image_height) - return (image_width, image_height, scale) + + return ImageCanvas( + image_widget=image, + x=0, + y=0, + width=image_width, + height=image_height, + scale=scale + ) + +# Internal Cell +class FitImage(ImageCanvasPrototype): + def prepare_canvas(self, canvas: Canvas, file: str) -> ImageCanvas: + image = Image.from_file(file) + + return ImageCanvas( + image_widget=image, + x=0, + y=0, + width=canvas.width, + height=canvas.height, + scale=1 + ) + +# Internal Cell +class ImageRenderer: + def __init__( + self, + clear: bool = False, + has_border: bool = False, + fit_canvas: bool = False + ): + self.clear = clear + self.has_border = has_border + self.fit_canvas = fit_canvas + if fit_canvas: + self._strategy = FitImage() # type: ImageCanvasPrototype + else: + self._strategy = ScaledImage() + + def render(self, canvas: Canvas, file: str) -> Tuple[int, int, float]: + with hold_canvas(canvas): + if self.clear: + canvas.clear() + + image_canvas = self._strategy.prepare_canvas(canvas, file) + + if self.has_border: + canvas.stroke_rect(x=0, y=0, width=image_canvas.width, height=image_canvas.height) + image_canvas.width -= 2 + image_canvas.height -= 2 + image_canvas.x, image_canvas.y = 1, 1 + + canvas.draw_image( + image_canvas.image_widget, + image_canvas.x, + image_canvas.y, + image_canvas.width, + image_canvas.height + ) + + return image_canvas.width, image_canvas.height, image_canvas.scale # Cell @@ -169,15 +297,6 @@ def points2bbox_coords(start_x, start_y, end_x, end_y) -> Dict[str, float]: return {'x': min_x, 'y': min_y, 'width': max_x - min_x, 'height': max_y - min_y} # Cell - -def coords_point2bbox(bbox_coords: Dict[str, float]) -> List[float]: - return [bbox_coords['x'], - bbox_coords['y'], - bbox_coords['width'], - bbox_coords['height']] - -# Cell - def coords_scaled(bbox_coords: List[float], image_scale: float): return [value * image_scale for value in bbox_coords] @@ -201,6 +320,7 @@ class BBoxCanvasState(BaseState): bbox_selected: Optional[int] height: Optional[int] width: Optional[int] + fit_canvas: bool = False @root_validator def set_height(cls, values): @@ -217,7 +337,12 @@ def set_height(cls, values): class BBoxCanvasGUI(HBox): debug_output = Output(layout={'border': '1px solid black'}) - def __init__(self, state: BBoxCanvasState, has_border: bool = False): + def __init__( + self, + state: BBoxCanvasState, + has_border: bool = False, + drawing_enabled: bool = True + ): super().__init__() self._state = state @@ -225,6 +350,7 @@ def __init__(self, state: BBoxCanvasState, has_border: bool = False): self.is_drawing = False self.has_border = has_border self.canvas_bbox_coords: Dict[str, Any] = {} + self.drawing_enabled = drawing_enabled # do not stick bbox to borders self.padding = 2 @@ -235,22 +361,31 @@ def __init__(self, state: BBoxCanvasState, has_border: bool = False): align_items='center', align_content='center', overflow='hidden')) - self.multi_canvas = MultiCanvas( - len(BBoxLayer), - width=self._state.width, - height=self._state.height - ) - self.im_name_box = Label() + if not drawing_enabled: + self.multi_canvas = MultiCanvas( + len(BBoxLayer), + width=self._state.width, + height=self._state.height + ) + self.children = [VBox([self.multi_canvas])] + else: + self.multi_canvas = MultiCanvas( + len(BBoxLayer), + width=self._state.width, + height=self._state.height + ) + + self.im_name_box = Label() - children = [VBox([self.multi_canvas, self.im_name_box])] - self.children = children - draw_bg(self.multi_canvas[BBoxLayer.bg]) + children = [VBox([self.multi_canvas, self.im_name_box])] + self.children = children + draw_bg(self.multi_canvas[BBoxLayer.bg]) - # link drawing events - self.multi_canvas[BBoxLayer.drawing].on_mouse_move(self._update_pos) - self.multi_canvas[BBoxLayer.drawing].on_mouse_down(self._start_drawing) - self.multi_canvas[BBoxLayer.drawing].on_mouse_up(self._stop_drawing) + # link drawing events + self.multi_canvas[BBoxLayer.drawing].on_mouse_move(self._update_pos) + self.multi_canvas[BBoxLayer.drawing].on_mouse_down(self._start_drawing) + self.multi_canvas[BBoxLayer.drawing].on_mouse_up(self._stop_drawing) @property def highlight(self) -> BboxCoordinate: @@ -279,7 +414,7 @@ def highlight(self, index: int): self._state.set_quietly('bbox_selected', index) - @debug_output.capture(clear_output=False) + @debug_output.capture(clear_output=True) def _update_pos(self, x, y): # print(f"-> BBoxCanvasGUI::_update_post({x}, {y})") if self.is_drawing: @@ -300,7 +435,7 @@ def _invalid_coords(self, x, y) -> bool: self.canvas_bbox_coords["x"] < self.padding or self.canvas_bbox_coords["y"] < self.padding) - @debug_output.capture(clear_output=False) + @debug_output.capture(clear_output=True) def _stop_drawing(self, x, y): # print(f"-> BBoxCanvasGUI::_stop_drawing({x}, {y})") self.is_drawing = False @@ -361,8 +496,13 @@ def observe_client_ready(self, cb=None): class BBoxVideoCanvasGUI(BBoxCanvasGUI): debug_output = Output(layout={'border': '1px solid black'}) - def __init__(self, state: BBoxCanvasState, has_border: bool = False): - super().__init__(state, has_border) + def __init__( + self, + state: BBoxCanvasState, + has_border: bool = False, + drawing_enabled: bool = True + ): + super().__init__(state, has_border, drawing_enabled) @property def highlight(self) -> BboxCoordinate: @@ -479,14 +619,20 @@ def clear_all_bbox(self): @debug_output.capture(clear_output=True) def _draw_image(self, image_path: str): - # print(f"-> _draw_image {image_path}") + print(f"-> _draw_image {image_path}") self.clear_all_bbox() - image_width, image_height, scale = draw_img( - self._gui.multi_canvas[BBoxLayer.image], - image_path, + + img_renderer_service = ImageRenderer( clear=True, - has_border=self._gui.has_border + has_border=self._gui.has_border, + fit_canvas=self._state.fit_canvas + ) + + image_width, image_height, scale = img_renderer_service.render( + self._gui.multi_canvas[BBoxLayer.image], + image_path ) + self._state.set_quietly('image_width', image_width) self._state.set_quietly('image_height', image_height) self._state.image_scale = scale @@ -528,12 +674,23 @@ class BBoxCanvas(BBoxCanvasGUI): Gives user an ability to draw a bbox with mouse. """ - def __init__(self, width, height, has_border: bool = False): + def __init__( + self, + width, + height, + has_border: bool = False, + fit_canvas: bool = False, + drawing_enabled: bool = True + ): self.state = BBoxCanvasState( uuid=str(id(self)), - **{'width': width, 'height': height} + **{'width': width, 'height': height, 'fit_canvas': fit_canvas} + ) + super().__init__( + state=self.state, + has_border=has_border, + drawing_enabled=drawing_enabled ) - super().__init__(state=self.state, has_border=has_border) self._controller = BBoxCanvasController(gui=self, state=self.state) self._bbox_history: List[Any] = [] @@ -565,13 +722,6 @@ def __init__(self, width, height, has_border: bool = False, drawing_enabled: boo **{'width': width, 'height': height} ) self.drawing_enabled = drawing_enabled - super().__init__(state=self.state, has_border=has_border) - if not drawing_enabled: - self.multi_canvas = MultiCanvas( - len(BBoxLayer), - width=self._state.width, - height=self._state.height - ) - self.children = [VBox([self.multi_canvas, self.im_name_box])] + super().__init__(state=self.state, has_border=has_border, drawing_enabled=drawing_enabled) self._controller = BBoxVideoCanvasController(gui=self, state=self.state) \ No newline at end of file diff --git a/ipyannotator/bbox_video_annotator.py b/ipyannotator/bbox_video_annotator.py index ef15618..cbc9c0f 100644 --- a/ipyannotator/bbox_video_annotator.py +++ b/ipyannotator/bbox_video_annotator.py @@ -60,6 +60,7 @@ def __init__( on_trajectory_enabled_clicked: Callable, on_btn_delete_clicked: Callable[[BboxVideoCoordinate], None] ): + self.on_label_changed = on_label_changed super().__init__( app_state, bbox_canvas_state, @@ -79,22 +80,26 @@ def __init__( if on_trajectory_enabled_clicked: self.trajectory_enabled_checkbox.observe(on_trajectory_enabled_clicked, names='value') + self._bbox_state.unsubscribe('drawing_enabled') pub.unsubscribe(super()._sync_labels, f'{bbox_canvas_state.root_topic}.bbox_coords') pub.unsubscribe(super()._refresh_children, f'{app_state.root_topic}.index') + self._init_bbox_list(self._bbox_state.drawing_enabled) + + bbox_canvas_state.subscribe(self._update_max_coord_input, 'image_scale') + + self.children = self._bbox_list.children + + def _init_bbox_list(self, drawing_enabled: bool): self._bbox_list = BBoxVideoList( btn_delete_enabled=drawing_enabled, - on_label_changed=on_label_changed, + on_label_changed=self.on_label_changed, on_btn_delete_clicked=self._on_btn_delete_clicked, - on_btn_select_clicked=on_btn_select_clicked, - classes=bbox_state.classes, + on_btn_select_clicked=self.on_btn_select_clicked, + classes=self._bbox_state.classes, on_checkbox_object_clicked=self._on_checkbox_object_clicked ) - bbox_canvas_state.subscribe(self._update_max_coord_input, 'image_scale') - - self.children = self._bbox_list.children - def _refresh_children(self, index: int): self._render( self._bbox_canvas_state.bbox_coords, @@ -184,7 +189,8 @@ def __init__( super().__init__( app_state=app_state, bbox_state=bbox_state, - on_save_btn_clicked=on_save_btn_clicked + on_save_btn_clicked=on_save_btn_clicked, + fit_canvas=False ) self._app_state = app_state @@ -192,6 +198,7 @@ def __init__( self.on_bbox_drawn = on_bbox_drawn self.bbox_trajectory = BBoxTrajectory() self.history = BboxVideoHistory() + self.on_label_changed = on_label_changed pub.unsubAll(f'{self._image_box.state.root_topic}.bbox_coords') @@ -208,7 +215,7 @@ def __init__( bbox_state=self._bbox_state, # type: ignore on_btn_select_clicked=self._highlight_bbox, on_btn_delete_clicked=self._remove_trajectory_history, - on_label_changed=on_label_changed, + on_label_changed=self.on_label_changed, drawing_enabled=drawing_enabled, on_trajectory_enabled_clicked=self.on_trajectory_enabled_clicked ) @@ -231,7 +238,7 @@ def __init__( self.btn_right_menu_enabled = ToggleButton( description="Menu", - tooltip="Disable right menu for a better navigation experience.", + tooltip="Disable right menu for a faster navigation experience.", icon="eye-slash", disabled=False, # Argument 1 to "render_right_menu" of "BBoxAnnotatorVideoGUI" has incompatible @@ -471,7 +478,8 @@ def __init__(self, *args, **kwargs): pub.unsubscribe(self.controller._idx_changed, f'{self.app_state.root_topic}.index') pub.unsubAll(f'{self.app_state.root_topic}.index') state_params = {**self.bbox_state.dict()} - state_params.pop('_uuid') + state_params.pop('_uuid', []) + state_params.pop('event_map', []) self.bbox_state = BBoxVideoState( uuid=self.bbox_state._uuid, **state_params @@ -502,8 +510,8 @@ def update_labels(self, change: dict, index: int): # "BBoxAnnotatorController" has no attribute "update_storage_labels" self.controller.update_storage_labels(change, index) # type: ignore - def on_save_btn_clicked(self, bbox_coords: Dict): - self.controller.save_current_annotations(bbox_coords) + def on_save_btn_clicked(self, bbox_coords: List[BboxVideoCoordinate]): + self.controller.save_current_annotations(bbox_coords) # type: ignore def _update_state_id(self, merged_ids: List[str], bbox_coords: List[BboxVideoCoordinate]): merged_id = "-".join(merged_ids) diff --git a/ipyannotator/capture_annotator.py b/ipyannotator/capture_annotator.py index 5ac979b..08e94db 100644 --- a/ipyannotator/capture_annotator.py +++ b/ipyannotator/capture_annotator.py @@ -1,100 +1,31 @@ # AUTOGENERATED! DO NOT EDIT! File to edit: nbs/06_capture_annotator.ipynb (unless otherwise specified). -__all__ = ['CaptureGrid', 'CaptureAnnotator'] +__all__ = ['CaptureAnnotator'] # Internal Cell import math import warnings -from functools import partial +from copy import deepcopy from pathlib import Path -from typing import Dict, Optional, List, Iterable, Callable +from typing import Dict, Optional, Callable, List -from IPython.core.display import display -from ipywidgets import (AppLayout, VBox, HBox, Button, GridBox, - Layout, Checkbox, HTML, IntText, Output) +from IPython.display import display +from ipywidgets import (AppLayout, HBox, Button, HTML, VBox, + Layout, Checkbox, Output) -from .base import BaseState, AppWidgetState -from .image_button import ImageButton +from .custom_widgets.grid_menu import GridMenu, Grid +from .base import BaseState, AppWidgetState, Annotator from .navi_widget import Navi +from .ipytyping.annotations import LabelStore, _label_store_to_image_button from .storage import JsonCaptureStorage # Internal Cell - class CaptureState(BaseState): - annotations: Dict[str, Optional[Dict[str, bool]]] = {} - disp_number: int = 9 + annotations: LabelStore = LabelStore() + grid: Grid question_value: str = '' all_none: bool = False - n_rows: int = 3 - n_cols: int = 3 - -# Cell - -class CaptureGrid(GridBox): - """ - Represents grid of `ImageButtons` with state. - - """ - debug_output = Output(layout={'border': '1px solid black'}) - - def __init__(self, grid_item=ImageButton, image_width=150, image_height=150, - n_rows=3, n_cols=3, display_label=False): - - self.image_width = image_width - self.image_height = image_height - self.n_rows = n_rows - self.n_cols = n_cols - self._screen_im_number = IntText(value=n_rows * n_cols, - description='screen_image_number', - disabled=False) - - self._labels = [grid_item( - display_label=display_label, image_width='%dpx' % self.image_width, - image_height='%dpx' % self.image_height) for _ in range(self._screen_im_number.value)] - - self.callback = None - - gap = 40 if display_label else 15 - - centered_settings = { - 'grid_template_columns': " ".join(["%dpx" % (self.image_width + gap) for i - in range(self.n_cols)]), - 'grid_template_rows': " ".join(["%dpx" % (self.image_height + gap) for i - in range(self.n_rows)]), - 'justify_content': 'center', - 'align_content': 'space-around' - } - - super().__init__(children=self._labels, layout=Layout(**centered_settings)) - - @debug_output.capture(clear_output=True) - def load_annotations_labels(self, annotations: Optional[Iterable[Dict]] = None): - # error: Argument 1 to "iter" has incompatible type - # "Optional[Iterable[Dict[Any, Any]]]"; expected "Iterable[Dict[Any, Any]]" - iter_state = iter(annotations) # type: ignore - - for label in self._labels: - p = next(iter_state, None) - if p: - label.image_path = str(p) # type: ignore - label.label_value = Path(p).stem # type: ignore - label.active = annotations[p].get('answer', False) # type: ignore - else: - label.clear() - - if self.callback: - self.register_on_click() - - def on_click(self, cb: Callable): - self.callback = cb - self.register_on_click() - - @debug_output.capture(clear_output=True) - def register_on_click(self): - for label in self._labels: - label.reset_callbacks() - label.on_click(partial(self.callback, name=label.name)) # Internal Cell @@ -108,6 +39,8 @@ class CaptureAnnotatorGUI(AppLayout): activated when the user navigates through the annotator """ + debug_output = Output(layout={'border': '1px solid black'}) + def __init__( self, app_state: AppWidgetState, @@ -123,12 +56,6 @@ def __init__( self._grid_box_clicked = grid_box_clicked self._select_none_changed = select_none_changed - self._screen_im_number = IntText( - value=self._capture_state.n_rows * self._capture_state.n_cols, - description='screen_image_number', - disabled=False - ) - self._navi = Navi() self._save_btn = Button(description="Save", @@ -152,13 +79,7 @@ def __init__( ) ) - self._grid_box = CaptureGrid( - image_width=self._app_state.size[0], - image_height=self._app_state.size[1], - n_rows=self._capture_state.n_rows, - n_cols=self._capture_state.n_cols, - display_label=False - ) + self._grid_box = GridMenu(capture_state.grid) self._grid_label = HTML() self._labels_box = VBox( @@ -174,9 +95,10 @@ def __init__( ) ) - self._navi.on_navi_clicked = on_navi_clicked + self.on_navi_clicked = on_navi_clicked + self._navi.on_navi_clicked = self._on_navi_clicked self._save_btn.on_click(self._btn_clicked) - self._grid_box.on_click(self._grid_clicked) + self._grid_box.on_click(self.on_grid_clicked) self._none_checkbox.observe(self._none_checkbox_changed, 'value') if self._capture_state.question_value: @@ -186,12 +108,12 @@ def __init__( self._set_navi_max_im_number(self._app_state.max_im_number) if self._capture_state.annotations: - self._grid_box.load_annotations_labels(self._capture_state.annotations) + self._load_menu(self._capture_state.annotations) self._capture_state.subscribe(self._set_none_checkbox, 'all_none') self._capture_state.subscribe(self._set_label, 'question_value') self._app_state.subscribe(self._set_navi_max_im_number, 'max_im_number') - self._capture_state.subscribe(self._grid_box.load_annotations_labels, 'annotations') + self._capture_state.subscribe(self._load_menu, 'annotations') super().__init__( header=None, @@ -202,6 +124,20 @@ def __init__( pane_widths=(2, 8, 0), pane_heights=(1, 4, 1)) + def _on_navi_clicked(self, index: int): + if self.on_navi_clicked: + self.on_navi_clicked(index) + + self._grid_box.load( + _label_store_to_image_button(self._capture_state.annotations) + ) + + @debug_output.capture(clear_output=True) + def _load_menu(self, annotations: LabelStore): + self._grid_box.load( + _label_store_to_image_button(annotations) + ) + def _set_none_checkbox(self, all_none: bool): self._none_checkbox.value = all_none @@ -222,7 +158,7 @@ def _none_checkbox_changed(self, change: dict): if self._select_none_changed: self._select_none_changed(change) - def _grid_clicked(self, event, name=None): + def on_grid_clicked(self, event, name=None): if self._grid_box_clicked: self._grid_box_clicked(event, name) else: @@ -288,9 +224,6 @@ def __init__( self.output_item = output_item self._last_index = 0 - self._capture_state.subscribe(self.update_state, 'disp_number') - self._capture_state.subscribe(self._calc_screens_num, 'disp_number') - self.images = self._storage.get_im_names(filter_files) self.current_im_number = len(self.images) @@ -298,11 +231,12 @@ def __init__( self._capture_state.question_value = ('

{question}

') - self.update_state(self._capture_state.disp_number) - self._calc_screens_num(self._capture_state.disp_number) + self.update_state() + self._calc_screens_num() - def update_state(self, disp_number: int): + def update_state(self): state_images = self._get_state_names(self._app_state.index) + tmp_annotations = deepcopy(self._capture_state.annotations) current_state = {} for im_path in state_images: @@ -313,7 +247,9 @@ def update_state(self, disp_number: int): # error: Incompatible types in assignment (expression has type # "Dict[str, Dict[Any, Any]]", variable has type # "Dict[str, Optional[Dict[str, bool]]]") - self._capture_state.annotations = current_state # type: ignore + tmp_annotations.clear() + tmp_annotations.update(current_state) + self._capture_state.annotations = tmp_annotations # type: ignore def _update_all_none_state(self, state_images: dict): self._capture_state.all_none = all( @@ -321,13 +257,13 @@ def _update_all_none_state(self, state_images: dict): ) def save_annotations(self, index: int): # to disk - state_images = self._capture_state.annotations + state_images = dict(self._capture_state.annotations) self._storage.update_annotations(state_images) def _get_state_names(self, index: int) -> List[str]: - start = index * self._capture_state.disp_number - end = start + self._capture_state.disp_number + start = index * self._capture_state.grid.disp_number + end = start + self._capture_state.grid.disp_number im_names = self.images[start:end] return im_names @@ -337,24 +273,21 @@ def idx_changed(self, index: int): ''' self._app_state.set_quietly('index', index) self.save_annotations(self._last_index) - self.update_state(self._capture_state.disp_number) + self.update_state() self._last_index = index - def _calc_screens_num(self, disp_number: int): + def _calc_screens_num(self): self._app_state.max_im_number = math.ceil( - self.current_im_number / self._capture_state.disp_number + self.current_im_number / self._capture_state.grid.disp_number ) @debug_output.capture(clear_output=False) def handle_grid_click(self, event: dict, name=None): p = self._storage.input_item_path / name - current_state = self._capture_state.annotations.copy() - + current_state = deepcopy(self._capture_state.annotations) if not p.is_dir(): - # error: Item "None" of "Optional[Dict[str, bool]]" - # has no attribute "get" state_answer = self._capture_state.annotations[ - str(p)].get('answer', False) # type: ignore + str(p)].get('answer', False) current_state[str(p)] = {'answer': not state_answer} for k, v in current_state.items(): @@ -364,7 +297,7 @@ def handle_grid_click(self, event: dict, name=None): if self._capture_state.all_none: self._capture_state.all_none = False else: - self._update_all_none_state(current_state) + self._update_all_none_state(dict(current_state)) else: return @@ -372,12 +305,17 @@ def handle_grid_click(self, event: dict, name=None): def select_none(self, change: dict): if self._capture_state.all_none: - self._capture_state.annotations = {p: { - 'answer': False} for p in self._capture_state.annotations} + tmp_annotations = deepcopy(self._capture_state.annotations) + tmp_annotations.clear() + tmp_annotations.update( + {p: { + 'answer': False} for p in self._capture_state.annotations} + ) + self._capture_state.annotations = tmp_annotations # Cell -class CaptureAnnotator: +class CaptureAnnotator(Annotator): debug_output = Output(layout={'border': '1px solid black'}) """ Represents capture annotator. @@ -397,7 +335,7 @@ def __init__( annotation_file_path, n_rows=3, n_cols=3, - disp_number: int = 9, + disp_number=9, question=None, filter_files=None ): @@ -411,21 +349,29 @@ def __init__( self._annotation_file_path = annotation_file_path self._n_rows = n_rows self._n_cols = n_cols - self._disp_number = disp_number self._question = question self._filter_files = filter_files - self.app_state = AppWidgetState( + app_state = AppWidgetState( uuid=str(id(self)), **{'size': (input_item.width, input_item.height)} ) + + super().__init__(app_state) + + grid = Grid( + width=input_item.width, + height=input_item.height, + n_rows=n_rows, + n_cols=n_cols, + display_label=False, + disp_number=disp_number + ) + self.capture_state = CaptureState( uuid=str(id(self)), - **{ - 'n_cols': n_cols, - 'n_rows': n_rows, - 'disp_number': disp_number - } + annotations=LabelStore(), + grid=grid ) self.storage = CaptureAnnotationStorage( diff --git a/ipyannotator/custom_input/buttons.py b/ipyannotator/custom_input/buttons.py index c0933c4..3a73b66 100644 --- a/ipyannotator/custom_input/buttons.py +++ b/ipyannotator/custom_input/buttons.py @@ -1,6 +1,162 @@ # AUTOGENERATED! DO NOT EDIT! File to edit: nbs/16_custom_buttons.ipynb (unless otherwise specified). -__all__ = [] +__all__ = ['ImageButton'] + +# Internal Cell +from pathlib import Path + +import attr +from ipyevents import Event +from ipywidgets import Image, VBox, Layout, Output, HTML +from traitlets import Bool, Unicode, HasTraits, observe +from typing import Optional, Union, Any + +# Internal Cell +@attr.define(slots=False) +class ImageButtonSetting: + im_path: Optional[str] = None + label: Optional[Union[HTML, str]] = None + im_name: Optional[str] = None + im_index: Optional[Any] = None + display_label: bool = True + image_width: str = '50px' + image_height: Optional[str] = None + +# Cell + +class ImageButton(VBox, HasTraits): + """ + Represents simple image with label and toggle button functionality. + + # Class methods + + - clear(): Clear image infos + + - on_click(p_function): Handle click events + + # Class methods + + - clear(): Clear image infos + + - on_click(p_function): Handle click events + + - reset_callbacks(): Reset event callbacks + """ + debug_output = Output(layout={'border': '1px solid black'}) + active = Bool() + image_path = Unicode() + label_value = Unicode() + + def __init__(self, setting: ImageButtonSetting): + + self.setting = setting + self.image = Image( + layout=Layout(display='flex', + justify_content='center', + align_items='center', + align_content='center', + width=setting.image_width, + margin='0 0 0 0', + height=setting.image_height), + ) + + if self.setting.display_label: # both image and label + self.setting.label = HTML( + value='?', + layout=Layout(display='flex', + justify_content='center', + align_items='center', + align_content='center'), + ) + else: # no label (capture image case) + self.im_name = self.setting.im_name + self.im_index = self.setting.im_index + self.image.layout.border = 'solid 1px gray' + self.image.layout.object_fit = 'contain' + self.image.margin = '0 0 0 0' + self.image.layout.overflow = 'hidden' + + super().__init__(layout=Layout(align_items='center', + margin='3px', + overflow='hidden', + padding='2px')) + if not setting.im_path: + self.clear() + + self.d = Event(source=self, watched_events=['click']) + + @observe('image_path') + def _read_image(self, change=None): + new_path = change['new'] + if new_path: + self.image.value = open(new_path, "rb").read() + if not self.children: + self.children = (self.image,) + if self.setting.display_label: + self.children += (self.setting.label,) + else: + #do not display image widget + self.children = tuple() + + @observe('label_value') + def _read_label(self, change=None): + new_label = change['new'] + + if isinstance(self.setting.label, HTML): + self.setting.label.value = new_label + else: + self.setting.label = new_label + + def clear(self): + if isinstance(self.setting.label, HTML): + self.setting.label.value = '' + else: + self.setting.label = '' + self.image_path = '' + self.active = False + + @observe('active') + def mark(self, ev): + # pad to compensate self size with border + if self.active: + if self.setting.display_label: + self.layout.border = 'solid 2px #1B8CF3' + self.layout.padding = '0px' + else: + self.image.layout.border = 'solid 3px #1B8CF3' + self.image.layout.padding = '0px' + else: + if self.setting.display_label: + self.layout.border = 'none' + self.layout.padding = '2px' + else: + self.image.layout.border = 'solid 1px gray' + + def __eq__(self, other): + equals = [ + other.image_path == self.image_path, + other.label_value == self.label_value, + other.active == self.active, + ] + + return all(equals) + + def update(self, other): + if self != other: + self.image_path = other.image_path + self.label_value = other.label_value + self.active = other.active + + @property + def value(self): + return Path(self.image_path).name + + @debug_output.capture(clear_output=False) + def on_click(self, cb): + self.d.on_dom_event(cb) + + def reset_callbacks(self): + self.d.reset_callbacks() # Internal Cell @@ -11,4 +167,11 @@ class ActionButton(Button): def __init__(self, value=None, **kwargs): super().__init__(**kwargs) - self.value = value \ No newline at end of file + self.value = value + + def reset_callbacks(self): + self.on_click(None, remove=True) + + def update(self, other): + self.value = other.value + self.layout = other.layout \ No newline at end of file diff --git a/ipyannotator/custom_input/coordinates.py b/ipyannotator/custom_input/coordinates.py index b9d970a..d1ba12f 100644 --- a/ipyannotator/custom_input/coordinates.py +++ b/ipyannotator/custom_input/coordinates.py @@ -17,9 +17,11 @@ def __init__( uuid: int = None, bbox_coord: BboxCoordinate = None, input_max: BboxCoordinate = None, - coord_changed: Optional[Callable] = None + coord_changed: Optional[Callable] = None, + disabled: bool = False ): super().__init__() + self.disabled = disabled self.uuid = uuid self._input_max = input_max self.coord_changed = coord_changed @@ -44,7 +46,8 @@ def inputs(self) -> list: min=0, max=None if self._input_max is None else getattr(self._input_max, in_p), layout=Layout(width="55px"), - continuous_update=False + continuous_update=False, + disabled=self.disabled ) widget_inputs.append(widget_input) widget_input.observe(self._on_coord_change, names="value") diff --git a/ipyannotator/custom_widgets/__init__.py b/ipyannotator/custom_widgets/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/ipyannotator/custom_widgets/grid_menu.py b/ipyannotator/custom_widgets/grid_menu.py new file mode 100644 index 0000000..7499b20 --- /dev/null +++ b/ipyannotator/custom_widgets/grid_menu.py @@ -0,0 +1,140 @@ +# AUTOGENERATED! DO NOT EDIT! File to edit: nbs/02b_grid_menu.ipynb (unless otherwise specified). + +__all__ = ['GridMenu'] + +# Internal Cell +from math import ceil +from functools import partial +from typing import Callable, Iterable, Optional, Tuple +import warnings +import attr +from ipywidgets import GridBox, Output, Layout + +# Internal Cell +@attr.define(slots=False) +class Grid: + width: int + height: int + n_rows: Optional[int] = 3 + n_cols: Optional[int] = 3 + disp_number: int = 9 + display_label: bool = False + + @property + def num_items(self) -> int: + row, col = self.area_adjusted(self.disp_number) + return row * col + + def area_adjusted(self, n_total: int) -> Tuple[int, int]: + """Returns the row and col automatic arranged""" + if self.n_cols is None: + if self.n_rows is None: # automatic arrange + label_cols = 3 + label_rows = ceil(n_total / label_cols) + else: # calc cols to show all labels + label_rows = self.n_rows + label_cols = ceil(n_total / label_rows) + else: + if self.n_rows is None: # calc rows to show all labels + label_cols = self.n_cols + label_rows = ceil(n_total / label_cols) + else: # user defined + label_cols = self.n_cols + label_rows = self.n_rows + + return label_rows, label_cols + +# Cell +class GridMenu(GridBox): + debug_output = Output(layout={'border': '1px solid black'}) + + def __init__( + self, + grid: Grid, + widgets: Optional[Iterable] = None, + ): + self.callback = None + self.gap = 40 if grid.display_label else 15 + self.grid = grid + + n_row, n_col = grid.area_adjusted(grid.disp_number) + column = grid.width + self.gap + row = grid.height + self.gap + centered_settings = { + 'grid_template_columns': " ".join([f'{(column)}px' for _ + in range(n_col)]), + 'grid_template_rows': " ".join([f'{row}px' for _ + in range(n_row)]), + 'justify_content': 'center', + 'align_content': 'space-around' + } + + super().__init__( + layout=Layout(**centered_settings) + ) + + if widgets: + self.load(widgets) + self.widgets = widgets + + def _fill_widgets(self, widgets: Iterable): + if self.widgets is None: + self.widgets = widgets + + self.children = self.widgets + + if self.callback: + self.register_on_click() + else: + iter_state = iter(widgets) + + for widget in self.widgets: + i_widget = next(iter_state, None) + if i_widget: + widget.update(i_widget) + else: + widget.clear() + + def _filter_widgets(self, widgets: Iterable) -> Iterable: + """Limit the number of widgets to be rendered + according to the grid's area""" + widgets_list = list(widgets) # Iterable don't have len() + num_widgets = len(widgets_list) + row, col = self.grid.area_adjusted(num_widgets) + num_items = row * col + + if num_widgets > num_items: + warnings.warn("!! Not all labels shown. Check n_cols, n_rows args !!") + return widgets_list[:num_items] + + return widgets + + @debug_output.capture(clear_output=False) + def load(self, widgets: Iterable, callback: Optional[Callable] = None): + widgets_filtered = self._filter_widgets(widgets) + self._fill_widgets(widgets_filtered) + + if callback: + self.on_click(callback) + + @debug_output.capture(clear_output=False) + def on_click(self, callback: Callable): + setattr(self, 'callback', callback) + self.register_on_click() + + @debug_output.capture(clear_output=False) + def register_on_click(self): + if self.widgets: + for widget in self.widgets: + widget.reset_callbacks() + + widget.on_click( + partial( + self.callback, + value=widget.value + ) + ) + + def clear(self): + self.widgets = None + self.children = tuple() \ No newline at end of file diff --git a/ipyannotator/datasets/factory_legacy.py b/ipyannotator/datasets/factory_legacy.py index 318b524..94c8272 100644 --- a/ipyannotator/datasets/factory_legacy.py +++ b/ipyannotator/datasets/factory_legacy.py @@ -59,7 +59,7 @@ def get_settings(dataset: DS): project_file = project_path / 'annotations.json' image_dir = 'images' label_dir = None - im_width = 50 + im_width = 200 im_height = im_width create_object_detection(path=project_path, n_samples=50, n_objects=1, size=(500, 500)) diff --git a/ipyannotator/doc_utils.py b/ipyannotator/doc_utils.py new file mode 100644 index 0000000..fc72824 --- /dev/null +++ b/ipyannotator/doc_utils.py @@ -0,0 +1,78 @@ +# AUTOGENERATED! DO NOT EDIT! File to edit: nbs/00d_doc_utils.ipynb (unless otherwise specified). + +__all__ = ['nbglob', 'hide', 'collapse_cells'] + +# Internal Cell +import os + +# Internal Cell +def is_building_docs() -> bool: + return 'DOCUTILSCONFIG' in os.environ + +# Internal Cell +import glob +from fastcore.all import L, compose, Path +from nbdev.export2html import _mk_flag_re, _re_cell_to_collapse_output, check_re +from nbdev.export import check_re_multi +import nbformat as nbf + +# Cell +def nbglob(fname='.', recursive=False, extension='.ipynb') -> L: + """Find all files in a directory matching an extension. + Ignores hidden directories and filenames starting with `_`""" + fname = Path(fname) + if fname.is_dir(): + abs_name = fname.absolute() + rec_path = f'{abs_name}/**/*{extension}' + non_rec_path = f'{abs_name}/*{extension}' + fname = rec_path if recursive else non_rec_path + fls = L( + glob.glob(str(fname), recursive=recursive) + ).filter( + lambda x: '/.' not in x + ).map(Path) + return fls.filter(lambda x: not x.name.startswith('_') and x.name.endswith(extension)) + +# Internal Cell +def upd_metadata(cell, tag): + cell_tags = list(set(cell.get('metadata', {}).get('tags', []))) + if tag not in cell_tags: + cell_tags.append(tag) + cell['metadata']['tags'] = cell_tags + +# Cell +def hide(cell): + """Hide inputs of `cell` that need to be hidden + if check_re_multi(cell, [_re_show_doc, *_re_hide_input]): upd_metadata(cell, 'remove-input') + elif check_re(cell, _re_hide_output): upd_metadata(cell, 'remove-output') + """ + regexes = ['#(.+|)hide', '%%ipytest'] + if check_re_multi(cell, regexes): + upd_metadata(cell, 'remove-cell') + + return cell + + +_re_cell_to_collapse_input = _mk_flag_re( + '(collapse_input|collapse-input)', 0, "Cell with #collapse_input") + + +def collapse_cells(cell): + "Add a collapse button to inputs or outputs of `cell` in either the open or closed position" + if check_re(cell, _re_cell_to_collapse_input): + upd_metadata(cell, 'hide-input') + elif check_re(cell, _re_cell_to_collapse_output): + upd_metadata(cell, 'hide-output') + return cell + +# Internal Cell +if __name__ == '__main__': + + _func = compose(hide, collapse_cells) + files = nbglob('nbs/') + + for file in files: + nb = nbf.read(file, nbf.NO_CONVERT) + for c in nb.cells: + _func(c) + nbf.write(nb, file) \ No newline at end of file diff --git a/ipyannotator/explore_annotator.py b/ipyannotator/explore_annotator.py index 9fa0651..69977c9 100644 --- a/ipyannotator/explore_annotator.py +++ b/ipyannotator/explore_annotator.py @@ -3,20 +3,16 @@ __all__ = ['ExploreAnnotatorState', 'ExploreAnnotator'] # Internal Cell -from typing import Optional - from .im2im_annotator import ImCanvas - -# Internal Cell -from .base import BaseState, AppWidgetState +from .base import BaseState, AppWidgetState, Annotator from .navi_widget import Navi from .storage import MapeableStorage, get_image_list_from_folder -from .mltypes import Input, Output +from .mltypes import InputImage, Output from abc import ABC, abstractmethod from IPython.display import display from pathlib import Path from ipywidgets import AppLayout, HBox, Layout -from typing import Any, List +from typing import Any, List, Optional # Cell @@ -27,7 +23,13 @@ class ExploreAnnotatorState(BaseState): class ExploreAnnotatorGUI(AppLayout): - def __init__(self, app_state: AppWidgetState, explorer_state: ExploreAnnotatorState): + def __init__( + self, + app_state: AppWidgetState, + explorer_state: ExploreAnnotatorState, + fit_canvas: bool = False, + has_border: bool = False + ): self._app_state = app_state self._state = explorer_state @@ -44,7 +46,9 @@ def __init__(self, app_state: AppWidgetState, explorer_state: ExploreAnnotatorSt self._image = ImCanvas( width=self._app_state.size[0], - height=self._app_state.size[1] + height=self._app_state.size[1], + fit_canvas=fit_canvas, + has_border=has_border ) # set the values already instantiated on state @@ -141,32 +145,38 @@ def _save_annotation(self, index: int): # Cell -class ExploreAnnotator: +class ExploreAnnotator(Annotator): def __init__( self, project_path: Path, - input_item: Input, + input_item: InputImage, output_item: Output, + has_border: bool = False, *args, **kwargs ): - self._app_state = AppWidgetState(uuid=str(id(self)), **{ + app_state = AppWidgetState(uuid=str(id(self)), **{ # "Input" has no attribute "width", "height" 'size': (input_item.width, input_item.height) # type: ignore }) + + super().__init__(app_state) + self._state = ExploreAnnotatorState(uuid=str(id(self))) # "Input" has no attribute "dir" self._storage = InMemoryStorage(project_path / input_item.dir) # type: ignore self._controller = ExploreAnnotatorController( - self._app_state, + self.app_state, self._state, self._storage ) self._view = ExploreAnnotatorGUI( - self._app_state, - self._state + self.app_state, + self._state, + fit_canvas=input_item.fit_canvas, + has_border=has_border ) def __repr__(self): diff --git a/ipyannotator/helpers.py b/ipyannotator/helpers.py index 66ba248..d5b32d8 100644 --- a/ipyannotator/helpers.py +++ b/ipyannotator/helpers.py @@ -5,6 +5,7 @@ # Internal Cell import pandas as pd +from typing import Union try: from collections.abc import Iterable @@ -101,7 +102,7 @@ def augment(sig): from .datasets.factory import DS as NDS from .datasets.factory_legacy import DS, _combine_train_test from pathlib import Path -from tqdm import tqdm +from tqdm.notebook import tqdm class Tutorial: @@ -110,7 +111,7 @@ class Tutorial: """ - def __init__(self, dataset: DS, project_path): + def __init__(self, dataset: Union[DS, NDS], project_path): self.dataset = dataset self.project_path = project_path if self.dataset not in [DS.ARTIFICIAL_CLASSIFICATION, DS.ARTIFICIAL_DETECTION, @@ -208,7 +209,7 @@ def fix_incorrect_bboxes(self, improver, creator): improver.capture_state.annotations[k] = {'answer': v_expl != v_cret} improver.view._navi._next_btn.click() - def annotate_video_bboxes(self, annotator): + def annotate_video_bboxes(self, annotator) -> dict: mot_gt = pd.read_csv(self.project_path / 'mot.csv') mot_gt.columns = [ 'frame', @@ -226,7 +227,8 @@ def annotate_video_bboxes(self, annotator): full_path = f'{self.project_path}/images' mot_gt['frame'] = mot_gt['frame'].apply(lambda x: full_path + '/' + x + '.jpg') mot_gt.index = mot_gt['frame'] - mot_gt = mot_gt[mot_gt.columns.drop(['frame', 'conf', 'label', 'vis'])] + mot_gt = mot_gt.drop(columns=['frame', 'conf', 'label', 'vis']) +# mot_gt = mot_gt[mot_gt.columns.drop(['frame', 'conf', 'label', 'vis'])] mot_gt = mot_gt.groupby('frame').apply(lambda x: x.to_json(orient='records')) result = mot_gt.to_json(orient='index') parsed = json.loads(result) @@ -259,6 +261,8 @@ def annotate_video_bboxes(self, annotator): with open(self.project_path / 'create_results/annotations.json', 'w+') as f: json.dump(annotations, f) + return annotations + def _mutate_id(self, bbox: dict, index: int) -> str: id = '2' if bbox['height'] == bbox['width']: diff --git a/ipyannotator/im2im_annotator.py b/ipyannotator/im2im_annotator.py index 9b6bf6f..34e254c 100644 --- a/ipyannotator/im2im_annotator.py +++ b/ipyannotator/im2im_annotator.py @@ -1,100 +1,141 @@ # AUTOGENERATED! DO NOT EDIT! File to edit: nbs/07_im2im_annotator.ipynb (unless otherwise specified). -__all__ = ['ImCanvas', 'Im2ImAnnotator'] +__all__ = ['Im2ImAnnotator'] # Internal Cell - +import io import warnings -from math import ceil from pathlib import Path -from typing import Optional, Dict, List, Callable +from copy import deepcopy +from typing import Optional, Callable, Union, Iterable from ipycanvas import Canvas -from ipywidgets import (AppLayout, VBox, HBox, Button, Layout, HTML, Output) +from ipywidgets import (AppLayout, VBox, HBox, Button, Layout, HTML, Output, Image) -from .base import BaseState, AppWidgetState -from .bbox_canvas import draw_img -from .capture_annotator import CaptureGrid -from .image_button import ImageButton +from .base import BaseState, AppWidgetState, Annotator, AnnotatorStep +from .bbox_canvas import ImageRenderer +from .mltypes import OutputImageLabel, OutputLabel, InputImage +from .ipytyping.annotations import LabelStore, LabelStoreCaster +from .custom_widgets.grid_menu import GridMenu, Grid from .navi_widget import Navi from .storage import JsonLabelStorage from IPython.display import display +from .doc_utils import is_building_docs +from PIL import Image as PILImage # Internal Cell class Im2ImState(BaseState): - annotations: Dict[str, Optional[List[str]]] = {} - disp_number: int = 9 + annotations: LabelStore = LabelStore() question_value: str = '' - n_rows: Optional[int] = 3 - n_cols: Optional[int] = 3 + grid: Grid image_path: Optional[str] im_width: int = 300 im_height: int = 300 - label_width: int = 150 - label_height: int = 150 # Cell +if is_building_docs(): + class ImCanvas(Image): + def __init__( + self, + width: int = 150, + height: int = 150, + has_border: bool = False, + fit_canvas: bool = False + ): + super().__init__(width=width, height=height) + image = PILImage.new('RGB', (100, 100), (255, 255, 255)) + b = io.BytesIO() + image.save(b, format='PNG') + self.value = b.getvalue() + + def _draw_image(self, image_path: str): + self.value = Image.from_file(image_path).value + + def _clear_image(self): + pass + + def observe_client_ready(self, cb=None): + pass +else: + class ImCanvas(HBox): # type: ignore + def __init__( + self, + width: int = 150, + height: int = 150, + has_border: bool = False, + fit_canvas: bool = False + ): + self.has_border = has_border + self.fit_canvas = fit_canvas + self._canvas = Canvas(width=width, height=height) + super().__init__([self._canvas]) + + def _draw_image(self, image_path: str): + img_render_strategy = ImageRenderer( + clear=True, + has_border=self.has_border, + fit_canvas=self.fit_canvas + ) -class ImCanvas(HBox): - def __init__(self, width=150, height=150, has_border=False): - self.has_border = has_border - self._canvas = Canvas(width=width, height=height) - super().__init__([self._canvas]) - - def _draw_image(self, image_path: str): - self._image_scale = draw_img( - self._canvas, - image_path, - clear=True, - has_border=self.has_border - ) + self._image_scale = img_render_strategy.render( + self._canvas, + image_path + ) - def _clear_image(self): - self._canvas.clear() + def _clear_image(self): + self._canvas.clear() - # needed to support voila - # https://ipycanvas.readthedocs.io/en/latest/advanced.html#ipycanvas-in-voila - def observe_client_ready(self, cb=None): - self._canvas.on_client_ready(cb) + # needed to support voila + # https://ipycanvas.readthedocs.io/en/latest/advanced.html#ipycanvas-in-voila + def observe_client_ready(self, cb=None): + self._canvas.on_client_ready(cb) # Internal Cell class Im2ImAnnotatorGUI(AppLayout): + debug_output = Output(layout={'border': '1px solid black'}) + def __init__( self, app_state: AppWidgetState, im2im_state: Im2ImState, + state_to_widget: LabelStoreCaster, label_autosize=False, - save_btn_clicked: Callable = None, - grid_box_clicked: Callable = None, - has_border: bool = False + on_save_btn_clicked: Callable = None, + on_grid_box_clicked: Callable = None, + on_navi_clicked: Callable = None, + has_border: bool = False, + fit_canvas: bool = False ): self._app_state = app_state self._im2im_state = im2im_state - self.save_btn_clicked = save_btn_clicked - self.grid_box_clicked = grid_box_clicked + self._on_save_btn_clicked = on_save_btn_clicked + self._on_navi_clicked = on_navi_clicked + self._on_grid_box_clicked = on_grid_box_clicked + self.state_to_widget = state_to_widget if label_autosize: if self._im2im_state.im_width < 100 or self._im2im_state.im_height < 100: - self._im2im_state.set_quietly('label_width', 10) - self._im2im_state.set_quietly('label_height', 10) + self._im2im_state.grid.width = 10 + self._im2im_state.grid.height = 10 elif self._im2im_state.im_width > 1000 or self._im2im_state.im_height > 1000: - self._im2im_state.set_quietly('label_width', 50) - self._im2im_state.set_quietly('label_height', 10) + self._im2im_state.grid.width = 50 + self._im2im_state.grid.height = 10 else: - label_width = min(self._im2im_state.im_width, self._im2im_state.im_height) / 10 - self._im2im_state.set_quietly('label_width', label_width) - self._im2im_state.set_quietly('label_height', label_width) + label_width = min(self._im2im_state.im_width, self._im2im_state.im_height) // 10 + self._im2im_state.grid.width = label_width + self._im2im_state.grid.height = label_width self._image = ImCanvas( width=self._im2im_state.im_width, height=self._im2im_state.im_height, - has_border=has_border + has_border=has_border, + fit_canvas=fit_canvas ) self._navi = Navi() - + self._navi.on_navi_clicked = self.on_navi_clicked self._save_btn = Button(description="Save", layout=Layout(width='auto')) @@ -108,13 +149,7 @@ def __init__( ) ) - self._grid_box = CaptureGrid( - grid_item=ImageButton, - image_width=self._im2im_state.label_width, - image_height=self._im2im_state.label_height, - n_rows=self._im2im_state.n_rows, - n_cols=self._im2im_state.n_cols - ) + self._grid_box = GridMenu(self._im2im_state.grid) self._grid_label = HTML(value="LABEL",) self._labels_box = VBox( @@ -122,51 +157,74 @@ def __init__( layout=Layout( display='flex', justify_content='center', - flex_wrap='wrap', align_items='center') ) + self._save_btn.on_click(self._on_btn_clicked) + self._grid_box.on_click(self.on_grid_clicked) + if self._app_state.max_im_number: self._set_navi_max_im_number(self._app_state.max_im_number) if self._im2im_state.annotations: - self._grid_box.load_annotations_labels(self._im2im_state.annotations) + self._grid_box.load( + self.state_to_widget(self._im2im_state.annotations) + ) if self._im2im_state.question_value: self._set_label(self._im2im_state.question_value) self._im2im_state.subscribe(self._set_label, 'question_value') self._im2im_state.subscribe(self._image._draw_image, 'image_path') - self._im2im_state.subscribe(self._grid_box.load_annotations_labels, 'annotations') - self._save_btn.on_click(self._btn_clicked) - self._grid_box.on_click(self._grid_clicked) + self._im2im_state.subscribe(self.load_menu, 'annotations') + + layout = Layout( + display='flex', + justify_content='center', + align_items='center' + ) + + im2im_display = HBox([ + VBox([self._image, self._controls_box]), + self._labels_box + ], layout=layout) super().__init__( header=None, - left_sidebar=VBox([self._image, self._controls_box], - layout=Layout(display='flex', justify_content='center', - flex_wrap='wrap', align_items='center')), - center=self._labels_box, + left_sidebar=None, + center=im2im_display, right_sidebar=None, footer=None, pane_widths=(6, 4, 0), pane_heights=(1, 1, 1)) + @debug_output.capture(clear_output=False) + def load_menu(self, annotations: LabelStore): + self._grid_box.load( + self.state_to_widget(annotations) + ) + + @debug_output.capture(clear_output=False) + def on_navi_clicked(self, index: int): + if self._on_navi_clicked: + self._on_navi_clicked(index) + def _set_navi_max_im_number(self, max_im_number: int): self._navi.max_im_num = max_im_number def _set_label(self, question_value: str): self._grid_label.value = question_value - def _btn_clicked(self, *args): - if self.save_btn_clicked: - self.save_btn_clicked(*args) + def _on_btn_clicked(self, *args): + if self._on_save_btn_clicked: + self._on_save_btn_clicked(*args) else: warnings.warn("Save button click didn't triggered any event.") - def _grid_clicked(self, event, name=None): - if self.grid_box_clicked: - self.grid_box_clicked(event, name) + @debug_output.capture(clear_output=False) + def on_grid_clicked(self, event, value=None): + if self._on_grid_box_clicked: + self._on_grid_box_clicked(event, value) else: warnings.warn("Grid box click didn't triggered any event.") @@ -178,9 +236,17 @@ def _label_state_to_storage_format(label_state): return [Path(k).name for k, v in label_state.items() if v['answer']] # Internal Cell -def _storage_format_to_label_state(storage_format, label_names, label_dir): - return {str(Path(label_dir) / label): { - 'answer': label in storage_format} for label in label_names} +def _storage_format_to_label_state( + storage_format, + label_names, + label_dir: str +): + try: + path = Path(label_dir) + return {str(path / label): { + 'answer': label in storage_format} for label in label_names} + except Exception: + return {label: {'answer': label in storage_format} for label in label_names} # Internal Cell @@ -212,39 +278,10 @@ def __init__( # Tracks the app_state.index history self._last_index = 0 - self._im2im_state.n_rows, self._im2im_state.n_cols = self._calc_num_labels( - self.labels_num, - # Argument 2 to "_calc_num_labels" of "Im2ImAnnotatorController" - # has incompatible type "Optional[int]"; expected "int" - self._im2im_state.n_rows, # type: ignore - self._im2im_state.n_cols # type: ignore - ) - if question: self._im2im_state.question_value = (f'

' f'{question}

') - def _calc_num_labels(self, n_total: int, n_rows: int, n_cols: int) -> tuple: - if n_cols is None: - if n_rows is None: # automatic arrange - label_cols = 3 - label_rows = ceil(n_total / label_cols) - else: # calc cols to show all labels - label_rows = n_rows - label_cols = ceil(n_total / label_rows) - else: - if n_rows is None: # calc rows to show all labels - label_cols = n_cols - label_rows = ceil(n_total / label_cols) - else: # user defined - label_cols = n_cols - label_rows = n_rows - - if label_cols * label_rows < n_total: - warnings.warn("!! Not all labels shown. Check n_cols, n_rows args !!") - - return label_rows, label_cols - def _update_im(self): # print('_update_im') index = self._app_state.index @@ -256,14 +293,17 @@ def _update_state(self, change=None): # from annotations if not image_path: return - + tmp_annotations = LabelStore() if image_path in self._storage: current_annotation = self._storage.get(str(image_path)) or {} - self._im2im_state.annotations = _storage_format_to_label_state( - storage_format=current_annotation or [], - label_names=self.labels, - label_dir=self._storage.label_dir + tmp_annotations.update( + _storage_format_to_label_state( + storage_format=current_annotation or [], + label_names=self.labels, + label_dir=self._storage.label_dir + ) ) + self._im2im_state.annotations = tmp_annotations def _update_annotations(self, index: int): # from screen # print('_update_annotations') @@ -293,14 +333,19 @@ def idx_changed(self, index: int): @debug_output.capture(clear_output=False) def handle_grid_click(self, event, name): # print('_handle_grid_click') - label_changed = self._storage.label_dir / name + label_changed = name - if label_changed.is_dir(): - # button without image - invalid - return + # check if the im2im is using the label as path + # otherwise it uses the iterable of labels + if isinstance(self._storage.label_dir, Path): + label_changed = self._storage.label_dir / name - label_changed = str(label_changed) - current_label_state = self._im2im_state.annotations.copy() + if label_changed.is_dir(): + # button without image - invalid + return + + label_changed = str(label_changed) + current_label_state = deepcopy(self._im2im_state.annotations) # inverse state current_label_state[label_changed] = { @@ -319,7 +364,7 @@ def to_dict(self, only_annotated: bool) -> dict: # Cell -class Im2ImAnnotator: +class Im2ImAnnotator(Annotator): """ Represents image-to-image annotator. @@ -332,8 +377,8 @@ class Im2ImAnnotator: def __init__( self, project_path: Path, - input_item, - output_item, + input_item: InputImage, + output_item: Union[OutputImageLabel, OutputLabel], annotation_file_path, n_rows=None, n_cols=None, @@ -344,23 +389,31 @@ def __init__( assert input_item, "WARNING: Provide valid Input" assert output_item, "WARNING: Provide valid Output" - self.app_state = AppWidgetState(uuid=str(id(self))) + self.project_path = project_path + self.input_item = input_item + self.output_item = output_item + app_state = AppWidgetState(uuid=str(id(self))) + + super().__init__(app_state) + + grid = Grid( + width=output_item.width, + height=output_item.height, + n_rows=n_rows, + n_cols=n_cols + ) self.im2im_state = Im2ImState( uuid=str(id(self)), - **{ - "im_height": input_item.height, - "im_width": input_item.width, - "label_width": output_item.width, - "label_height": output_item.height, - "n_rows": n_rows, - "n_cols": n_cols, - } + grid=grid, + annotations=LabelStore(), + im_height=input_item.height, + im_width=input_item.width ) self.storage = JsonLabelStorage( im_dir=project_path / input_item.dir, - label_dir=project_path / output_item.dir, + label_dir=self._get_label_dir(), annotation_file_path=annotation_file_path ) @@ -373,22 +426,47 @@ def __init__( question=question, ) + self.state_to_widget = LabelStoreCaster(output_item) + self.view = Im2ImAnnotatorGUI( app_state=self.app_state, im2im_state=self.im2im_state, + state_to_widget=self.state_to_widget, label_autosize=label_autosize, - has_border=has_border + on_navi_clicked=self.controller.idx_changed, + on_save_btn_clicked=self.controller.save_annotations, + on_grid_box_clicked=self.controller.handle_grid_click, + has_border=has_border, + fit_canvas=input_item.fit_canvas ) - self.view.save_btn_clicked = self.controller.save_annotations - self.view.grid_box_clicked = self.controller.handle_grid_click - - # link current image index from controls to annotator model - self.view._navi.on_navi_clicked = self.controller.idx_changed + self.app_state.subscribe(self._on_annotation_step_change, 'annotation_step') # draw current image and bbox only when client is ready self.view.on_client_ready(self.controller.handle_client_ready) + def _on_annotation_step_change(self, annotation_step: AnnotatorStep): + if annotation_step == AnnotatorStep.EXPLORE: + self.state_to_widget.widgets_disabled = True + self.view._grid_box.clear() + elif self.state_to_widget.widgets_disabled: + self.state_to_widget.widgets_disabled = False + + # forces annotator to have img loaded + self.controller._update_im() + self.controller._update_state() + self.view.load_menu(self.im2im_state.annotations) + + def _get_label_dir(self) -> Union[Iterable[str], Path]: + if isinstance(self.output_item, OutputImageLabel): + return self.project_path / self.output_item.dir + elif isinstance(self.output_item, OutputLabel): + return self.output_item.class_labels + else: + raise ValueError( + "output_item should have type OutputLabel or OutputImageLabel" + ) + def __repr__(self): display(self.view) return "" diff --git a/ipyannotator/image_button.py b/ipyannotator/image_button.py deleted file mode 100644 index 19a0f80..0000000 --- a/ipyannotator/image_button.py +++ /dev/null @@ -1,130 +0,0 @@ -# AUTOGENERATED! DO NOT EDIT! File to edit: nbs/05_image_button.ipynb (unless otherwise specified). - -__all__ = ['ImageButton'] - -# Internal Cell -from pathlib import Path - -from ipyevents import Event -from ipywidgets import Image, VBox, Layout, Output, HTML -from traitlets import Bool, Unicode, HasTraits, observe - -# Cell - -class ImageButton(VBox, HasTraits): - """ - Represents simple image with label and toggle button functionality. - - # Class methods - - - clear(): Clear image infos - - - on_click(p_function): Handle click events - - # Class methods - - - clear(): Clear image infos - - - on_click(p_function): Handle click events - - - reset_callbacks(): Reset event callbacks - """ - debug_output = Output(layout={'border': '1px solid black'}) - active = Bool() - image_path = Unicode() - label_value = Unicode() - - def __init__(self, im_path=None, label=None, - im_name=None, im_index=None, - display_label=True, image_width='50px', image_height=None): - - self.display_label = display_label - self.label = 'None' - self.image = Image( - layout=Layout(display='flex', - justify_content='center', - align_items='center', - align_content='center', - width=image_width, - height=image_height), - ) - - if self.display_label: # both image and label - self.label = HTML( - value='?', - layout=Layout(display='flex', - justify_content='center', - align_items='center', - align_content='center'), - ) - else: # no label (capture image case) - self.im_name = im_name - self.im_index = im_index - self.image.layout.border = 'solid 1px gray' - self.image.layout.object_fit = 'contain' - - super().__init__(layout=Layout(align_items='center', - margin='3px', - padding='2px')) - if not im_path: - self.clear() - - self.d = Event(source=self, watched_events=['click']) - - @observe('image_path') - def _read_image(self, change=None): - new_path = change['new'] - if new_path: - self.image.value = open(new_path, "rb").read() - if not self.children: - self.children = (self.image,) - if self.display_label: - self.children += (self.label,) - else: - #do not display image widget - self.children = [] - - @observe('label_value') - def _read_label(self, change=None): - new_label = change['new'] - - if isinstance(self.label, HTML): - self.label.value = new_label - else: - self.label = new_label - - def clear(self): - if isinstance(self.label, HTML): - self.label.value = '' - else: - self.label = '' - self.image_path = '' - self.active = False - - @observe('active') - def mark(self, ev): - # pad to compensate self size with border - if self.active: - if self.display_label: - self.layout.border = 'solid 2px #1B8CF3' - self.layout.padding = '0px' - else: - self.image.layout.border = 'solid 3px #1B8CF3' - self.image.layout.padding = '0px' - else: - if self.display_label: - self.layout.border = 'none' - self.layout.padding = '2px' - else: - self.image.layout.border = 'solid 1px gray' - - @property - def name(self): - return Path(self.image_path).name - - @debug_output.capture(clear_output=False) - def on_click(self, cb): - self.d.on_dom_event(cb) - - def reset_callbacks(self): - self.d.reset_callbacks() \ No newline at end of file diff --git a/ipyannotator/ipytyping/__init__.py b/ipyannotator/ipytyping/__init__.py new file mode 100644 index 0000000..e69de29 diff --git a/ipyannotator/ipytyping/annotations.py b/ipyannotator/ipytyping/annotations.py new file mode 100644 index 0000000..416af21 --- /dev/null +++ b/ipyannotator/ipytyping/annotations.py @@ -0,0 +1,136 @@ +# AUTOGENERATED! DO NOT EDIT! File to edit: nbs/00c_annotation_types.ipynb (unless otherwise specified). + +__all__ = [] + +# Internal Cell +from pathlib import Path +from collections.abc import MutableMapping +from typing import Dict, Optional, Iterable, Any, Union +from ipywidgets import Layout +from ..mltypes import OutputImageLabel, OutputLabel +from ..custom_input.buttons import ImageButton, ImageButtonSetting, ActionButton + +# Internal Cell +class AnnotationStore(MutableMapping): + def __init__(self, annotations: Optional[Dict] = None): + self._annotations = annotations or {} + + def __getitem__(self, key: str): + return self._annotations[key] + + def __delitem__(self, key: str): + if key in self: + del self._annotations[key] + + def __setitem__(self, key: str, value: Any): + self._annotations[key] = value + + def __iter__(self): + return iter(self._annotations) + + def __len__(self): + return len(self._annotations) + + def __repr__(self): + return "{}({!r})".format(self.__class__.__name__, self._annotations) + +# Internal Cell +class LabelStore(AnnotationStore): + def __getitem__(self, key: str): + assert isinstance(key, str) + return self._annotations[key] + + def __delitem__(self, key: str): + assert isinstance(key, str) + if key in self: + del self._annotations[key] + + def __setitem__(self, key: str, value: Optional[Dict[str, bool]]): + assert isinstance(key, str) + if value: + assert isinstance(value, dict) + self._annotations[key] = value + +# Internal Cell +def _label_store_to_image_button( + annotation: LabelStore, + width: int = 150, + height: int = 150, + disabled: bool = False +) -> Iterable[ImageButton]: + button_setting = ImageButtonSetting( + display_label=False, + image_width=f'{width}px', + image_height=f'{height}px' + ) + + buttons = [] + + for path, value in annotation.items(): + image_button = ImageButton(button_setting) + image_button.image_path = str(path) + image_button.label_value = Path(path).stem + image_button.active = value.get('answer', False) + image_button.disabled = disabled + buttons.append(image_button) + + return buttons + +# Internal Cell +def _label_store_to_button( + annotation: LabelStore, + disabled: bool +) -> Iterable[ActionButton]: + layout = { + 'width': 'auto', + 'height': 'auto' + } + buttons = [] + + for label, value in annotation.items(): + button = ActionButton(layout=Layout(**layout)) + button.description = label + button.value = label + button.tooltip = label + button.disabled = disabled + if value.get('answer', True): + button.layout.border = 'solid 2px #1B8CF3' + buttons.append(button) + + return buttons + +# Internal Cell +class LabelStoreCaster: # pylint: disable=too-few-public-methods + """Factory that casts the correctly widget + accordingly with the input""" + + def __init__( + self, + output: Union[OutputImageLabel, OutputLabel], + width: int = 150, + height: int = 150, + widgets_disabled: bool = False + ): + self.width = width + self.height = height + self.output = output + self.widgets_disabled = widgets_disabled + + def __call__(self, annotation: LabelStore) -> Iterable: + if isinstance(self.output, OutputImageLabel): + return _label_store_to_image_button( + annotation, + self.width, + self.height, + self.widgets_disabled + ) + + if isinstance(self.output, OutputLabel): + return _label_store_to_button( + annotation, + disabled=self.widgets_disabled + ) + + raise ValueError( + f"output should have type OutputImageLabel or OutputLabel. {type(self.output)} given" + ) \ No newline at end of file diff --git a/ipyannotator/mltypes.py b/ipyannotator/mltypes.py index 06f24cc..c53c96d 100644 --- a/ipyannotator/mltypes.py +++ b/ipyannotator/mltypes.py @@ -1,9 +1,10 @@ # AUTOGENERATED! DO NOT EDIT! File to edit: nbs/00b_mltypes.ipynb (unless otherwise specified). __all__ = ['Coordinate', 'BboxCoordinate', 'BboxVideoCoordinate', 'Input', 'Output', 'InputImage', 'OutputImageLabel', - 'OutputImageBbox', 'OutputVideoBbox', 'OutputGridBox', 'NoOutput'] + 'OutputLabel', 'OutputImageBbox', 'OutputVideoBbox', 'OutputGridBox', 'NoOutput'] # Internal Cell +import warnings import random import uuid import attr @@ -62,10 +63,20 @@ class InputImage(Input): Height of the image """ - def __init__(self, image_dir='pics', image_width=100, image_height=100): + def __init__( + self, + image_dir: str = 'pics', + image_width: int = 100, + image_height: int = 100, + fit_canvas: bool = False + ): self.width = image_width self.height = image_height self.dir = image_dir + self.fit_canvas = fit_canvas + + if fit_canvas: + warnings.warn("Image size will be ignored since fit_canvas is activated") # Cell class OutputImageLabel(Output): @@ -84,10 +95,18 @@ def __init__(self, label_dir=None, label_width=50, label_height=50): else: self.dir = label_dir +# Cell +class OutputLabel(Output): + def __init__(self, class_labels: List[str], label_width=50, label_height=50): + self.width = label_width + self.height = label_height + self.class_labels = class_labels + # Cell class OutputImageBbox(Output): def __init__(self, classes: List[str] = None): self.classes = classes or [] + self.drawing_enabled = True # Cell class OutputVideoBbox(OutputImageBbox): diff --git a/ipyannotator/right_menu_widget.py b/ipyannotator/right_menu_widget.py index 580943e..79ebe76 100644 --- a/ipyannotator/right_menu_widget.py +++ b/ipyannotator/right_menu_widget.py @@ -20,28 +20,32 @@ def __init__( bbox_coord: BboxCoordinate, max_coord_input_values: Optional[BboxCoordinate], index: int, - options: List[str] = None + options: List[str] = None, + readonly: bool = False ): super().__init__() + self.readonly = readonly self.bbox_coord = bbox_coord self.index = index self._max_coord_input_values = max_coord_input_values self.layout = Layout(display='flex', overflow='hidden') - self.btn_delete = self._btn_delete(index) self.dropdown_classes = self._dropdown_classes(options) self.btn_select = self._btn_select(index) self.input_coordinates = self._coordinate_inputs(bbox_coord) - self.children = [ - HBox([ - self.btn_select, - self.dropdown_classes, - self.input_coordinates, - self.btn_delete - ]) + elements = [ + self.btn_select, + self.dropdown_classes, + self.input_coordinates, ] + if not self.readonly: + self.btn_delete = self._btn_delete(index) + elements.append(self.btn_delete) + + self.children = [HBox(elements)] + def _btn_delete(self, index: int) -> ActionButton: return ActionButton( layout=Layout(width='auto'), @@ -54,7 +58,8 @@ def _dropdown_classes(self, options: Optional[List[str]], value: str = None) -> return Dropdown( layout=Layout(width='auto'), options=options, - value=value + value=value, + disabled=self.readonly ) def _btn_select(self, index: int) -> ActionButton: @@ -67,7 +72,8 @@ def _btn_select(self, index: int) -> ActionButton: def _coordinate_inputs(self, bbox_coord: BboxCoordinate): return CoordinateInput( bbox_coord=bbox_coord, - input_max=self._max_coord_input_values + input_max=self._max_coord_input_values, + disabled=self.readonly ) # Internal Cell @@ -80,10 +86,11 @@ def __init__( label: List[str], options: List[str], selected: bool = False, - btn_delete_enabled: bool = True + btn_delete_enabled: bool = True, + readonly: bool = False ): super(VBox, self).__init__() # type: ignore - + self.readonly = readonly self.selected = selected self.bbox_video_coord = bbox_video_coord self.object_checkbox = self._object_checkbox() @@ -127,7 +134,8 @@ def __init__( on_coords_changed: Optional[Callable], on_label_changed: Callable, on_btn_delete_clicked: Callable, - on_btn_select_clicked: Optional[Callable] + on_btn_select_clicked: Optional[Callable], + readonly: bool = False ): super().__init__() self._classes = classes @@ -136,6 +144,7 @@ def __init__( self._on_btn_delete_clicked = on_btn_delete_clicked self._on_label_changed = on_label_changed self._on_btn_select_clicked = on_btn_select_clicked + self.readonly = readonly @property def max_coord_input_values(self) -> Optional[BboxCoordinate]: @@ -157,9 +166,11 @@ def render_btn_list(self, bbox_coords: List[BboxCoordinate], classes: List[List[ options=self._classes, bbox_coord=coord, max_coord_input_values=self._max_coord_input_values, + readonly=self.readonly ) - bbox_item.btn_delete.on_click(self.del_element) + if not self.readonly: + bbox_item.btn_delete.on_click(self.del_element) bbox_item.input_coordinates.uuid = index bbox_item.input_coordinates.coord_changed = self._on_coords_changed bbox_item.btn_select.on_click(self._on_btn_select_clicked) @@ -174,6 +185,9 @@ def render_btn_list(self, bbox_coords: List[BboxCoordinate], classes: List[List[ self.children = [*list(self.children), *elements] # type: ignore + def __getitem__(self, index: int): + return self.children[index] + def clear(self): self.children = [] @@ -209,7 +223,7 @@ def __init__( classes: list, on_label_changed: Callable, on_btn_delete_clicked: Callable, - on_btn_select_clicked: Callable, + on_btn_select_clicked: Optional[Callable], on_checkbox_object_clicked: Callable, btn_delete_enabled: bool = True ): @@ -225,9 +239,6 @@ def __init__( self._btn_delete_enabled = btn_delete_enabled self._on_checkbox_object_clicked = on_checkbox_object_clicked - def __getitem__(self, index: int): - return self.children[index] - # error: Signature of "render_btn_list" incompatible with supertype "BBoxList" def render_btn_list( # type: ignore self, diff --git a/ipyannotator/services/bbox_trajectory.py b/ipyannotator/services/bbox_trajectory.py index 0072fc1..3e33553 100644 --- a/ipyannotator/services/bbox_trajectory.py +++ b/ipyannotator/services/bbox_trajectory.py @@ -5,33 +5,24 @@ # Internal Cell from ipycanvas import Canvas from typing import List -from collections.abc import MutableMapping +from ..ipytyping.annotations import AnnotationStore from ..mltypes import BboxCoordinate # Internal Cell -class TrajectoryStore(MutableMapping): - def __init__(self): - self.track = {} - +class TrajectoryStore(AnnotationStore): def __getitem__(self, key: str): assert isinstance(key, str) - return self.track[key] + return self._annotations[key] def __delitem__(self, key: str): assert isinstance(key, str) if key in self: - del self.track[key] + del self._annotations[key] def __setitem__(self, key: str, value: List[BboxCoordinate]): assert isinstance(key, str) assert isinstance(value, list) - self.track[key] = value - - def __iter__(self): - return iter(self.track) - - def __len__(self): - return len(self.track) + self._annotations[key] = value # Internal Cell class BBoxTrajectory: diff --git a/ipyannotator/storage.py b/ipyannotator/storage.py index 23c63c1..f6e38c5 100644 --- a/ipyannotator/storage.py +++ b/ipyannotator/storage.py @@ -3,10 +3,11 @@ __all__ = ['MapeableStorage', 'AnnotationStorage', 'AnnotationStorageIterator', 'AnnotationDBStorage'] # Internal Cell - +import warnings import copy import json import os +from typing import List, Union, Iterable from collections import defaultdict from collections.abc import MutableMapping from pathlib import Path @@ -25,16 +26,15 @@ def construct_annotation_path(project_path=None, file_name=None, results_dir=Non if file_name is not None: annotation_file_path = Path(file_name) results_dir = annotation_file_path.parent - print(f"WARNING: `results_dir` is deduced from `file_name` path: {results_dir}") elif project_path is not None: results_dir = Path( project_path, 'results') if results_dir is None else Path(project_path, results_dir) annotation_file_path = Path(results_dir, 'annotations.json') if annotation_file_path.is_file(): - raise ValueError(f"Error: Annotations file already exists in {results_dir}!" - "\n If you want to create annotations from scratch" - " - use empty dir!") + warnings.warn(f"Error: Annotations file already exists in {results_dir}!" + "\n If you want to create annotations from scratch" + " - use empty dir!") else: raise ValueError("You must define `project_path` or `file_name`!") @@ -78,7 +78,7 @@ def setup_project_paths(project_path, file_name=None, image_dir='pics', import glob -def get_image_list_from_folder(image_dir, strip_path=False): +def get_image_list_from_folder(image_dir) -> Iterable[Path]: ''' Scans to construct list of existing images as objects ''' # if no files in `image_dir` assume all images are under `class_name` directories @@ -88,10 +88,12 @@ def get_image_list_from_folder(image_dir, strip_path=False): path_list = [Path(image_dir, f) for f in os.listdir(image_dir) if os.path.isfile(os.path.join(image_dir, f))] - if strip_path: - path_list = [p.name for p in path_list] return path_list + +def strip_path(paths: Iterable[Path]) -> Iterable[str]: + return [p.name for p in paths] + # Cell class MapeableStorage(MutableMapping): @@ -148,46 +150,45 @@ def load(self, file_name): self.mapping = json.load(data_file) # Internal Cell - from .helpers import flatten, reconstruct_class_images -import warnings class JsonLabelStorage(AnnotationStorage): - def __init__(self, im_dir: Path, label_dir: Path, annotation_file_path): + def __init__(self, im_dir: Path, label_dir: Union[Iterable[str], Path], annotation_file_path): self.annotation_file_path = annotation_file_path self.label_dir = label_dir self.has_annotation_file = True if (annotation_file_path is not None and annotation_file_path.is_file()) else False - print(f'has anno file: {self.has_annotation_file}') self.images = get_image_list_from_folder(im_dir) - # artificialy generate labels if no class images given (TODO: temorary workaround) - if 'class_autogenerated_' in str(label_dir): - print(f'autotgenerated: {label_dir}') - label_dir.mkdir(parents=True, exist_ok=True) + if isinstance(label_dir, Path): + # artificialy generate labels if no class images given (TODO: temorary workaround) + if 'class_autogenerated_' in str(label_dir): + label_dir.mkdir(parents=True, exist_ok=True) - if self.has_annotation_file: - print('reconstruct: FROM annotation file') - reconstruct_class_images(label_dir, annotation_file_path, lbl_w=50, lbl_h=50) - else: - warnings.warn("Annotation file should be provided" - " to generate labels automatically!") + if self.has_annotation_file: + reconstruct_class_images(label_dir, annotation_file_path, lbl_w=50, lbl_h=50) + else: + warnings.warn("Annotation file should be provided" + " to generate labels automatically!") - self.labels = get_image_list_from_folder(label_dir, strip_path=True) + self.labels = strip_path(get_image_list_from_folder(label_dir)) + elif isinstance(label_dir, Iterable): + self.labels = label_dir + else: + raise ValueError("label_dir should have str or Path type") if self.has_annotation_file: # init from json - print('load') self.load() else: # init storage from folder - print('save') super().__init__(self.images) self.save() def get_im_names(self, filter_files=None): - images = sorted(k for k in self.images if str(k) in self.keys()) + keys = self.keys() + images = sorted([k for k in self.images if str(k) in keys]) if not images: raise UserWarning("!! No Images to dipslay !!") @@ -199,14 +200,17 @@ def get_im_names(self, filter_files=None): raise UserWarning("!! No image files to display. Check filter !!") return images - def get_labels(self): + def get_labels(self) -> List[Union[Path, str]]: if not self.labels: warnings.warn("!! No labels to display !!") return [] - if self.has_annotation_file: - return sorted(v for v in self.labels if str(v) in set(flatten(self.values()))) - else: # create mod -> display all labels from folder, not json - return sorted(self.labels) + + if self.has_annotation_file and isinstance(self.label_dir, Path): + values = set(flatten(self.values())) + return sorted([v for v in self.labels if str(v) in values]) + + # create mod -> display all labels from folder, not json + return sorted(self.labels) def save(self): super().save(self.annotation_file_path) diff --git a/make.bat b/make.bat new file mode 100644 index 0000000..153be5e --- /dev/null +++ b/make.bat @@ -0,0 +1,35 @@ +@ECHO OFF + +pushd %~dp0 + +REM Command file for Sphinx documentation + +if "%SPHINXBUILD%" == "" ( + set SPHINXBUILD=sphinx-build +) +set SOURCEDIR=. +set BUILDDIR=_build + +if "%1" == "" goto help + +%SPHINXBUILD% >NUL 2>NUL +if errorlevel 9009 ( + echo. + echo.The 'sphinx-build' command was not found. Make sure you have Sphinx + echo.installed, then set the SPHINXBUILD environment variable to point + echo.to the full path of the 'sphinx-build' executable. Alternatively you + echo.may add the Sphinx directory to PATH. + echo. + echo.If you don't have Sphinx installed, grab it from + echo.https://www.sphinx-doc.org/ + exit /b 1 +) + +%SPHINXBUILD% -M %1 %SOURCEDIR% %BUILDDIR% %SPHINXOPTS% %O% +goto end + +:help +%SPHINXBUILD% -M help %SOURCEDIR% %BUILDDIR% %SPHINXOPTS% %O% + +:end +popd diff --git a/nbs/00_base.ipynb b/nbs/00_base.ipynb index e9ba7d4..1f0784d 100644 --- a/nbs/00_base.ipynb +++ b/nbs/00_base.ipynb @@ -9,6 +9,13 @@ "# default_exp base" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# Base" + ] + }, { "cell_type": "code", "execution_count": null, @@ -27,15 +34,44 @@ "outputs": [], "source": [ "#exporti\n", - "\n", "import json\n", "import random\n", "from pubsub import pub\n", "from pathlib import Path\n", + "from enum import Enum, auto\n", "from typing import NamedTuple, Optional, Tuple, Any, Callable\n", + "from abc import ABC\n", "from pydantic import BaseModel, BaseSettings" ] }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "import ipytest\n", + "import pytest\n", + "ipytest.autoconfig(raise_on_error=True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Ipyannotator base\n", + "\n", + "The current notebook define the classes, enum and helper functions that will be used on the whole application." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## State" + ] + }, { "cell_type": "code", "execution_count": null, @@ -44,13 +80,59 @@ "source": [ "#exporti\n", "\n", - "def validate_project_path(project_path):\n", - " project_path = Path(project_path)\n", - " assert project_path.exists(), \"WARNING: Project path should point to \" \\\n", - " \"existing directory\"\n", - " assert project_path.is_dir(), \"WARNING: Project path should point to \" \\\n", - " \"existing directory\"\n", - " return project_path" + "class StateSettings(BaseSettings):\n", + " class Config:\n", + " validate_assignment = True\n", + "\n", + "\n", + "class BaseState(StateSettings, BaseModel):\n", + " def __init__(self, uuid: str = None, *args, **kwargs):\n", + " super().__init__(*args, **kwargs)\n", + " self.set_quietly('_uuid', uuid)\n", + " self.set_quietly('event_map', {})\n", + "\n", + " def set_quietly(self, key: str, value: Any):\n", + " \"\"\"\n", + " Assigns a value to a state's attribute.\n", + "\n", + " This function can be used to avoid that\n", + " the state dispatches a PyPubSub event.\n", + " It's very usefull to avoid event recursion,\n", + " ex: a component is listening for an event A\n", + " but it also changes the state that dispatch\n", + " the event A. Using set_quietly to set the\n", + " value at the component will avoid the recursion.\n", + " \"\"\"\n", + " object.__setattr__(self, key, value)\n", + "\n", + " @property\n", + " def root_topic(self) -> str:\n", + " if hasattr(self, '_uuid') and self._uuid: # type: ignore\n", + " return f'{self._uuid}.{type(self).__name__}' # type: ignore\n", + "\n", + " return type(self).__name__\n", + "\n", + " def subscribe(self, change: Callable, attribute: str):\n", + " key = f'{self.root_topic}.{attribute}'\n", + " self.event_map[key] = change # type: ignore\n", + " pub.subscribe(change, key)\n", + "\n", + " def unsubscribe(self, attribute: str):\n", + " key = self.topic_attribute(attribute)\n", + " pub.unsubscribe(self.event_map[key], key) # type: ignore\n", + " del self.event_map[key] # type: ignore\n", + "\n", + " def topic_attribute(self, attribute: str):\n", + " return f'{self.root_topic}.{attribute}'\n", + "\n", + " def is_subscribed(self, attribute: str) -> bool:\n", + " return attribute in self.event_map # type: ignore\n", + "\n", + " def __setattr__(self, key: str, value: Any):\n", + " self.set_quietly(key, value)\n", + "\n", + " if key != '__class__':\n", + " pub.sendMessage(f'{self.root_topic}.{key}', **{key: value})" ] }, { @@ -59,8 +141,132 @@ "metadata": {}, "outputs": [], "source": [ - "# hide\n", - "im2im_proj_path = validate_project_path('../data/projects/im2im1/')" + "%%ipytest\n", + "def test_it_can_unsubscribe():\n", + " count = 0\n", + " class Increment(BaseState):\n", + " inc: int = 1\n", + "\n", + " def incrementing(inc):\n", + " nonlocal count\n", + " count += inc\n", + "\n", + " state = Increment()\n", + " state.subscribe(incrementing, 'inc')\n", + " state.inc = 1\n", + " assert count == 1\n", + " state.unsubscribe('inc')\n", + " state.inc = 1\n", + " assert count == 1" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Annotator\n", + "\n", + "All annotator share some states and types, the next cells will design this shared features." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Ipyannotator's uses a `create`, `explore`, `improve` steps when handling data in it's annotators. This enum will be used across the application to check and change the annotators on every step change" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "class AnnotatorStep(Enum):\n", + " EXPLORE = auto()\n", + " CREATE = auto()\n", + " IMPROVE = auto()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "\n", + "class AppWidgetState(BaseState):\n", + " annotation_step: AnnotatorStep = AnnotatorStep.CREATE\n", + " size: Tuple[int, int] = (640, 400)\n", + " max_im_number: int = 1\n", + " index: int = 0" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The following cells will define a common interface for all annotators. Every annotator has a `app_state` that should be implemented." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "class Annotator(ABC):\n", + " def __init__(self, app_state: AppWidgetState):\n", + " self.app_state = app_state" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_raises_not_implemented_app_state():\n", + " with pytest.raises(TypeError):\n", + " class Anno(Annotator):\n", + " pass\n", + "\n", + " anno = Anno()\n", + " anno.app_state" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_not_raises_if_implemented_app_state():\n", + " try:\n", + " class Anno(Annotator):\n", + " def __init__(self):\n", + " self._app_state = AppWidgetState()\n", + " \n", + " @property\n", + " def app_state(self):\n", + " return self._app_state\n", + "\n", + " anno = Anno()\n", + " anno.app_state\n", + " except:\n", + " pytest.fail(\"Anno couldn't call app_state\")" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Helpers" ] }, { @@ -130,46 +336,13 @@ "outputs": [], "source": [ "#exporti\n", - "\n", - "class StateSettings(BaseSettings):\n", - " class Config:\n", - " validate_assignment = True\n", - "\n", - "\n", - "class BaseState(StateSettings, BaseModel):\n", - " def __init__(self, uuid: str = None, *args, **kwargs):\n", - " super().__init__(*args, **kwargs)\n", - " self.set_quietly('_uuid', uuid)\n", - "\n", - " def set_quietly(self, key: str, value: Any):\n", - " \"\"\"\n", - " Assigns a value to a state's attribute.\n", - "\n", - " This function can be used to avoid that\n", - " the state dispatches a PyPubSub event.\n", - " It's very usefull to avoid event recursion,\n", - " ex: a component is listening for an event A\n", - " but it also changes the state that dispatch\n", - " the event A. Using set_quietly to set the\n", - " value at the component will avoid the recursion.\n", - " \"\"\"\n", - " object.__setattr__(self, key, value)\n", - "\n", - " @property\n", - " def root_topic(self) -> str:\n", - " if hasattr(self, '_uuid') and self._uuid: # type: ignore\n", - " return f'{self._uuid}.{type(self).__name__}' # type: ignore\n", - "\n", - " return type(self).__name__\n", - "\n", - " def subscribe(self, change: Callable, attribute: str):\n", - " pub.subscribe(change, f'{self.root_topic}.{attribute}')\n", - "\n", - " def __setattr__(self, key: str, value: Any):\n", - " self.set_quietly(key, value)\n", - "\n", - " if key != '__class__':\n", - " pub.sendMessage(f'{self.root_topic}.{key}', **{key: value})" + "def validate_project_path(project_path):\n", + " project_path = Path(project_path)\n", + " assert project_path.exists(), \"WARNING: Project path should point to \" \\\n", + " \"existing directory\"\n", + " assert project_path.is_dir(), \"WARNING: Project path should point to \" \\\n", + " \"existing directory\"\n", + " return project_path" ] }, { @@ -178,12 +351,8 @@ "metadata": {}, "outputs": [], "source": [ - "#exporti\n", - "\n", - "class AppWidgetState(BaseState):\n", - " size: Tuple[int, int] = (640, 400)\n", - " max_im_number: int = 1\n", - " index: int = 0" + "# hide\n", + "im2im_proj_path = validate_project_path('../data/projects/im2im1/')" ] }, { @@ -207,7 +376,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/00a_annotator.ipynb b/nbs/00a_annotator.ipynb index f4064ef..796340f 100644 --- a/nbs/00a_annotator.ipynb +++ b/nbs/00a_annotator.ipynb @@ -15,11 +15,19 @@ "metadata": {}, "outputs": [], "source": [ - "# hide\n", "%load_ext autoreload\n", "%autoreload 2" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# Annotator Factory\n", + "\n", + "The current notebook will develop the annotator factory. Given an input and output, the factory will return the corresponding annotator. Once called the user can choose between the three actions available: explore, create or improve." + ] + }, { "cell_type": "code", "execution_count": null, @@ -30,14 +38,14 @@ "import json\n", "from abc import ABC, abstractmethod\n", "from pathlib import Path\n", - "from typing import Tuple, Type\n", + "from typing import Tuple, Type, List\n", "\n", "from skimage import io\n", - "from tqdm import tqdm\n", + "from tqdm.notebook import tqdm\n", "\n", - "from ipyannotator.base import generate_subset_anno_json, Settings\n", + "from ipyannotator.base import generate_subset_anno_json, Settings, AnnotatorStep\n", "from ipyannotator.mltypes import (Input, Output, OutputVideoBbox,\n", - " InputImage, OutputImageLabel,\n", + " InputImage, OutputImageLabel, OutputLabel,\n", " OutputImageBbox, OutputGridBox, NoOutput)\n", "from ipyannotator.bbox_annotator import BBoxAnnotator\n", "from ipyannotator.bbox_video_annotator import BBoxVideoAnnotator\n", @@ -48,6 +56,24 @@ "from ipyannotator.storage import (construct_annotation_path, group_files_by_class)" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The next cell defines the actual factory implementation, expecting the pair of input/output. The following classes defines all supported annotators with correct input/output pairs for internal use." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#hide\n", + "import ipytest\n", + "ipytest.autoconfig(raise_on_error=True)" + ] + }, { "cell_type": "code", "execution_count": null, @@ -56,14 +82,19 @@ "source": [ "#exporti\n", "class AnnotatorFactory(ABC):\n", - " io: Tuple[Type[Input], Type[Output]]\n", + " io: Tuple[Type[Input], List[Type[Output]]]\n", "\n", " @abstractmethod\n", " def get_annotator(self):\n", " pass\n", "\n", " def __new__(cls, input_item, output_item):\n", - " subclass_map = {subclass.io: subclass for subclass in cls.__subclasses__()}\n", + " subclass_map = {}\n", + "\n", + " for subclass in cls.__subclasses__():\n", + " for subclass_output in subclass.io[1]:\n", + " subclass_map[(subclass.io[0], subclass_output)] = subclass\n", + "\n", " try:\n", " subclass = subclass_map[(type(input_item), type(output_item))]\n", " instance = super(AnnotatorFactory, subclass).__new__(subclass)\n", @@ -71,51 +102,75 @@ " except KeyError:\n", " print(f\"Pair {(input_item, output_item)} is not supported!\")\n", "\n", - "#\n", - "# Define all supported annotators with correct Input/Output pairs for internal use below:\n", - "#\n", - "\n", "\n", "class Bboxer(AnnotatorFactory):\n", - " io = (InputImage, OutputImageBbox)\n", + " io = (InputImage, [OutputImageBbox])\n", "\n", " def get_annotator(self):\n", " return BBoxAnnotator\n", "\n", "\n", "class Im2Imer(AnnotatorFactory):\n", - " io = (InputImage, OutputImageLabel)\n", + " io = (InputImage, [OutputImageLabel, OutputLabel])\n", "\n", " def get_annotator(self):\n", " return Im2ImAnnotator\n", "\n", "\n", "class Capturer(AnnotatorFactory):\n", - " io = (InputImage, OutputGridBox)\n", + " io = (InputImage, [OutputGridBox])\n", "\n", " def get_annotator(self):\n", " return CaptureAnnotator\n", "\n", "\n", "class ImExplorer(AnnotatorFactory):\n", - " io = (InputImage, NoOutput)\n", + " io = (InputImage, [NoOutput])\n", "\n", " def get_annotator(self):\n", " return ExploreAnnotator\n", "\n", "\n", "class VideoBboxer(AnnotatorFactory):\n", - " io = (InputImage, OutputVideoBbox)\n", + " io = (InputImage, [OutputVideoBbox])\n", "\n", " def get_annotator(self):\n", " return BBoxVideoAnnotator" ] }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_get_im2im_annotator_from_output_image_label():\n", + " inp = InputImage()\n", + " outp = OutputImageLabel()\n", + " factory = AnnotatorFactory(inp, outp).get_annotator()\n", + " assert factory == Im2ImAnnotator" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_get_im2im_annotator_from_output_label():\n", + " inp = InputImage()\n", + " outp = OutputLabel(class_labels=('A', 'B'))\n", + " factory = AnnotatorFactory(inp, outp).get_annotator()\n", + " assert factory == Im2ImAnnotator" + ] + }, { "cell_type": "markdown", "metadata": {}, "source": [ - "# Annotator" + "The following cell uses the factory designed before and define the actions that can be used with the factory." ] }, { @@ -150,13 +205,18 @@ "\n", " annotator = AnnotatorFactory(self.input_item, self.output_item).get_annotator()\n", "\n", - " return annotator(project_path=self.settings.project_path,\n", - " input_item=self.input_item,\n", - " output_item=self.output_item,\n", - " annotation_file_path=anno_,\n", - " n_cols=self.settings.n_cols,\n", - " question=\"Classification \",\n", - " has_border=True)\n", + " self.output_item.drawing_enabled = False\n", + " annotator = annotator(project_path=self.settings.project_path,\n", + " input_item=self.input_item,\n", + " output_item=self.output_item,\n", + " annotation_file_path=anno_,\n", + " n_cols=self.settings.n_cols,\n", + " question=\"Classification \",\n", + " has_border=True)\n", + "\n", + " annotator.app_state.annotation_step = AnnotatorStep.EXPLORE\n", + "\n", + " return annotator\n", "\n", " def create(self):\n", " anno_ = construct_annotation_path(project_path=self.settings.project_path,\n", @@ -165,13 +225,18 @@ "\n", " annotator = AnnotatorFactory(self.input_item, self.output_item).get_annotator()\n", "\n", - " return annotator(project_path=self.settings.project_path,\n", - " input_item=self.input_item,\n", - " output_item=self.output_item,\n", - " annotation_file_path=anno_,\n", - " n_cols=self.settings.n_cols,\n", - " question=\"Classification \",\n", - " has_border=True)\n", + " self.output_item.drawing_enabled = True\n", + " annotator = annotator(project_path=self.settings.project_path,\n", + " input_item=self.input_item,\n", + " output_item=self.output_item,\n", + " annotation_file_path=anno_,\n", + " n_cols=self.settings.n_cols,\n", + " question=\"Classification \",\n", + " has_border=True)\n", + "\n", + " annotator.app_state.annotation_step = AnnotatorStep.CREATE\n", + "\n", + " return annotator\n", "\n", " def improve(self):\n", " # open labels from create step\n", @@ -185,7 +250,6 @@ " if type(self.output_item) == OutputImageLabel:\n", "\n", " #Construct multiple Capturers for each class\n", - " #\n", " out = []\n", " for class_name, class_anno in tqdm(\n", " group_files_by_class(loaded_image_annotations).items()):\n", @@ -226,7 +290,6 @@ " captured_path = Path(self.settings.project_path) / \"captured\"\n", "\n", " # Save annotated images on disk\n", - " #\n", " for im, bbx in tqdm(di.items()):\n", " # use captured_path instead image_dir, keeping the folder structure\n", " old_im_path = Path(im)\n", @@ -257,6 +320,13 @@ "\n", " else:\n", " raise Exception(f\"Improve is not supported for {self.output_item}\")\n", + " if isinstance(out, list):\n", + " def update_step(anno):\n", + " anno.app_state.annotation_step = AnnotatorStep.IMPROVE\n", + " return anno\n", + " out = [update_step(anno) for anno in out]\n", + " else:\n", + " out.app_state.annotation_step = AnnotatorStep.IMPROVE\n", "\n", " return out" ] @@ -282,7 +352,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/00b_mltypes.ipynb b/nbs/00b_mltypes.ipynb index 1babf26..6260a88 100644 --- a/nbs/00b_mltypes.ipynb +++ b/nbs/00b_mltypes.ipynb @@ -9,6 +9,13 @@ "# default_exp mltypes" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# Mltypes" + ] + }, { "cell_type": "code", "execution_count": null, @@ -27,12 +34,25 @@ "outputs": [], "source": [ "#exporti\n", + "import warnings\n", "import random\n", "import uuid\n", "import attr\n", "from typing import List" ] }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "import ipytest\n", + "import pytest\n", + "ipytest.autoconfig(raise_on_error=True)" + ] + }, { "cell_type": "markdown", "metadata": {}, @@ -138,10 +158,74 @@ " Height of the image\n", " \"\"\"\n", "\n", - " def __init__(self, image_dir='pics', image_width=100, image_height=100):\n", + " def __init__(\n", + " self,\n", + " image_dir: str = 'pics',\n", + " image_width: int = 100,\n", + " image_height: int = 100,\n", + " fit_canvas: bool = False\n", + " ):\n", " self.width = image_width\n", " self.height = image_height\n", - " self.dir = image_dir" + " self.dir = image_dir\n", + " self.fit_canvas = fit_canvas\n", + "\n", + " if fit_canvas:\n", + " warnings.warn(\"Image size will be ignored since fit_canvas is activated\")" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_warn_if_fit_canvas_is_activate_with_size():\n", + " with warnings.catch_warnings(record=True) as w:\n", + " inp_img = InputImage(image_width = 300, image_height = 300, fit_canvas=True)\n", + "\n", + " assert bool(w) is True" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_doesnt_warn_if_fit_canvas_is_deactivate_with_size():\n", + " with warnings.catch_warnings(record=True) as w:\n", + " inp_img = InputImage(image_width = 300, image_height = 300, fit_canvas=False)\n", + "\n", + " assert bool(w) is False" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_warn_if_fit_canvas_is_activate_with_size_none():\n", + " with warnings.catch_warnings(record=True) as w:\n", + " inp_img = InputImage(image_width=None, image_height=None, fit_canvas=True)\n", + " assert bool(w) is True" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_doesnt_warn_if_fit_canvas_is_deactivate_with_size_none():\n", + " with warnings.catch_warnings(record=True) as w:\n", + " inp_img = InputImage(image_width=None, image_height=None, fit_canvas=False)\n", + " assert bool(w) is False" ] }, { @@ -179,6 +263,20 @@ " self.dir = label_dir" ] }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#export\n", + "class OutputLabel(Output):\n", + " def __init__(self, class_labels: List[str], label_width=50, label_height=50):\n", + " self.width = label_width\n", + " self.height = label_height\n", + " self.class_labels = class_labels" + ] + }, { "cell_type": "code", "execution_count": null, @@ -214,7 +312,8 @@ "#export\n", "class OutputImageBbox(Output):\n", " def __init__(self, classes: List[str] = None):\n", - " self.classes = classes or []" + " self.classes = classes or []\n", + " self.drawing_enabled = True" ] }, { diff --git a/nbs/00c_annotation_types.ipynb b/nbs/00c_annotation_types.ipynb new file mode 100644 index 0000000..43b26e7 --- /dev/null +++ b/nbs/00c_annotation_types.ipynb @@ -0,0 +1,371 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": null, + "id": "dc763abf", + "metadata": {}, + "outputs": [], + "source": [ + "# default_exp ipytyping.annotations" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "06fa07ff", + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "%load_ext autoreload\n", + "%autoreload 2" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "3f14dc56", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "from pathlib import Path\n", + "from collections.abc import MutableMapping\n", + "from typing import Dict, Optional, Iterable, Any, Union\n", + "from ipywidgets import Layout\n", + "from ipyannotator.mltypes import OutputImageLabel, OutputLabel\n", + "from ipyannotator.custom_input.buttons import ImageButton, ImageButtonSetting, ActionButton" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "be5e23da", + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "import ipytest\n", + "import pytest\n", + "ipytest.autoconfig(raise_on_error=True)" + ] + }, + { + "cell_type": "markdown", + "id": "491d358f", + "metadata": {}, + "source": [ + "## Annotation Types\n", + "\n", + "The current notebook store the annotation data typing. Every annotator stores its data in a particular way, this notebook designs the store and it's casting types." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "196e5fde", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "class AnnotationStore(MutableMapping):\n", + " def __init__(self, annotations: Optional[Dict] = None):\n", + " self._annotations = annotations or {}\n", + "\n", + " def __getitem__(self, key: str):\n", + " return self._annotations[key]\n", + "\n", + " def __delitem__(self, key: str):\n", + " if key in self:\n", + " del self._annotations[key]\n", + "\n", + " def __setitem__(self, key: str, value: Any):\n", + " self._annotations[key] = value\n", + "\n", + " def __iter__(self):\n", + " return iter(self._annotations)\n", + "\n", + " def __len__(self):\n", + " return len(self._annotations)\n", + "\n", + " def __repr__(self):\n", + " return \"{}({!r})\".format(self.__class__.__name__, self._annotations)" + ] + }, + { + "cell_type": "markdown", + "id": "698063e8", + "metadata": {}, + "source": [ + "### LabelStore Data Type\n", + "\n", + "The `LabelStore` stores a path as a key it's answer in the format: `{'': {'answer': }`." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "16418a55", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "class LabelStore(AnnotationStore):\n", + " def __getitem__(self, key: str):\n", + " assert isinstance(key, str)\n", + " return self._annotations[key]\n", + "\n", + " def __delitem__(self, key: str):\n", + " assert isinstance(key, str)\n", + " if key in self:\n", + " del self._annotations[key]\n", + "\n", + " def __setitem__(self, key: str, value: Optional[Dict[str, bool]]):\n", + " assert isinstance(key, str)\n", + " if value:\n", + " assert isinstance(value, dict)\n", + " self._annotations[key] = value" + ] + }, + { + "cell_type": "markdown", + "id": "7971dff5", + "metadata": {}, + "source": [ + "The following cell will define a cast from the annotation to a custom widget called `ImageButton`." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "ff3545e4", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "def _label_store_to_image_button(\n", + " annotation: LabelStore,\n", + " width: int = 150,\n", + " height: int = 150,\n", + " disabled: bool = False\n", + ") -> Iterable[ImageButton]:\n", + " button_setting = ImageButtonSetting(\n", + " display_label=False,\n", + " image_width=f'{width}px',\n", + " image_height=f'{height}px'\n", + " )\n", + "\n", + " buttons = []\n", + "\n", + " for path, value in annotation.items():\n", + " image_button = ImageButton(button_setting)\n", + " image_button.image_path = str(path)\n", + " image_button.label_value = Path(path).stem\n", + " image_button.active = value.get('answer', False)\n", + " image_button.disabled = disabled\n", + " buttons.append(image_button)\n", + "\n", + " return buttons" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "351f50fb", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "def _label_store_to_button(\n", + " annotation: LabelStore,\n", + " disabled: bool\n", + ") -> Iterable[ActionButton]:\n", + " layout = {\n", + " 'width': 'auto',\n", + " 'height': 'auto'\n", + " }\n", + " buttons = []\n", + "\n", + " for label, value in annotation.items():\n", + " button = ActionButton(layout=Layout(**layout))\n", + " button.description = label\n", + " button.value = label\n", + " button.tooltip = label\n", + " button.disabled = disabled\n", + " if value.get('answer', True):\n", + " button.layout.border = 'solid 2px #1B8CF3'\n", + " buttons.append(button)\n", + "\n", + " return buttons" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "b67c7dcf", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "class LabelStoreCaster: # pylint: disable=too-few-public-methods\n", + " \"\"\"Factory that casts the correctly widget\n", + " accordingly with the input\"\"\"\n", + "\n", + " def __init__(\n", + " self,\n", + " output: Union[OutputImageLabel, OutputLabel],\n", + " width: int = 150,\n", + " height: int = 150,\n", + " widgets_disabled: bool = False\n", + " ):\n", + " self.width = width\n", + " self.height = height\n", + " self.output = output\n", + " self.widgets_disabled = widgets_disabled\n", + "\n", + " def __call__(self, annotation: LabelStore) -> Iterable:\n", + " if isinstance(self.output, OutputImageLabel):\n", + " return _label_store_to_image_button(\n", + " annotation,\n", + " self.width,\n", + " self.height,\n", + " self.widgets_disabled\n", + " )\n", + "\n", + " if isinstance(self.output, OutputLabel):\n", + " return _label_store_to_button(\n", + " annotation,\n", + " disabled=self.widgets_disabled\n", + " )\n", + "\n", + " raise ValueError(\n", + " f\"output should have type OutputImageLabel or OutputLabel. {type(self.output)} given\"\n", + " )" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "55895901", + "metadata": {}, + "outputs": [], + "source": [ + "@pytest.fixture\n", + "def str_label_fixture():\n", + " return {\n", + " 'A': {'answer': False},\n", + " 'B': {'answer': True}\n", + " }" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "b848452c", + "metadata": {}, + "outputs": [], + "source": [ + "@pytest.fixture\n", + "def img_label_fixture():\n", + " return {\n", + " '../data/projects/capture1/pics/pink25x25.png': {'answer': False},\n", + " }" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "6dbe324d", + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_cast_label_store_to_image_button(img_label_fixture):\n", + " label_store = LabelStore()\n", + " label_store.update(img_label_fixture)\n", + " \n", + " output = OutputImageLabel()\n", + " caster = LabelStoreCaster(output)\n", + " image_buttons = caster(label_store)\n", + "\n", + " for image_button in image_buttons:\n", + " assert isinstance(image_button, ImageButton)\n", + " assert len(image_buttons) == 1" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "24c1d615", + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_cast_label_store_to_button(str_label_fixture): \n", + " label_store = LabelStore()\n", + " label_store.update(str_label_fixture)\n", + " \n", + " output = OutputLabel(class_labels=list(str_label_fixture.keys()))\n", + " caster = LabelStoreCaster(output)\n", + " buttons = caster(label_store)\n", + "\n", + " assert len(buttons) == 2\n", + " for button in buttons:\n", + " assert isinstance(button, ActionButton)\n", + " assert buttons[0].description == 'A'\n", + " assert buttons[1].description == 'B'\n", + " assert buttons[0].value == 'A'\n", + " assert buttons[1].value == 'B'" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "5be283a2", + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_can_disable_widgets(str_label_fixture):\n", + " label_store = LabelStore()\n", + " label_store.update(str_label_fixture)\n", + " \n", + " output = OutputLabel(class_labels=list(str_label_fixture.keys()))\n", + " caster = LabelStoreCaster(output, widgets_disabled=True)\n", + " buttons = caster(label_store)\n", + " for button in buttons:\n", + " assert button.disabled is True" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "4ab48220", + "metadata": {}, + "outputs": [], + "source": [ + "#hide\n", + "from nbdev.export import notebook2script\n", + "notebook2script()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "8fc4bbc5", + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3", + "language": "python", + "name": "python3" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +} diff --git a/nbs/00d_doc_utils.ipynb b/nbs/00d_doc_utils.ipynb new file mode 100644 index 0000000..6403628 --- /dev/null +++ b/nbs/00d_doc_utils.ipynb @@ -0,0 +1,217 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": null, + "id": "e15c1e2d", + "metadata": {}, + "outputs": [], + "source": [ + "# default_exp doc_utils" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "73b3212f", + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "%load_ext autoreload\n", + "%autoreload 2" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "2cd1d0c4", + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "from nbdev import *" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "ffcac7ef", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "import os" + ] + }, + { + "cell_type": "markdown", + "id": "449065ec", + "metadata": {}, + "source": [ + "# Doc Utils\n", + "\n", + "This notebook develops helper modules to build Ipyannotator's static documentation." + ] + }, + { + "cell_type": "markdown", + "id": "77fc616b", + "metadata": {}, + "source": [ + "The next cell design a helper function that check if the documentation it's been built. This is specially helpful to mock some behaviors that doesn't work well on static docs." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "f340d2fb", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "def is_building_docs() -> bool:\n", + " return 'DOCUTILSCONFIG' in os.environ" + ] + }, + { + "cell_type": "markdown", + "id": "b9608c7c", + "metadata": {}, + "source": [ + "## Docs metadata\n", + "\n", + "The following cells was extracted from [jb-nbdev](https://github.com/fastai/jb-nbdev) and will perform some changes on our metadata to integrate nbdev and mynb-st." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "08024850", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "import glob\n", + "from fastcore.all import L, compose, Path\n", + "from nbdev.export2html import _mk_flag_re, _re_cell_to_collapse_output, check_re\n", + "from nbdev.export import check_re_multi\n", + "import nbformat as nbf" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "a9e6bca1", + "metadata": {}, + "outputs": [], + "source": [ + "#export\n", + "def nbglob(fname='.', recursive=False, extension='.ipynb') -> L:\n", + " \"\"\"Find all files in a directory matching an extension.\n", + " Ignores hidden directories and filenames starting with `_`\"\"\"\n", + " fname = Path(fname)\n", + " if fname.is_dir():\n", + " abs_name = fname.absolute()\n", + " rec_path = f'{abs_name}/**/*{extension}'\n", + " non_rec_path = f'{abs_name}/*{extension}'\n", + " fname = rec_path if recursive else non_rec_path\n", + " fls = L(\n", + " glob.glob(str(fname), recursive=recursive)\n", + " ).filter(\n", + " lambda x: '/.' not in x\n", + " ).map(Path)\n", + " return fls.filter(lambda x: not x.name.startswith('_') and x.name.endswith(extension))" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "de3f485d", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "def upd_metadata(cell, tag):\n", + " cell_tags = list(set(cell.get('metadata', {}).get('tags', [])))\n", + " if tag not in cell_tags:\n", + " cell_tags.append(tag)\n", + " cell['metadata']['tags'] = cell_tags" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "06737dbe", + "metadata": {}, + "outputs": [], + "source": [ + "#export\n", + "def hide(cell):\n", + " \"\"\"Hide inputs of `cell` that need to be hidden\n", + " if check_re_multi(cell, [_re_show_doc, *_re_hide_input]): upd_metadata(cell, 'remove-input')\n", + " elif check_re(cell, _re_hide_output): upd_metadata(cell, 'remove-output')\n", + " \"\"\"\n", + " regexes = ['#(.+|)hide', '%%ipytest']\n", + " if check_re_multi(cell, regexes):\n", + " upd_metadata(cell, 'remove-cell')\n", + "\n", + " return cell\n", + "\n", + "\n", + "_re_cell_to_collapse_input = _mk_flag_re(\n", + " '(collapse_input|collapse-input)', 0, \"Cell with #collapse_input\")\n", + "\n", + "\n", + "def collapse_cells(cell):\n", + " \"Add a collapse button to inputs or outputs of `cell` in either the open or closed position\"\n", + " if check_re(cell, _re_cell_to_collapse_input):\n", + " upd_metadata(cell, 'hide-input')\n", + " elif check_re(cell, _re_cell_to_collapse_output):\n", + " upd_metadata(cell, 'hide-output')\n", + " return cell" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "b1844edf", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "if __name__ == '__main__':\n", + "\n", + " _func = compose(hide, collapse_cells)\n", + " files = nbglob('nbs/')\n", + "\n", + " for file in files:\n", + " nb = nbf.read(file, nbf.NO_CONVERT)\n", + " for c in nb.cells:\n", + " _func(c)\n", + " nbf.write(nb, file)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "cf24e1bf", + "metadata": {}, + "outputs": [], + "source": [ + "#hide\n", + "from nbdev.export import notebook2script\n", + "notebook2script()" + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3", + "language": "python", + "name": "python3" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +} diff --git a/nbs/01_bbox_canvas.ipynb b/nbs/01_bbox_canvas.ipynb index b56d874..3fccde5 100644 --- a/nbs/01_bbox_canvas.ipynb +++ b/nbs/01_bbox_canvas.ipynb @@ -37,6 +37,8 @@ "outputs": [], "source": [ "# exporti\n", + "import io\n", + "import attr\n", "from math import log\n", "from pubsub import pub\n", "from attr import asdict\n", @@ -45,11 +47,95 @@ "from enum import IntEnum\n", "from typing import Dict, Optional, List, Any, Tuple\n", "\n", + "from abc import ABC, abstractmethod\n", "from pydantic import root_validator\n", "from ipyannotator.base import BaseState\n", + "from ipyannotator.doc_utils import is_building_docs\n", "from ipyannotator.mltypes import BboxCoordinate, BboxVideoCoordinate\n", - "from ipycanvas import MultiCanvas, Canvas, hold_canvas\n", - "from ipywidgets import Image, Label, Layout, HBox, VBox, Output" + "from ipycanvas import MultiCanvas as IMultiCanvas, Canvas, hold_canvas\n", + "from ipywidgets import Image, Label, Layout, HBox, VBox, Output\n", + "from PIL import Image as PILImage" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# Bounding Box Canvas\n", + "\n", + "This notebook develops a drawnable canvas where users can draw bounding boxes." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The next cell will override the ipywidget's MultiCanvas when the docs is built, otherwise the original MultiCanvas will be used.\n", + "\n", + "This is a plug in replacement for ipycanvas that can be used to build docs. The difference is that the replacement holds the image state so that it can be included into the docs. Even through the replacement is feature equivalent it lacks mouse events and it's much slower, which consequently doesn't allow its use in lived annotation settings." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "if not is_building_docs():\n", + " class MultiCanvas(IMultiCanvas):\n", + " pass\n", + "else:\n", + " class MultiCanvas(Image): # type: ignore\n", + " def __init__(self, *args, **kwargs):\n", + " super().__init__(**kwargs)\n", + " image = PILImage.new('RGB', (100, 100), (255, 255, 255))\n", + " b = io.BytesIO()\n", + " image.save(b, format='PNG')\n", + " self.value = b.getvalue()\n", + "\n", + " def __getitem__(self, key):\n", + " return self\n", + "\n", + " def draw_image(self, image, x=0, y=0, width=None, height=None):\n", + " self.value = image.value\n", + " self.width = width\n", + " self.height = height\n", + "\n", + " def __getattr__(self, name):\n", + " ignored = [\n", + " 'flush',\n", + " 'fill_rect',\n", + " 'stroke_rect',\n", + " 'stroke_rects',\n", + " 'on_mouse_move',\n", + " 'on_mouse_down',\n", + " 'on_mouse_up',\n", + " 'clear',\n", + " 'on_client_ready',\n", + " 'stroke_styled_line_segments'\n", + " ]\n", + "\n", + " if name in ignored:\n", + " def wrapper(*args, **kwargs):\n", + " return self._ignored(*args, **kwargs)\n", + " return wrapper\n", + " return object.__getattr__(self, name)\n", + "\n", + " @property\n", + " def caching(self):\n", + " return False\n", + "\n", + " @caching.setter\n", + " def caching(self, value):\n", + " pass\n", + "\n", + " @property\n", + " def size(self):\n", + " return (self.width, self.height)\n", + "\n", + " def _ignored(self, *args, **kwargs):\n", + " pass" ] }, { @@ -66,11 +152,12 @@ "outputs": [], "source": [ "# hide\n", - "\n", + "sprite2 = Image.from_file(\"../data/projects/bbox/pics/red400x640.png\")\n", "# Create a multi-layer canvas with 4 layers\n", "multi_canvas = MultiCanvas(4, width=200, height=200)\n", "multi_canvas[0] # Access first layer (background)\n", "multi_canvas[3] # Access last layer\n", + "multi_canvas[3].draw_image(sprite2, 100, 100, width=40, height=40)\n", "multi_canvas" ] }, @@ -102,7 +189,6 @@ "outputs": [], "source": [ "#exporti\n", - "\n", "def draw_bg(canvas, color='rgb(236,240,241)'):\n", " with hold_canvas(canvas):\n", " canvas.fill_style = color\n", @@ -116,7 +202,6 @@ "outputs": [], "source": [ "# hide\n", - "\n", "canvas = Canvas(width=300, height=20)\n", "draw_bg(canvas)\n", "canvas" @@ -143,7 +228,6 @@ "outputs": [], "source": [ "#exporti\n", - "\n", "def draw_bounding_box(canvas, coord: BboxCoordinate, color='white', line_width=1,\n", " border_ratio=2, clear=False, stroke_color='black'):\n", " with hold_canvas(canvas):\n", @@ -211,7 +295,6 @@ "outputs": [], "source": [ "#exporti\n", - "\n", "class BoundingBox:\n", " def __init__(self):\n", " self.color = 'white'\n", @@ -312,20 +395,6 @@ "canvas" ] }, - { - "attachments": { - "canvas.png": { - "image/png": "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" - } - }, - "cell_type": "markdown", - "metadata": {}, - "source": [ - "This code example will produce the following image:\n", - "\n", - "![./images/canvas.png](attachment:canvas.png)" - ] - }, { "cell_type": "code", "execution_count": null, @@ -374,9 +443,11 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "### Draw resized Image\n", + "### Draw Image\n", "\n", - "`ipyannotar` will rezise the image so it fits inside the Canvas. Over the process, the dimensions ratio will mantain constans." + "This section will develop how a image is drawn over a canvas.\n", + "\n", + "The next cell will define how the data behave on top of a canvas. This representation uses the `Image` abstraction from `ipywidgets` in the `image_widget` property, this will allow us to draw on top of the canvas. " ] }, { @@ -385,48 +456,194 @@ "metadata": {}, "outputs": [], "source": [ - "#export\n", - "\n", - "def draw_img(canvas, file, clear=False, has_border=False) -> Tuple[int, int, float]:\n", - " \"\"\"\n", - " draws resized image on canvas and returns scale used\n", - " \"\"\"\n", - " with hold_canvas(canvas):\n", - " if clear:\n", - " canvas.clear()\n", - "\n", - " sprite1 = Image.from_file(file)\n", - "\n", - " width_canvas, height_canvas = canvas.width, canvas.height\n", - " width_img, height_img = get_image_size(file)\n", + "#exporti\n", + "@attr.define\n", + "class ImageCanvas:\n", + " image_widget: Image\n", + " x: int\n", + " y: int\n", + " width: int\n", + " height: int\n", + " scale: float" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "A simple strategy pattern will be develop in the next cells to switch the algorithm used to calculate the image." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "class ImageCanvasPrototype(ABC):\n", + " @abstractmethod\n", + " def prepare_canvas(self, canvas: Canvas, file: str) -> ImageCanvas:\n", + " pass" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "One of the strategies used by `ipyannotator` will resize the image so it fits inside the Canvas. Over the process, the dimensions ratio will mantain constants.\n", "\n", + "The next cells will develop:\n", + "- A mixin to calculate the image scale based on the canvas and images size\n", + "- A concrete strategy application that resizes the image on the canvas" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "class CanvasScaleMixin:\n", + " def _calc_scale(\n", + " self,\n", + " width_canvas: int,\n", + " height_canvas: int,\n", + " width_img: float,\n", + " height_img: float\n", + " ) -> float:\n", " ratio_canvas = float(width_canvas) / height_canvas\n", " ratio_img = float(width_img) / height_img\n", "\n", " if ratio_img > ratio_canvas:\n", " # wider then canvas, scale to canvas width\n", - " scale = width_canvas / width_img\n", - " else:\n", - " # taller then canvas, scale to canvas hight\n", - " scale = height_canvas / height_img\n", + " return width_canvas / width_img\n", + "\n", + " # taller then canvas, scale to canvas height\n", + " return height_canvas / height_img" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "class ScaledImage(ImageCanvasPrototype, CanvasScaleMixin):\n", + " def prepare_canvas(self, canvas: Canvas, file: str) -> ImageCanvas:\n", + " image = Image.from_file(file)\n", + " width_img, height_img = get_image_size(file)\n", + "\n", + " scale = self._calc_scale(\n", + " int(canvas.width),\n", + " int(canvas.height),\n", + " width_img,\n", + " height_img\n", + " )\n", "\n", " image_width = width_img * min(1, scale)\n", " image_height = height_img * min(1, scale)\n", - " image_x = 0\n", - " image_y = 0\n", "\n", - " if has_border:\n", - " canvas.stroke_rect(x=0, y=0, width=image_width, height=image_height)\n", - " image_width -= 2\n", - " image_height -= 2\n", - " image_x, image_y = 1, 1\n", + " return ImageCanvas(\n", + " image_widget=image,\n", + " x=0,\n", + " y=0,\n", + " width=image_width,\n", + " height=image_height,\n", + " scale=scale\n", + " )" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The other concrete strategy will force the image to have the same size as the canvas and it's developed in the following cell." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "class FitImage(ImageCanvasPrototype):\n", + " def prepare_canvas(self, canvas: Canvas, file: str) -> ImageCanvas:\n", + " image = Image.from_file(file)\n", + "\n", + " return ImageCanvas(\n", + " image_widget=image,\n", + " x=0,\n", + " y=0,\n", + " width=canvas.width,\n", + " height=canvas.height,\n", + " scale=1\n", + " )" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The draw image on canvas can be done in two ways. The default one resizes the image to fit the canvas, the second one forces the image to have the same size as the canvas. You can calibrate the strategy using the `fit_canvas` property.\n", + "\n", + "If `fit_canvas = True` then it will use the `FitImage` strategy, otherwise `ScaledImage` it will be used. The default behavior is to use the `CanvasScaleMixin`." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "class ImageRenderer:\n", + " def __init__(\n", + " self,\n", + " clear: bool = False,\n", + " has_border: bool = False,\n", + " fit_canvas: bool = False\n", + " ):\n", + " self.clear = clear\n", + " self.has_border = has_border\n", + " self.fit_canvas = fit_canvas\n", + " if fit_canvas:\n", + " self._strategy = FitImage() # type: ImageCanvasPrototype\n", + " else:\n", + " self._strategy = ScaledImage()\n", "\n", - " canvas.draw_image(sprite1,\n", - " image_x,\n", - " image_y,\n", - " width=image_width,\n", - " height=image_height)\n", - " return (image_width, image_height, scale)" + " def render(self, canvas: Canvas, file: str) -> Tuple[int, int, float]:\n", + " with hold_canvas(canvas):\n", + " if self.clear:\n", + " canvas.clear()\n", + "\n", + " image_canvas = self._strategy.prepare_canvas(canvas, file)\n", + "\n", + " if self.has_border:\n", + " canvas.stroke_rect(x=0, y=0, width=image_canvas.width, height=image_canvas.height)\n", + " image_canvas.width -= 2\n", + " image_canvas.height -= 2\n", + " image_canvas.x, image_canvas.y = 1, 1\n", + "\n", + " canvas.draw_image(\n", + " image_canvas.image_widget,\n", + " image_canvas.x,\n", + " image_canvas.y,\n", + " image_canvas.width,\n", + " image_canvas.height\n", + " )\n", + "\n", + " return image_canvas.width, image_canvas.height, image_canvas.scale" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The next example will show how the `DrawImage` behave without forcing the image to fit on canvas." ] }, { @@ -439,8 +656,30 @@ "file = \"../data/projects/bbox/pics/red400x640.png\"\n", "canvas = Canvas(width=300, height=300)\n", "draw_bg(canvas)\n", - "_, _, scale = draw_img(canvas, file)\n", - "\n", + "_, _, scale = ImageRenderer().render(canvas, file)\n", + "print(scale)\n", + "canvas" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The following example will be the same as before, but forcing the image to fit the canvas size" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "file = \"../data/projects/bbox/pics/red400x640.png\"\n", + "canvas = Canvas(width=300, height=300)\n", + "draw_bg(canvas)\n", + "_, _, scale = ImageRenderer(fit_canvas=True).render(canvas, file)\n", + "print(scale)\n", "canvas" ] }, @@ -454,7 +693,7 @@ "file = \"../data/projects/bbox/pics/green640x400.png\"\n", "canvas = Canvas(width=300, height=300)\n", "draw_bg(canvas)\n", - "_, _, scale = draw_img(canvas, file)\n", + "_, _, scale = ImageRenderer().render(canvas, file)\n", "print(scale)\n", "canvas" ] @@ -484,7 +723,7 @@ "file = \"../data/projects/bbox/pics/green640x400.png\"\n", "canvas = Canvas(width=300, height=300)\n", "draw_bg(canvas)\n", - "_, _, scale = draw_img(canvas, file)\n", + "_, _, scale = ImageRenderer().render(canvas, file)\n", "print(scale)" ] }, @@ -585,22 +824,6 @@ "outputs": [], "source": [ "#export\n", - "\n", - "def coords_point2bbox(bbox_coords: Dict[str, float]) -> List[float]:\n", - " return [bbox_coords['x'],\n", - " bbox_coords['y'],\n", - " bbox_coords['width'],\n", - " bbox_coords['height']]" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#export\n", - "\n", "def coords_scaled(bbox_coords: List[float], image_scale: float):\n", " return [value * image_scale for value in bbox_coords]" ] @@ -638,6 +861,7 @@ " bbox_selected: Optional[int]\n", " height: Optional[int]\n", " width: Optional[int]\n", + " fit_canvas: bool = False\n", "\n", " @root_validator\n", " def set_height(cls, values):\n", @@ -661,7 +885,12 @@ "class BBoxCanvasGUI(HBox):\n", " debug_output = Output(layout={'border': '1px solid black'})\n", "\n", - " def __init__(self, state: BBoxCanvasState, has_border: bool = False):\n", + " def __init__(\n", + " self,\n", + " state: BBoxCanvasState,\n", + " has_border: bool = False,\n", + " drawing_enabled: bool = True\n", + " ):\n", " super().__init__()\n", "\n", " self._state = state\n", @@ -669,6 +898,7 @@ " self.is_drawing = False\n", " self.has_border = has_border\n", " self.canvas_bbox_coords: Dict[str, Any] = {}\n", + " self.drawing_enabled = drawing_enabled\n", "\n", " # do not stick bbox to borders\n", " self.padding = 2\n", @@ -679,22 +909,31 @@ " align_items='center',\n", " align_content='center',\n", " overflow='hidden'))\n", - " self.multi_canvas = MultiCanvas(\n", - " len(BBoxLayer),\n", - " width=self._state.width,\n", - " height=self._state.height\n", - " )\n", "\n", - " self.im_name_box = Label()\n", + " if not drawing_enabled:\n", + " self.multi_canvas = MultiCanvas(\n", + " len(BBoxLayer),\n", + " width=self._state.width,\n", + " height=self._state.height\n", + " )\n", + " self.children = [VBox([self.multi_canvas])]\n", + " else:\n", + " self.multi_canvas = MultiCanvas(\n", + " len(BBoxLayer),\n", + " width=self._state.width,\n", + " height=self._state.height\n", + " )\n", + "\n", + " self.im_name_box = Label()\n", "\n", - " children = [VBox([self.multi_canvas, self.im_name_box])]\n", - " self.children = children\n", - " draw_bg(self.multi_canvas[BBoxLayer.bg])\n", + " children = [VBox([self.multi_canvas, self.im_name_box])]\n", + " self.children = children\n", + " draw_bg(self.multi_canvas[BBoxLayer.bg])\n", "\n", - " # link drawing events\n", - " self.multi_canvas[BBoxLayer.drawing].on_mouse_move(self._update_pos)\n", - " self.multi_canvas[BBoxLayer.drawing].on_mouse_down(self._start_drawing)\n", - " self.multi_canvas[BBoxLayer.drawing].on_mouse_up(self._stop_drawing)\n", + " # link drawing events\n", + " self.multi_canvas[BBoxLayer.drawing].on_mouse_move(self._update_pos)\n", + " self.multi_canvas[BBoxLayer.drawing].on_mouse_down(self._start_drawing)\n", + " self.multi_canvas[BBoxLayer.drawing].on_mouse_up(self._stop_drawing)\n", "\n", " @property\n", " def highlight(self) -> BboxCoordinate:\n", @@ -723,7 +962,7 @@ "\n", " self._state.set_quietly('bbox_selected', index)\n", "\n", - " @debug_output.capture(clear_output=False)\n", + " @debug_output.capture(clear_output=True)\n", " def _update_pos(self, x, y):\n", " # print(f\"-> BBoxCanvasGUI::_update_post({x}, {y})\")\n", " if self.is_drawing:\n", @@ -744,7 +983,7 @@ " self.canvas_bbox_coords[\"x\"] < self.padding or\n", " self.canvas_bbox_coords[\"y\"] < self.padding)\n", "\n", - " @debug_output.capture(clear_output=False)\n", + " @debug_output.capture(clear_output=True)\n", " def _stop_drawing(self, x, y):\n", " # print(f\"-> BBoxCanvasGUI::_stop_drawing({x}, {y})\")\n", " self.is_drawing = False\n", @@ -833,8 +1072,13 @@ "class BBoxVideoCanvasGUI(BBoxCanvasGUI):\n", " debug_output = Output(layout={'border': '1px solid black'})\n", "\n", - " def __init__(self, state: BBoxCanvasState, has_border: bool = False):\n", - " super().__init__(state, has_border)\n", + " def __init__(\n", + " self,\n", + " state: BBoxCanvasState,\n", + " has_border: bool = False,\n", + " drawing_enabled: bool = True\n", + " ):\n", + " super().__init__(state, has_border, drawing_enabled)\n", "\n", " @property\n", " def highlight(self) -> BboxCoordinate:\n", @@ -983,14 +1227,20 @@ "\n", " @debug_output.capture(clear_output=True)\n", " def _draw_image(self, image_path: str):\n", - " # print(f\"-> _draw_image {image_path}\")\n", + " print(f\"-> _draw_image {image_path}\")\n", " self.clear_all_bbox()\n", - " image_width, image_height, scale = draw_img(\n", - " self._gui.multi_canvas[BBoxLayer.image],\n", - " image_path,\n", + "\n", + " img_renderer_service = ImageRenderer(\n", " clear=True,\n", - " has_border=self._gui.has_border\n", + " has_border=self._gui.has_border,\n", + " fit_canvas=self._state.fit_canvas\n", " )\n", + "\n", + " image_width, image_height, scale = img_renderer_service.render(\n", + " self._gui.multi_canvas[BBoxLayer.image],\n", + " image_path\n", + " )\n", + "\n", " self._state.set_quietly('image_width', image_width)\n", " self._state.set_quietly('image_height', image_height)\n", " self._state.image_scale = scale\n", @@ -1046,12 +1296,23 @@ " Gives user an ability to draw a bbox with mouse.\n", " \"\"\"\n", "\n", - " def __init__(self, width, height, has_border: bool = False):\n", + " def __init__(\n", + " self,\n", + " width,\n", + " height,\n", + " has_border: bool = False,\n", + " fit_canvas: bool = False,\n", + " drawing_enabled: bool = True\n", + " ):\n", " self.state = BBoxCanvasState(\n", " uuid=str(id(self)),\n", - " **{'width': width, 'height': height}\n", + " **{'width': width, 'height': height, 'fit_canvas': fit_canvas}\n", + " )\n", + " super().__init__(\n", + " state=self.state,\n", + " has_border=has_border,\n", + " drawing_enabled=drawing_enabled\n", " )\n", - " super().__init__(state=self.state, has_border=has_border)\n", " self._controller = BBoxCanvasController(gui=self, state=self.state)\n", " self._bbox_history: List[Any] = []\n", "\n", @@ -1097,14 +1358,7 @@ " **{'width': width, 'height': height}\n", " )\n", " self.drawing_enabled = drawing_enabled\n", - " super().__init__(state=self.state, has_border=has_border)\n", - " if not drawing_enabled:\n", - " self.multi_canvas = MultiCanvas(\n", - " len(BBoxLayer),\n", - " width=self._state.width,\n", - " height=self._state.height\n", - " )\n", - " self.children = [VBox([self.multi_canvas, self.im_name_box])]\n", + " super().__init__(state=self.state, has_border=has_border, drawing_enabled=drawing_enabled)\n", "\n", " self._controller = BBoxVideoCanvasController(gui=self, state=self.state)" ] @@ -1120,6 +1374,29 @@ "gui" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "You can't draw on the following canvas" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "BBoxCanvas(width=100, height=100, has_border=True, drawing_enabled=False)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "You can't draw on the following canvas" + ] + }, { "cell_type": "code", "execution_count": null, @@ -1185,7 +1462,8 @@ "source": [ "# hide\n", "# gui._state.image_path = \"../data/projects/bbox/pics/red400x640.png\"\n", - "gui._state.image_path = '../data/projects/im2im1/class_images/blocks_1.png'" + "gui._state.image_path = '../data/projects/im2im1/class_images/blocks_1.png'\n", + "gui._controller._draw_image('../data/projects/im2im1/class_images/blocks_1.png')" ] }, { diff --git a/nbs/01_helpers.ipynb b/nbs/01_helpers.ipynb index fae5ebc..21c04c3 100644 --- a/nbs/01_helpers.ipynb +++ b/nbs/01_helpers.ipynb @@ -31,13 +31,6 @@ "from IPython.display import display" ] }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "# Helpers" - ] - }, { "cell_type": "code", "execution_count": null, @@ -47,6 +40,7 @@ "#exporti\n", "\n", "import pandas as pd\n", + "from typing import Union\n", "\n", "try:\n", " from collections.abc import Iterable\n", @@ -255,7 +249,7 @@ "from ipyannotator.datasets.factory import DS as NDS\n", "from ipyannotator.datasets.factory_legacy import DS, _combine_train_test\n", "from pathlib import Path\n", - "from tqdm import tqdm\n", + "from tqdm.notebook import tqdm\n", "\n", "\n", "class Tutorial:\n", @@ -264,7 +258,7 @@ "\n", " \"\"\"\n", "\n", - " def __init__(self, dataset: DS, project_path):\n", + " def __init__(self, dataset: Union[DS, NDS], project_path):\n", " self.dataset = dataset\n", " self.project_path = project_path\n", " if self.dataset not in [DS.ARTIFICIAL_CLASSIFICATION, DS.ARTIFICIAL_DETECTION,\n", @@ -362,7 +356,7 @@ " improver.capture_state.annotations[k] = {'answer': v_expl != v_cret}\n", " improver.view._navi._next_btn.click()\n", "\n", - " def annotate_video_bboxes(self, annotator):\n", + " def annotate_video_bboxes(self, annotator) -> dict:\n", " mot_gt = pd.read_csv(self.project_path / 'mot.csv')\n", " mot_gt.columns = [\n", " 'frame',\n", @@ -380,7 +374,8 @@ " full_path = f'{self.project_path}/images'\n", " mot_gt['frame'] = mot_gt['frame'].apply(lambda x: full_path + '/' + x + '.jpg')\n", " mot_gt.index = mot_gt['frame']\n", - " mot_gt = mot_gt[mot_gt.columns.drop(['frame', 'conf', 'label', 'vis'])]\n", + " mot_gt = mot_gt.drop(columns=['frame', 'conf', 'label', 'vis'])\n", + "# mot_gt = mot_gt[mot_gt.columns.drop(['frame', 'conf', 'label', 'vis'])]\n", " mot_gt = mot_gt.groupby('frame').apply(lambda x: x.to_json(orient='records'))\n", " result = mot_gt.to_json(orient='index')\n", " parsed = json.loads(result)\n", @@ -413,6 +408,8 @@ " with open(self.project_path / 'create_results/annotations.json', 'w+') as f:\n", " json.dump(annotations, f)\n", "\n", + " return annotations\n", + "\n", " def _mutate_id(self, bbox: dict, index: int) -> str:\n", " id = '2'\n", " if bbox['height'] == bbox['width']:\n", @@ -446,7 +443,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/01a_datasets.ipynb b/nbs/01a_datasets.ipynb index 3c85802..a7cffdd 100644 --- a/nbs/01a_datasets.ipynb +++ b/nbs/01a_datasets.ipynb @@ -953,7 +953,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/01a_datasets_download.ipynb b/nbs/01a_datasets_download.ipynb index 1adc1de..93cd98c 100644 --- a/nbs/01a_datasets_download.ipynb +++ b/nbs/01a_datasets_download.ipynb @@ -306,7 +306,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/01a_datasets_factory.ipynb b/nbs/01a_datasets_factory.ipynb index e0c7015..1cb5dc9 100644 --- a/nbs/01a_datasets_factory.ipynb +++ b/nbs/01a_datasets_factory.ipynb @@ -196,7 +196,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/01b_dataset_video.ipynb b/nbs/01b_dataset_video.ipynb index 2fa1aa0..b4819ad 100644 --- a/nbs/01b_dataset_video.ipynb +++ b/nbs/01b_dataset_video.ipynb @@ -38,7 +38,7 @@ "id": "0745ce2f", "metadata": {}, "source": [ - "## Generators for MOT data\n", + "# Generators for MOT data\n", "\n", "MOT data format can be [found here](https://github.com/JonathonLuiten/TrackEval/blob/master/docs/MOTChallenge-Official/Readme.md#data-format)." ] @@ -122,7 +122,7 @@ "id": "a05b7045", "metadata": {}, "source": [ - "### Create mot gt and render corresponding frames" + "## Create mot gt and render corresponding frames" ] }, { @@ -510,7 +510,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/01b_tutorial_image_classification.ipynb b/nbs/01b_tutorial_image_classification.ipynb index 81ce51d..81bab3e 100644 --- a/nbs/01b_tutorial_image_classification.ipynb +++ b/nbs/01b_tutorial_image_classification.ipynb @@ -32,27 +32,52 @@ "import ipywidgets as widgets\n", "\n", "from pathlib import Path\n", - "from tqdm import tqdm\n", + "from tqdm.notebook import tqdm\n", "\n", "from ipyannotator.base import Settings\n", - "from ipyannotator.mltypes import InputImage, OutputImageLabel\n", + "from ipyannotator.mltypes import InputImage, OutputImageLabel, NoOutput, OutputLabel\n", "from ipyannotator.annotator import Annotator\n", "from ipyannotator.datasets.factory_legacy import DS, get_settings, _combine_train_test\n", - "from ipyannotator.helpers import Tutorial" + "from ipyannotator.helpers import Tutorial\n", + "from ipyannotator.doc_utils import is_building_docs" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "# Tutorial: Image classification" + "# Image classification - Assigning meaning to images via classes\n", + "\n", + "The current tutorial will illustrate how to use Ipyannotator to classify images. \n", + "\n", + "The task of identifying what an image represents is called image classification.\n", + "\n", + "**Ipyannotator** allows users to **explore** an entire set of images and specific labels; manually **create** their datasets associating labels to images; **improve** existing annotations.\n", + "\n", + "This tutorial is divided in the following steps:\n", + "\n", + "- [Select dataset](#Select-Dataset)\n", + "- [Setup annotator](#Setup-annotator)\n", + "- [Explore](#Explore)\n", + "- [Create](#Create)\n", + "- [Improve](#Improve)\n", + "- [Postprocessing](#Postprocessing)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "## Select Dataset" + "## Select dataset\n", + "\n", + "For this tutorial you can select four different datasets:\n", + "\n", + "- **Artificial Classification** is a minimal dataset generated by Ipyannotator with 50 images in 3 classes to be labeled. It doesn't require downloading and is used by default for this tutorial.\n", + "- [Cifar10](https://www.cs.toronto.edu/~kriz/cifar.html) is a dataset with 60000 images in 10 classes of animals and passenger transporation vessels to be labeled.\n", + "- [Oxford102](https://www.tensorflow.org/datasets/catalog/oxford_flowers102) is a dataset with 6149 images in 102 classes of flowers to be labeled.\n", + "- [Cub200](http://www.vision.caltech.edu/visipedia/CUB-200.html) is a dataset with 6033 images in 200 classes of birds to be labeled.\n", + "\n", + "You can choose between the datasets uncommenting the following cell." ] }, { @@ -61,10 +86,8 @@ "metadata": {}, "outputs": [], "source": [ - "# You can choose between 3 datasets ['cifar10', 'oxford_flowers', 'CUB_200'] that you can download.\n", - "# We use a artifical generated classification dataset by default that doesn't require downloading.\n", "dataset = DS.ARTIFICIAL_CLASSIFICATION\n", - "# dataset = DS.CIFAR10\n", + "#dataset = DS.CIFAR10\n", "# dataset = DS.OXFORD102\n", "# dataset = DS.CUB200" ] @@ -73,7 +96,18 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "## Setup annotator" + "You don't need to download the data manually, it will be done automatically in the next step for datasets other than `DS.ARTIFICIAL_CLASSIFICATION`." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Setup annotator\n", + "\n", + "This section will set up the paths and input/output data needed to classify the images.\n", + "\n", + "The following cell imports the project file and directory where the images were downloaded (or generated). For this tutorial, we simplify the process using the `get_settings` function instead of hardcoding the paths." ] }, { @@ -88,6 +122,17 @@ "settings_.project_file, settings_.image_dir" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Ipyannotator uses pairs of input/output data to set up the annotation. \n", + "\n", + "The image classification annotator uses `InputImage` and `OutputImageLabel`as the pair to set up the annotator.\n", + "\n", + "The `InputImage` function provides information about the directory that contains the images to be classified, and the images itself. The `OutputImageLabel` function provides information about the directory that contains the classes that can be associated with the images and labels itself." + ] + }, { "cell_type": "code", "execution_count": null, @@ -104,6 +149,15 @@ "input_.dir, output_.dir" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The final part in setting up the Ipyannotator is the configuration of the `Annotator` factory with the pair of input/output data. \n", + "\n", + "The factory allows three types of annotator tools: explore, create, improve. The next sections will guide you through every step." + ] + }, { "cell_type": "code", "execution_count": null, @@ -117,14 +171,14 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "## EXPLORE" + "## Explore" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "You can explore dataset with `next/previous` buttons to check visualized labels." + "The **explore** option allows users to navigate across the images in the dataset using `next/previous` buttons. In case the dataset was already labeled, the labeling results can also be displayed. This function is used for data visualization only, improvement and addition of labels is done in the next steps. " ] }, { @@ -133,28 +187,41 @@ "metadata": {}, "outputs": [], "source": [ - "# explorer = anni.explore(k=3)\n", "explorer = anni.explore()\n", "explorer" ] }, { - "attachments": { - "Screenshot%20from%202022-01-28%2011-55-14.png": { - "image/png": "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" - } - }, "cell_type": "markdown", "metadata": {}, "source": [ - "![Screenshot%20from%202022-01-28%2011-55-14.png](attachment:Screenshot%20from%202022-01-28%2011-55-14.png)" + "Sometimes the classes are not defined yet or incomplete. To explore the input images without worring about any classes you can use the `NoOutput` option on the annotator factory which is done in the following:" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "unlabel_factory = Annotator(input_, NoOutput(), settings_)\n", + "unlabel_factory.explore()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "## CREATE" + "## Create" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The **create** option allows users to manually create their annotated datasets. \n", + "\n", + "The next cell removes already created results for any dataset that can be chosen in this tutorial." ] }, { @@ -177,9 +244,9 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "You can try to annotate by hand,\n", - "but be sure to have some pieces incorrectly annotated,\n", - "thus you prepare good set for `improve` step below" + "The next cell initializes the **create** option. \n", + "\n", + "For this tutorial, a function was defined that imitates human work. You can choose between performing the annotation manually yourself or letting the function do the work for you." ] }, { @@ -189,27 +256,16 @@ "outputs": [], "source": [ "creator = anni.create()\n", - "\n", "creator" ] }, - { - "attachments": { - "image_classification_one.png": { - "image/png": "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" - } - }, - "cell_type": "markdown", - "metadata": {}, - "source": [ - "![image_classification_one.png](attachment:image_classification_one.png)" - ] - }, { "cell_type": "markdown", "metadata": {}, "source": [ - "#### To imitate human work on currnt step, let's randomly annotate all the images automatically:" + "The next cell imitates human work by automatically annotating all images randomly. If you want to manually annotate then skip the next step.\n", + "\n", + "If you choose to annotate manually be sure to have some images incorrectly annotated. In this way you prepare a good dataset for the **improve** step below." ] }, { @@ -218,8 +274,6 @@ "metadata": {}, "outputs": [], "source": [ - "#SKIP THIS STEP IF YOU ANNOTATE MANUALLY\n", - "\n", "HELPER = Tutorial(dataset, settings_.project_path)\n", "HELPER.annotate_randomly(creator)" ] @@ -228,7 +282,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Let's get first 3 images marked for one class, depending on dataset (RED, CAT, etc..)" + "The example above shows how to set up the image classification using already predefined labels. Occasionally, you may want to create a dataset with your own text labels, for example 'Circle' and 'Rectangle'. You can create a dataset with new output labels like this:" ] }, { @@ -237,55 +291,47 @@ "metadata": {}, "outputs": [], "source": [ - "if dataset == DS.ARTIFICIAL_CLASSIFICATION:\n", - " #RED\n", - " print([k for k, v in creator.to_dict().items() if 'red.jpg' in v][:3])\n", - "elif dataset == DS.CIFAR10:\n", - " #CAT\n", - " print([k for k, v in creator.to_dict().items() if 'cat.jpg' in v][:3])\n", - "else:\n", - " pass" + "output_label = OutputLabel(class_labels=['Circle', 'Rectangle'])" ] }, { - "cell_type": "markdown", + "cell_type": "code", + "execution_count": null, "metadata": {}, + "outputs": [], "source": [ - "## IMPROVE" + "text_label_factory = Annotator(input_, output_label, settings_)\n", + "text_label_factory.create()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "Lets group some annotators together, so we can go through all annotated images but for each classs separately.\n", + "## Improve\n", + "\n", + "The **improve** feature allows users to refine the annotated dataset. This feature groups the annotated images according to their class and edits each class separately. This means that if your dataset has 3 labeling classes, 3 annotators instances are initiated to improve each class separately.\n", + "\n", + "As before, for the purpose of the tutorial, a function can be used to performe the annotation and you don't have to annotate manually. If you want to annotate manually then make sure to __mark all errors__ (images, which belongs to __different__ class).\n", "\n", - "Each grid shows images belonging to the __same__ class. \n", "\n", - "You should __mark all errors__ (images, which belongs to __different__ class)" + "If you chose to annotate manually don't forget to click the __SAVE__ button when finished with each class." ] }, { - "cell_type": "markdown", + "cell_type": "code", + "execution_count": null, "metadata": {}, + "outputs": [], "source": [ - "!! Dont forget to click __SAVE__ button when finished with each class:" + "all_improvers = anni.improve()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "#### improve" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "all_improvers = anni.improve()" + "Check the number of classes:" ] }, { @@ -301,7 +347,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "let's select first two classes to mark the errors" + "Let's select the first two classes to mark incorrectly labeled images:" ] }, { @@ -313,23 +359,11 @@ "all_improvers[:2]" ] }, - { - "attachments": { - "image_classification_two.png": { - "image/png": "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" - } - }, - "cell_type": "markdown", - "metadata": {}, - "source": [ - "![image_classification_two.png](attachment:image_classification_two.png)" - ] - }, { "cell_type": "markdown", "metadata": {}, "source": [ - "#### To imitate human work on current step, let's fix the incorrect annotations automatically:" + "The next cell imitates the human work. If you chose to annotate manually make sure to skip this cell." ] }, { @@ -338,7 +372,6 @@ "metadata": {}, "outputs": [], "source": [ - "#SKIP THIS STEP IF YOU CHECK ANNOTATIONS MANUALLY\n", "HELPER.fix_incorrect_annotations(all_improvers)" ] }, @@ -346,7 +379,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Now we can get list of all marked images, which should be reclassified:" + "Now we obtain a list of all marked images that need to be reclassified:" ] }, { @@ -365,16 +398,16 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "### Postprocessing.." + "### Postprocessing" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "Also, automatically generarted json file for each class can be used for further processing.\n", + "This section exemplifies how to process the data after the improve feature has been applied. By default, the improve feature creates a `missed` folder in your storage with a folder for every class available in the dataset.\n", "\n", - "Let's load one json for a random class and show filenames marked as incorrect on previous step for this class:" + "The next cell loads one JSON file for a random class and displays the filenames of images marked as incorrectly labeled in the previous step." ] }, { @@ -392,11 +425,12 @@ "\n", "random_class_annotation = pd.read_json(Path(random_class) / 'annotations.json').T\n", "\n", - "random_misssed = list(\n", - " random_class_annotation[random_class_annotation['answer'] == True].index.values) # noqa: E712\n", + "anwered_missed = random_class_annotation[random_class_annotation['answer'] == True] # noqa: E712\n", + "\n", + "random_missed = list(anwered_missed.index.values)\n", "\n", - "# show 10 files with incorrect label for the random class\n", - "random_misssed[:10]\n", + "# shows selected random class and 10 files with incorrect labels within that class\n", + "random_class, random_missed[:10]\n", "\n", "# result may be empty, if all annotations are correct" ] diff --git a/nbs/01c_tutorial_bbox.ipynb b/nbs/01c_tutorial_bbox.ipynb index 12f0709..96e0bc1 100644 --- a/nbs/01c_tutorial_bbox.ipynb +++ b/nbs/01c_tutorial_bbox.ipynb @@ -39,7 +39,7 @@ "\n", "from ipyannotator.annotator import Annotator\n", "from ipyannotator.base import Settings\n", - "from ipyannotator.mltypes import InputImage, OutputImageBbox, OutputImageLabel\n", + "from ipyannotator.mltypes import InputImage, OutputImageBbox, OutputImageLabel, NoOutput\n", "from ipyannotator.datasets.factory_legacy import DS, get_settings, _combine_train_test\n", "from ipyannotator.datasets.generators import create_object_detection, xyxy_to_xywh, xywh_to_xyxy\n", "from ipyannotator.helpers import Tutorial\n", @@ -50,14 +50,30 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "# Tutorial: BBox" + "# Bounding Box Annotator - Identifying objects in images through boxes \n", + "\n", + "The current tutorial will illustrate how to use Ipyannotator to annotate images using bounding boxes.\n", + "\n", + "The task of identifying what an image represents is called image classification. Often an image contains multiple objects which can be identified individually.\n", + "\n", + "**Ipyannotator** allows users to **explore** an entire set of images and given labels; manually **create** datasets by associating labels to bounding boxes drawn on top of the images; **improve** existing annotations.\n", + "\n", + "This tutorial is divided in the following steps:\n", + "\n", + "- [Select dataset](#Select-Dataset)\n", + "- [Setup annotator](#Setup-annotator)\n", + "- [Explore](#Explore)\n", + "- [Create](#Create)\n", + "- [Improve](#Improve)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "## Select Dataset" + "## Select dataset\n", + "\n", + "For this tutorial an artificial minimal dataset is generated by Ipyannotator with 50 images in 2 classes to be labeled (circle and rectangle)." ] }, { @@ -66,8 +82,6 @@ "metadata": {}, "outputs": [], "source": [ - "# We use an artifical generated classification dataset by default that doesn't require downloading.\n", - "\n", "dataset = DS.ARTIFICIAL_DETECTION" ] }, @@ -75,7 +89,11 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "## Setup annotator" + "## Setup annotator\n", + "\n", + "This section will set up the paths and the input/output pair needed to classify the images.\n", + "\n", + "The following cell imports the project file and directory where the images were generated. For this tutorial we simplify the process using the `get_settings` function instead of hardcoding the paths." ] }, { @@ -84,12 +102,21 @@ "metadata": {}, "outputs": [], "source": [ - "# get special project settings for selected dataset\n", - "\n", "settings_ = get_settings(dataset)\n", "settings_.project_file, settings_.image_dir" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Ipyannotator uses pairs of input/output data to set up the annotation. \n", + "\n", + "The Bounding Box annotator uses `InputImage` and `OutputImageBox`as the pair to set up the annotator.\n", + "\n", + "The `InputImage` function provides information about the directory that contains the images to be labeled, and the images itself. The `OutputImageBox` function provides information about the directory that contains the classes that can be associated with the bounding boxes drawn on the images." + ] + }, { "cell_type": "code", "execution_count": null, @@ -105,6 +132,15 @@ "input_.dir" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The final part in setting up the Ipyannotator is the configuration of the `Annotator` factory with the pair of input/output data. \n", + "\n", + "The factory allows three types of annotator tools: explore, create, improve. The next sections will guide you through every step." + ] + }, { "cell_type": "code", "execution_count": null, @@ -118,8 +154,9 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "## EXPLORE\n", - "You can explore dataset with next/previous buttons to check visualized bounding boxes." + "## Explore\n", + "\n", + "The **explore** option allows users to navigate across the images in the dataset using `next/previous` buttons. In case the dataset was already labeled, the labeling results can also be displayed. This function is used for data visualization only, improvement and addition of labels is done in the next steps. " ] }, { @@ -133,15 +170,10 @@ ] }, { - "attachments": { - "bbox_one.png": { - "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgQAAACnCAIAAADhd9OoAAAACXBIWXMAAA7EAAAOxAGVKw4bAAAXeklEQVR4nO3da1gTZ9oH8Hs4VkqCgEACCmipeCigdlsQXUSLqKXWooBCbVkFDyhUUbiqeBUrnisg6hbBBdzVsq6HqisqWl1b0eIJt1TARQTEIBipEkI4yCGZ98Psm4tyiIAExP3/Ps1Mnpm5J63zn3meYcKwLEsAAPC/TaOvCwAAgL6HMAAAAIQBAAAgDAAAgBAGAABACAMAACCEAQAAEMIAAAAIYQAAAESk1bObk8vljx492r59e1lZma6ubje2wLIswzAWFhZRUVE8Hq9nywMAgHb1cBiwLFtXVycWi01MTN5+++36+vquboFhmMrKytzc3Obm5p6tDQAAOtLDYcAxMjKytra2trauklbpvKH1lH2mIIWqFRgiloyZQdVPq42MjGxtbfPz8xmGUUdtAADQllrCgIiKioqeVT2t0axuNm8sVBTLqZk6PrezxBJp2DA2j66WjRSMmjhxopqqAgCAdqkrDIjI0MpA/12d2/ZZb5IewyiIOn49KsswRJWsWKOB+BK++koCAIB2qS8M2Afsw4eKEl2GYUj+grYMyxKxDCtlq2RstdpKAgCA9qnr0dJGtqmeahs06rq0loJhXzC6AAAAaqCuMGAZhYJRUJfP7CyLYWMAgF6nljBgCT+fBgDQn6glDBSkYFUMF6uEGwMAgN6npjsDlc8OAQDAKwbvJgIAAHWFATp7AAD6E7WEgQYxyAMAgH5ELWHAkAaiAACgH1FTGDDMf+8MEAoAAP2A+gaQGWK71VmEp5AAAHqdut5NpEM6mv9951CnMCyxDEvEEoM0AADobeoKA4YYUxJok3Y5+7Bzv0zAyLWa3tR4U5/FW0sBAHqbGl9hrVGppfeLvpHE5CF7v5lYtoNIYImIFAxpjNQe9aToaZNuk/pKAgCAdqkrDExNTS1NLKW1UqHITI99o5lYai8M/r9LSMGQxiiNd8x4T01MTPEbZwAAvayHw4BhGA0NjcbGRg0NDV1dXa1aLarRGk52nVx90CATPT290tJSHR2dni0MAABU6Pk7Az09Pblcfu7cOU1NzXYbKH+xoMUTqL/z5ptv2tnZdbQ6AAD0OIbt6bdNy+Xympoaufz3v27Gcj90TAwxRCwxzH9fa8qNJPw+ERiG0dLS0tfXR38RAEDv6PkwAACAfgdvLQUAAIQBAAAgDAAAgBAGAABACAMAACCEAQAAEMIAAAAIYQAAAIQwAAAAQhgAAAAhDAAAgBAGAABACAMAACCEAQAAEMIAAAAIYQAAAIQwAAAAQhgAAAAhDAAAgBAGAABACAMAACCEAQAAEMIAAAAIYQAAAIQwAAAAQhgAAAAhDAAAgBAGAABACAMAACCEAQAAEMIAAAAIYQAAANQ2DHx9fXft2tXju7ly5cq4ceOMjIzmzZtXVVXV6tPDhw9PnTq1x3cKAACd1DoMgoKC3N3de3YfNTU1M2fOXL58eW5ubm1t7cqVK1s1cHR0DA8P79mdAgBA5zEsy6p7H0lJSfv27bt58yYR5eXlvffee0+ePOHxeOreLwAAdFLrO4PPPvuM6yYKCgpatmyZjY0Nj8f7+OOPs7Kyxo8fb2Ji4u/vX19fT0RlZWXTp08fOHDgjBkzFi9evGHDho72kZeX5+TkxE2PHj1aS0vrwYMHLRsou4nGjh27devWIUOG8Pn8Dz/88NGjR1yD+Ph4Kysra2vrhIQEhmF67vABAIBI9QDy0aNHz507l5eXl5WV9fHHHx86dOj69eunTp1KS0sjonnz5o0ZM6aoqOizzz5LSkpSsR2xWGxsbKycNTIyevLkSUeN9+zZc/To0YKCAiMjI29vbyK6ePHi1q1bT548eeXKldTU1C4fIgAAvIiWis98fX1tbGyIyMXFZfDgwdbW1kQ0YcIEiURSWFiYnZ39r3/9S0dHx8/Pb+/evSq2o1AoWi1pbm7uqPHq1au524j4+HgTE5PCwsK//vWvy5YtGzt2LBF9/fXXbm5unT04AADoHFV3BsrLeW1tbaFQqJwmIpFIJBQKdXR0uIVcTnTEzMyssrJSOSuRSAQCQUeNufghIj6fLxAIRCKRSCSysrLqzI4AAKB7uvl3BkKhUCwWNzU1cbOlpaUqGjs4OGRlZXHTBQUFTU1NyjN+W0VFRdxEbW3tkydPLCwshEKhSCTqzI4AAKB7uhkGI0aMGD16dFRUVHV19bFjx65du6ZiXNfb2/vu3bsnT56sr6+PioqaO3cu9yjR6dOn//Of/7RqHBsbe/v27WfPnoWEhNjb2w8fPnzBggV79+7Ny8sTi8VbtmzBADIAQI/rZhgwDHPkyJHMzEwrK6v09HRPT08jI6OOGvP5/DNnzmzcuNHS0lIul//5z3/mloeEhJw6dapV48WLFwcGBg4dOrS8vPzYsWMMw0yfPn316tVubm6urq7z5s1TsSMAAOie1gPIBw8e5CZajgkrFxLRiRMniEgul589ezYtLU1PT4+IXFxczM3NVezG2dn59u3brRa2esCU4+joGBkZ2XLJjRs3xowZ8/jxYyK6dOmS6h0BAEA3qHqaSAVNTc3du3eXlZWtWrUqIyPjzp077u7ukZGR7Z7fW2ZJWyzL5ubmqvgbNJFItHr16nPnzhkbG2/bts3Ly6t7NQMAQEe6GQZEdPjw4VWrVtnb27/11lv//Oc/9fX1o6KiurGdmJiY8+fPx8TEENHnn3/e9nkhb2/v/Px8T09PTU3N6dOnh4WFdbtmAABoV2+8jgIAAF5xeIU1AAAgDAAAAGEAAACEMAAAAEIYAAAAIQwAAIAQBgAAQAgDAAAghAEAABDCAAAACGEAAACEMAAAAEIYAAAAIQwAAIAQBgAAQAgDAAAghAEAABDCAAAACGEAAACEMAAAAEIYAAAAIQwAAIAQBgAAQAgDAAAghAEAABDCAAAACGEAAACEMAAAAEIYAAAAIQwAAIAQBgAAQAgDAAAghAEAABDCAAAACGEAAACEMAAAAEIYAAAAIQwAAIAQBgAAQAgDAAAghAEAABDCAAAACGEAAACEMAAAAEIYAAAAIQwAAICItFrNSySSPqmjDxkaGvZ1CQDQjv51Omp5JnkFK3/hiQ53BgAAgDAAAACEAQAAEMIAAAAIYQAAANT2aSKAVioqKgoLCx8/flxdXU1EfD5fKBTa2NiYmpr2dWkA0GN6+M6gurp6Sgve3t4JCQlyubzbG5wzZ05RUVEPVtjLbt++zTDMy3wD3fC3v/1t4MCBVVVV3GxtbS33VJlEIvH29jY0NHRxcbl//35nNpWZmZmRkcHj8dzc3BYsWLBgwQI3Nzcej5eRkZGZmanGY4BXQ3x8/LBhwywsLCIjIxUKRV+X8zsKhWLKlCmxsbEtF4aGhp49e7avSuq8oqKiNWvWfPLJJ97e3pGRkcXFxX1dkXq6ifbu3Xvy5Mnvv/8+NDT0xIkTaWlp3d5USEgIrkC7QSqVxsTEtFoYGBhobGx87do1T0/Pjz76qLm5WfVGzp8/T0ReXl52dnYGBgbcQgMDAzs7Oy8vL2UDeF1lZGSsW7fu73//+8WLFw8ePHjgwIG+rqgdZ86cuXfvXrdXf/z4cVBQUA/W0xm1tbXh4eFDhw7dvXv3X/7yF3t7+7CwMLFY3MtltKKWMODxeHw+39DQ0NnZ2cnJ6d///nfLTxsbGzu/KVdXVx6P19MFvv6mTZuWmJj47Nkz5ZKqqqqrV6/u2bNnxIgRoaGhZmZmP//8s4otZGZm8ng8Z2fnjho4OzvzeDzcH7zG9u3bt2jRIicnp5EjR4aFhSUlJfV1Re1wc3PbuXNnt+9ampube/8snJ+fL5fLlyxZYmlpOXDgQC8vr2HDhvX5P6WXCgOWZbl+ZBWUnSTR0dEnTpxYt27d9u3biSgzMzMwMHDWrFnr169/9OhRU1PTrFmzsrOzubUyMjL8/PxYlvXx8eG6iXJycpYtW+bh4bFo0aKbN29yzebMmVNSUsJN7969OyUlhYhKSkpWrlw5Y8aMuXPnvsxNSY8bOXLkV199ZWJiYmRkFBERcfz4cVtbWysrq61bt3IN0tLSHBwc9PT0rK2tv/32W27hlStXHBwcBg0aFBwc7OzsfPnyZSKSSqULFiwwMTFxcHCIjo5uu68hQ4b4+fnt2LFDuUQkEg0dOlRbW1tZzMOHDzsqtaKiory8XEUScJydncvLyysqKrryNUC/kZeX5+TkxE07OTndvXu3b+tpl5+fX3V19enTp1stVygUqampfn5+Pj4+sbGxDQ0NYrF41qxZ3Bnj7t27Xl5eZWVlK1asqK6u9vHxaWhoUK4rk8nU2ic2ePBgmUyWlpamvDuPjo6ePXs2ET18+DAsLMzDw+OTTz6JjY2Vy+WHDx9euXKlct2FCxeeP3++7dG9fFXdDIOamppjx475+/tfv3697ad1dXW1tbUymezmzZvXr18fM2YMt/zo0aPe3t4RERHFxcUbN25cunTpd999N2LEiIiIiObm5gkTJly5coVr+eOPP06bNo1hGG5WIpGsWbNmxowZ//jHPz799NPIyMjHjx93VNvOnTuHDx9+6NCh5cuXx8XFSaXS7h2jOmRkZBQVFR0+fHjr1q0JCQm3b9/euXNnRERERUWFTCbz9vZeuXJlaWlpTEzMF198IZVKKysrZ82aFRkZmZ+fr6mpee3aNW47gYGBMpksJycnNTU1JSUlMTGx7b7Wrl2bkpKiPFMLhcKSkhLl6MWDBw8sLCw6qrOwsNDW1rYzR2Rra1tYWNi1bwH6CbFYbGxszE0bGxtLJJIu3db3jjfeeCMkJCQ5OVk5SMY5derUDz/8sGXLloSEhOrq6m+++UYgEAQEBOzYsaOxsXHHjh2hoaEWFha7du3i8/lHjhzR1dVVrpubmzt//vzU1NRW2+wpZmZmy5Yt27dvn5eXV1RUVHp6uvKLjY6OHjx48MGDB2NiYq5evXrjxo3Jkyfn5uZy77coKSkRi8UuLi5tj+7lq+ry00TFxcUnT568efPm5MmTt23bZm5u3rbN4sWLuQl9fX13d3cu8Yho7NixXDCkp6e7urr+4Q9/ICJfX9+zZ89mZ2d/8MEH27dvDw4ObmxsvHHjhnIjRHT16lUbG5uZM2cSkaur68WLFy9duvTpp5+2W2FTU5NUKn3+/LmLi0t6erryWvhVsGLFCj6fP3XqVDMzsy+++EJfX3/27Nn6+voSicTQ0PDOnTvDhw+Xy+UCgYBhmKqqqgsXLrz33ntz5swhIi4/iEgikRw/frygoEAgEAgEgoiIiLi4uCVLlrTal5mZ2cKFC7dt27Zx40YiMjExGTly5Pr168PDw8+cOZOfnz9x4sSO6nz8+PHo0aM7c0SWlpYXLly4dOlS978U6MCUKVP6toC2V8fNzc06Ojp9UowK48ePf+eddxITE7/88kvlwjNnzvj6+lpbWxPR8uXL582bt3r16pkzZ/7000/BwcG2trYTJkxQscFhw4adOnVq8eLF48aN8/T07OS1Ued5eXl5eHhkZ2ffunUrMTExNTX166+/trGxWbt27aBBg3R0dBoaGng8nkwmMzU1HTlyZGZmpoeHx+XLlydNmjRgwIB2j05PT+9lSupaGKxZs6aiosLX1zc4OFjF/xPfffdduyFhZmbGTZSXl+fk5Pj4+Cg/ev78uaOjY1NTU35+vlgsHj58uFAoVH769OnTlhu0sLD47bffOtr7unXrDhw4sHTpUj09valTp/r7+yvvMPqc8jpLW1tbeYBaWlpEpKmpeeLEiZSUFB0dndGjR2toaBCRSCSysrLimunp6XFj6Q8ePCAiV1dX5Wbb/baJKDw8fNSoUcrxsf3798+dO3fz5s1vvfXW0aNHW14KtVJdXa0cMVbNwMBAJpMZGRl1pjH0L2ZmZpWVldy0RCLh8/kvebpRn5CQkAULFnh4eCiXlJeX79u3j+s6JiJjY+Oamho9PT0fH5+IiIiwsDDVGzQzM1u0aJG/v//ly5ejo6O1tbXj4+N7qtp79+5pa2sPGzZs/Pjx48ePX7JkSURExP79+zdv3lxSUrJx48aamhoLCwtlJ9LkyZOvXLni4eHx448/rlixQsXRvUxVXQsDBweHtLS069evm5mZ2dvbd3Vnmpqa3ISRkZGPj8/8+fO52fz8fCsrKw0NjUmTJmVkZJSXl0+bNq3lioMGDfr111+Vs2Kx+O233+amlf0excXF9vb2CoXi6dOnYWFhWlpaBQUF69evHz169Pvvv9/VUntfWlrarl27rl+/bmlpSUTcf1ehUKgcIHn+/PnTp0+JSCAQ6OrqFhcXczc9YrG45UBxS8bGxkuXLt2yZQs3a21tfePGjfr6+gEDBvTCEUF/5+DgkJWV5enpSUS3bt3qxj/5XiMQCObPnx8XF6c8IRobG4eHh9vZ2RFRQ0NDYWGhqalpU1NTUlLS+++/v3fv3tjY2BdeJpaWlubl5dXV1U2ePLkHq83IyBCJRNwtOxHp6uqOGzcuJydHJpNFRkZu37793XffJaI1a9ZwDSZNmpSYmHjnzp2Ghgauc6Xdo3vJqro2ZuDr63vw4ME//vGPycnJAQEBaWlpz58/78Ze3dzcjh8//ssvv8hksiNHjqxZs4a7EJ48efJPP/3066+/trzsJaKJEyfev38/PT29vr4+IyPj5s2bXAM+n//DDz80NjZmZ2fn5OQQkYaGxqZNm06cONHQ0KClpdXQ0PDCByhfERKJRFdXd8CAAfX19Zs2baqvr6+rq5s9e/atW7fS0tKkUulXX33V3NzMMIy5ubmjo2NwcHBFRUVeXp6bm1vb0TOlVatWnTlzhohYln327NmzZ8/q6uq4iaampo7W4vP5nRxrkUqleNzrdRUQEJCUlHT//v0nT57ExcUFBgb2dUWq+Pj4NDY25uXlcbMffPBBQkLCo0ePfvvttx07dhw8eJCIDhw4MGTIkM2bN9fU1Cj/1TQ2NrbqEGtsbLxw4UJISEh0dPSoUaP279/fs8c+bdq0X375JTk5WSQSVVZW3rhx4/vvv3dzc6utrWVZVk9PT6FQZGZmcmd/IjI2Nh41alRMTIy7uzsXYO0e3Uvq8piBpqamq6urq6trUVHRyZMnGYb56KOPuroRBweHoKCguLi4iooKGxub7du3c10WdnZ2zc3Njo6OrS5dDQ0Nt23bFh8fv2fPHgsLiw0bNgwePJiIgoODDxw4MHPmzIEDB06dOpVrHBER8e233yYnJxsYGMyePfuFj8S8Inx9fdPT021sbIRC4YoVKz7//PMPP/zwwYMHx48fDwoKkkgka9euFQgEXIfMoUOHQkJCRowYoaOj4+/vr+Ke18DAIDQ09JtvvqmoqBAIBC0/On36dMvb6paEQqFIJOKuO1QTiUTm5ub95UuGLnF1dd2wYcP06dObmpoCAgL8/f37uiJVtLW1V6xYER4ezs36+fk1NDSsWrWqrq7u3Xff/fLLL+/du3f69OmUlBRNTc2wsLDw8HAnJyehUGhtbT1r1qxjx44pO06vXbuWlZW1fPnyESNGqKNUS0vL+Pj4xMTEsLCw2tpac3PzgIAANzc3Ilq4cGFERIS2tvakSZMWLlyYnJxsb29vbW09ZcqUnTt3uru7d3R0L18Vw7Jsy/lX8DcZ1O1V/nGb0tLSn3/+ed68eURUX19vaGhYVlamHHhQn4qKioyMDO4vy1Q7duyYi4sL/jAQ1KF/nY7w4zagRhoaGn/6059Onz4tlUo3bdrk7OzcC0lARKampubm5i/8K5jMzExzc3MkAcBrAGHwSrOwsDh06ND69evHjh1bUFCwf//+Xtu1s7OzTCZTkQeZmZkymQwdRACvB3QTvdLdRH0uMzOzvLzc1tbW0tKSe9hUKpWKRKJ79+5hqADUrX+djvp7NxHCAGHwAniFNfSV/nU66u9hgN8zgBcwNTXFeR/gtYcxAwAAQBgAAADCAAAACGEAAACEMAAAAGr7aCkAAPwPwp0BAAAgDAAAAGEAAACEMAAAAEIYAAAAIQwAAIAQBgAAQAgDAAAghAEAABDCAAAACGEAAABE9H8aYqxyBGezJAAAAABJRU5ErkJggg==" - } - }, "cell_type": "markdown", "metadata": {}, "source": [ - "![bbox_one.png](attachment:bbox_one.png)" + "Sometimes the classes are not defined yet or incomplete. To explore the input images without worring about any classes you can use the `NoOutput` option on the annotator factory which is done in the following:" ] }, { @@ -150,15 +182,24 @@ "metadata": {}, "outputs": [], "source": [ - "# todo: differs from im2im\n", - "# explorer._controller.images[:3]" + "unlabel_factory = Annotator(input_, NoOutput(), settings_)\n", + "unlabel_factory.explore()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "## CREATE" + "## Create\n", + "The **create** option allows users to manually create their annotated datasets. \n", + "\n", + "The bbox annotator allows users to create multiple bounding boxes on the image and associate labels to the bboxes. Additionally features are the following:\n", + "\n", + "- The lamp button can be used to highlight the annotation \n", + "- The coordinate inputs can be changed to improve the annotated bounding box\n", + "- The trash button can delete the annotation\n", + "\n", + "The next cell removes already created annotation files to create a new dataset." ] }, { @@ -172,33 +213,29 @@ ] }, { - "cell_type": "code", - "execution_count": null, + "cell_type": "markdown", "metadata": {}, - "outputs": [], "source": [ - "creator = anni.create()\n", + "The next cell initializes the **create** option. \n", "\n", - "creator" + "For this tutorial, a function was defined that imitates human work. You can choose between performing the annotation manually yourself or letting the function do the work for you. Use the mouse to draw on the canvas." ] }, { - "attachments": { - "bbox_two.png": { - "image/png": "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" - } - }, - "cell_type": "markdown", + "cell_type": "code", + "execution_count": null, "metadata": {}, + "outputs": [], "source": [ - "![bbox_two.png](attachment:bbox_two.png)" + "creator = anni.create()\n", + "creator" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "#### To imitate human work on the current step, let's randomly annotate all the images automatically:" + "The next cell imitates human work by randomly annotating all images automatically. If you want to manually annotate then skip the next step." ] }, { @@ -207,8 +244,6 @@ "metadata": {}, "outputs": [], "source": [ - "#SKIP THIS STEP IF YOU ANNOTATE MANUALLY\n", - "\n", "HELPER = Tutorial(dataset, settings_.project_path)\n", "HELPER.add_random_bboxes(creator)" ] @@ -217,14 +252,18 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "## IMPROVE" + "## Improve" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "On this step we explore annotated images, selecting all __missed__ or __incorrect__ bounding boxes." + "The **improve** feature allows users to refine the annotated dataset. \n", + "\n", + "As before, for the purpose of the tutorial, a function can be used to performe the annotation and you don't have to annotate manually. If you want to annotate manually then make sure selecting all __missed__ or __incorrect__ bounding boxes.\n", + "\n", + "If you chose to annotate manually don't forget to click the __SAVE__ button when finished with each class." ] }, { @@ -256,15 +295,10 @@ ] }, { - "attachments": { - "bbox_three.png": { - "image/png": "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" - } - }, "cell_type": "markdown", "metadata": {}, "source": [ - "![bbox_three.png](attachment:bbox_three.png)" + "The next cell imitates the human work. If you chose to annotate manually make sure to skip this cell." ] }, { @@ -273,8 +307,6 @@ "metadata": {}, "outputs": [], "source": [ - "#SKIP THIS STEP IF YOU FIX MANUALLY\n", - "\n", "HELPER.fix_incorrect_bboxes(improver, creator)" ] }, @@ -282,7 +314,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Get the list of images with incorrect bboxes" + "Now we obtain a list of all images with incorrect bboxes:" ] }, { diff --git a/nbs/01d_tutorial_video_annotator.ipynb b/nbs/01d_tutorial_video_annotator.ipynb index b012b8e..d0b3cee 100644 --- a/nbs/01d_tutorial_video_annotator.ipynb +++ b/nbs/01d_tutorial_video_annotator.ipynb @@ -31,6 +31,7 @@ "outputs": [], "source": [ "# hide\n", + "import json\n", "from ipyannotator.annotator import Annotator\n", "from ipyannotator.base import Settings\n", "from ipyannotator.mltypes import InputImage, OutputVideoBbox, NoOutput\n", @@ -43,11 +44,21 @@ "id": "53d61168", "metadata": {}, "source": [ - "# Tutorial: Video Annotator\n", + "# Video Annotator - Tracking objects through video frames\n", "\n", - "The current notebook will demonstrate how to use Ipyannotator to explore, create and improve video annotation.\n", + "The current tutorial illustrates how to use Ipyannotator to classify video data.\n", "\n", - "It's used an artifical video dataset that follows [MOT data format](https://github.com/JonathonLuiten/TrackEval/blob/master/docs/MOTChallenge-Official/Readme.md#data-format)." + "The task of identifying objects in a video frame is called video classification. \n", + "\n", + "**Ipyannotator** allows users to explore an entire set of video frames and specific labels; manually **create** their datasets drawing bounding boxes and associating labels across the frames; **improve** existing annotations.\n", + "\n", + "This tutorial is divided in the following steps:\n", + "\n", + "- [Select dataset](#Select-dataset)\n", + "- [Setup annotator](#Setup-annotator)\n", + "- [Explore](#Explore)\n", + "- [Create](#Create)\n", + "- [Improve](#Improve)" ] }, { @@ -55,7 +66,9 @@ "id": "55531473", "metadata": {}, "source": [ - "## Select dataset" + "## Select dataset\n", + "\n", + "This tutorial uses a minimal artificial video dataset generated by Ipyannotator. The dataset follows [MOT data format](https://github.com/JonathonLuiten/TrackEval/blob/master/docs/MOTChallenge-Official/Readme.md#data-format). It contains 20 images with 2 classes (`rectangle` and `circle`) and doesn't need to be downloaded." ] }, { @@ -75,7 +88,10 @@ "source": [ "## Setup annotator\n", "\n", - "This section will instantiate the BBoxVideoAnnotator using Ipyannotator's input/output factory. This annotator uses a image as input and output's a bbox video UI that handles labelling (in this case with the classes `circle` and `rectangle`)." + "\n", + "This section will set up the paths and the input/output pair needed to classify the images.\n", + "\n", + "The following cell will import the project file and directory where the images were generated. For this tutorial we simplify the process using the `get_settings` function instead of hardcoding the paths." ] }, { @@ -89,6 +105,18 @@ "settings_.project_file, settings_.image_dir" ] }, + { + "cell_type": "markdown", + "id": "5c822ce9", + "metadata": {}, + "source": [ + "Ipyannotator uses pairs of input/output data to set up the annotation. \n", + "\n", + "The video image classification annotator uses `InputImage` and `OutputVideoBox`as the pair to set up the annotator.\n", + "\n", + "The `InputImage` function provides information about the directory that contains the images to be classified, and the images itself. The `OutputImageBox` function provides information about the directory that contains the classes that can be associated with the images." + ] + }, { "cell_type": "code", "execution_count": null, @@ -105,6 +133,16 @@ "input_.dir" ] }, + { + "cell_type": "markdown", + "id": "1f5a4ee8", + "metadata": {}, + "source": [ + "The final part in setting up the Ipyannotator is the configuration of the `Annotator` factory with the pair of input/output data. \n", + "\n", + "The factory allows three types of annotator tools: explore, create, improve. The next sections will guide you through every step." + ] + }, { "cell_type": "code", "execution_count": null, @@ -121,8 +159,9 @@ "metadata": {}, "source": [ "## Explore\n", + "The **explore** option allows users to navigate across the images in the dataset using `next/previous` buttons. This function is used for data visualization only, improvement and additional labeling is done in the next steps. \n", "\n", - "Navigate the images generated by the artificial dataset." + "When exploring the artificial dataset used in this tutorial you will see a red circle and a gray rectangle as the objects to be tracked. The black square represents an occlusion on the objects and is used to illustrate how the **improve** step works." ] }, { @@ -142,31 +181,15 @@ "metadata": {}, "source": [ "## Create\n", + "The **create** option allows users to manually create their annotated datasets. Please be aware that\n", "\n", - "Annotate and label every object in the image. Ipyannotator generates the objects created using indexed labels that starts from 0.\n", - "\n", - "All data is stored as json in the following format:\n", - "\n", - "```json\n", - "{\n", - " '../path/to/image1': {\n", - " 'bbox': [\n", - " {'x': 1, 'y:' 1, 'width': 1, 'height': 1, 'id': '0'},\n", - " {'x': 2, 'y:' 2, 'width': 2, 'height': 2, 'id': '1'},\n", - " ], \n", - " 'labels': [['Label A'], ['Label B']],\n", - " },\n", - " '../path/to/image2': {\n", - " 'bbox': [\n", - " {'x': 1, 'y:' 1, 'width': 1, 'height': 1, 'id': '0'},\n", - " ], \n", - " 'labels': [['Label B']],\n", - " },\n", - " ...\n", - "}\n", + "```{warning}\n", + "The video annotator create option is a beta version\n", "```\n", "\n", - "Every frame has its annotations mapped by the path of the image. Then every bounding box draw in the annotators has it's `x`, `y`, `width`, `height`, `id` properties (as part of the `bbox`) and a label (mapped as `labels`, but with the same index as the object mapped in the `bbox`)." + "Currently, video annotation allows users to draw multiple bounding boxes in every frame and associate a label to every annotated object bounding box. Ipyannotator generates the objects creating indexed labels that start from 0.\n", + "\n", + "The next cell removes already created annotation files to create a new dataset." ] }, { @@ -184,7 +207,9 @@ "id": "ae6b2f91", "metadata": {}, "source": [ - "**To imitate human work on the current step, let's randomly annotate all the images automatically:**" + "The next cell initializes the **create** option. \n", + "\n", + "For this tutorial, a function was defined that imitates human work, annotating the images automatically." ] }, { @@ -195,7 +220,16 @@ "outputs": [], "source": [ "anni.output_item = output_\n", - "creator = anni.create()" + "creator = anni.create()\n", + "creator" + ] + }, + { + "cell_type": "markdown", + "id": "bd4bd327-6a32-4793-bd0f-f3ab6e56022a", + "metadata": {}, + "source": [ + "The next cell imitate human work annotating all images automatically." ] }, { @@ -205,13 +239,35 @@ "metadata": {}, "outputs": [], "source": [ - "#SKIP THIS STEP IF YOU ANNOTATE MANUALLY\n", - "\n", - "# Argument 1 to \"Tutorial\" has incompatible type\n", - "# \"ipyannotator.datasets.factory.DS\"; expected\n", - "# \"ipyannotator.datasets.factory_legacy.DS\"\n", - "HELPER = Tutorial(dataset, settings_.project_path) # type: ignore\n", - "HELPER.annotate_video_bboxes(creator)" + "HELPER = Tutorial(dataset, settings_.project_path)\n", + "annotations = HELPER.annotate_video_bboxes(creator)" + ] + }, + { + "cell_type": "markdown", + "id": "051fd83f", + "metadata": {}, + "source": [ + "All data is stored in a file formatted as JSON in the following format:" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "8f4e6e83", + "metadata": {}, + "outputs": [], + "source": [ + "data_format = {k: v for i, (k, v) in enumerate(annotations.items()) if i == 0}\n", + "print(json.dumps(data_format, indent=2))" + ] + }, + { + "cell_type": "markdown", + "id": "4ce2a176", + "metadata": {}, + "source": [ + "Note that in the JSON file above the annotations of each frame is mapped by the path of the image. Every bounding box drawn in the annotators has the properties: `x`, `y`, `width`, `height`, `id` as part of the `bbox` field. The annotation labels are mapped in the `labels` field in the JSON file. Every index of the `labels` array corresponds to the object mapped in the `bbox` property." ] }, { @@ -221,12 +277,22 @@ "source": [ "## Improve\n", "\n", - "Ipyannotator's video annotation tool allows users to:\n", + "The **improve** feature in the Ipyannotator video annotation allows users to refine the annotated dataset. This includes:\n", "\n", "- Select objects across the frames and join the trajectories drawn.\n", - "- Update labels the labels across the whole annotation.\n", + "- Update labels across the entire annotation.\n", + "\n", + "In the example below we have an occlusion illustrated by a black square. The rectangle disappears behind the occluding object and appears again with a new object id. The video annotator allows users to join the trajectories of different objects into a new object. \n", + "\n", + "Joining trajectory:\n", "\n", - "In the example below we have an occlusion colored as black. The rectangle dissapear after the occlusion and appears again with a new object id. The video annotator allow users to join the trajectories of different objects into a new and single one." + "- Navigate across the annotator\n", + "- Note that the gray rectangle disappears\n", + "- Note that the gray rectangle reappears but with a new id\n", + "- Select the rectangle with a new id (marking the checkbox)\n", + "- Navigate back until you see the old gray rectangle id\n", + "- Select the rectangle with the old id (marking the checkbox)\n", + "- Click on the join button\n" ] }, { @@ -243,7 +309,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/02_navi_widget.ipynb b/nbs/02_navi_widget.ipynb index 58a3a4f..dc195fc 100644 --- a/nbs/02_navi_widget.ipynb +++ b/nbs/02_navi_widget.ipynb @@ -274,7 +274,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/02a_right_menu_widget.ipynb b/nbs/02a_right_menu_widget.ipynb index 5420890..d8c8fd3 100644 --- a/nbs/02a_right_menu_widget.ipynb +++ b/nbs/02a_right_menu_widget.ipynb @@ -62,6 +62,14 @@ "ipytest.autoconfig(raise_on_error=True)" ] }, + { + "cell_type": "markdown", + "id": "c9197437", + "metadata": {}, + "source": [ + "# Right menu Widget" + ] + }, { "cell_type": "code", "execution_count": null, @@ -79,28 +87,32 @@ " bbox_coord: BboxCoordinate,\n", " max_coord_input_values: Optional[BboxCoordinate],\n", " index: int,\n", - " options: List[str] = None\n", + " options: List[str] = None,\n", + " readonly: bool = False\n", " ):\n", " super().__init__()\n", "\n", + " self.readonly = readonly\n", " self.bbox_coord = bbox_coord\n", " self.index = index\n", " self._max_coord_input_values = max_coord_input_values\n", " self.layout = Layout(display='flex', overflow='hidden')\n", - " self.btn_delete = self._btn_delete(index)\n", " self.dropdown_classes = self._dropdown_classes(options)\n", " self.btn_select = self._btn_select(index)\n", " self.input_coordinates = self._coordinate_inputs(bbox_coord)\n", "\n", - " self.children = [\n", - " HBox([\n", - " self.btn_select,\n", - " self.dropdown_classes,\n", - " self.input_coordinates,\n", - " self.btn_delete\n", - " ])\n", + " elements = [\n", + " self.btn_select,\n", + " self.dropdown_classes,\n", + " self.input_coordinates,\n", " ]\n", "\n", + " if not self.readonly:\n", + " self.btn_delete = self._btn_delete(index)\n", + " elements.append(self.btn_delete)\n", + "\n", + " self.children = [HBox(elements)]\n", + "\n", " def _btn_delete(self, index: int) -> ActionButton:\n", " return ActionButton(\n", " layout=Layout(width='auto'),\n", @@ -113,7 +125,8 @@ " return Dropdown(\n", " layout=Layout(width='auto'),\n", " options=options,\n", - " value=value\n", + " value=value,\n", + " disabled=self.readonly\n", " )\n", "\n", " def _btn_select(self, index: int) -> ActionButton:\n", @@ -126,7 +139,8 @@ " def _coordinate_inputs(self, bbox_coord: BboxCoordinate):\n", " return CoordinateInput(\n", " bbox_coord=bbox_coord,\n", - " input_max=self._max_coord_input_values\n", + " input_max=self._max_coord_input_values,\n", + " disabled=self.readonly\n", " )" ] }, @@ -165,10 +179,11 @@ " label: List[str],\n", " options: List[str],\n", " selected: bool = False,\n", - " btn_delete_enabled: bool = True\n", + " btn_delete_enabled: bool = True,\n", + " readonly: bool = False\n", " ):\n", " super(VBox, self).__init__() # type: ignore\n", - "\n", + " self.readonly = readonly\n", " self.selected = selected\n", " self.bbox_video_coord = bbox_video_coord\n", " self.object_checkbox = self._object_checkbox()\n", @@ -238,7 +253,8 @@ " on_coords_changed: Optional[Callable],\n", " on_label_changed: Callable,\n", " on_btn_delete_clicked: Callable,\n", - " on_btn_select_clicked: Optional[Callable]\n", + " on_btn_select_clicked: Optional[Callable],\n", + " readonly: bool = False\n", " ):\n", " super().__init__()\n", " self._classes = classes\n", @@ -247,6 +263,7 @@ " self._on_btn_delete_clicked = on_btn_delete_clicked\n", " self._on_label_changed = on_label_changed\n", " self._on_btn_select_clicked = on_btn_select_clicked\n", + " self.readonly = readonly\n", "\n", " @property\n", " def max_coord_input_values(self) -> Optional[BboxCoordinate]:\n", @@ -268,9 +285,11 @@ " options=self._classes,\n", " bbox_coord=coord,\n", " max_coord_input_values=self._max_coord_input_values,\n", + " readonly=self.readonly\n", " )\n", "\n", - " bbox_item.btn_delete.on_click(self.del_element)\n", + " if not self.readonly:\n", + " bbox_item.btn_delete.on_click(self.del_element)\n", " bbox_item.input_coordinates.uuid = index\n", " bbox_item.input_coordinates.coord_changed = self._on_coords_changed\n", " bbox_item.btn_select.on_click(self._on_btn_select_clicked)\n", @@ -285,6 +304,9 @@ "\n", " self.children = [*list(self.children), *elements] # type: ignore\n", "\n", + " def __getitem__(self, index: int):\n", + " return self.children[index]\n", + "\n", " def clear(self):\n", " self.children = []\n", "\n", @@ -321,10 +343,6 @@ "outputs": [], "source": [ "#hide\n", - "\n", - "from typing import Any\n", - "\n", - "\n", "def f(x):\n", " return x\n", "\n", @@ -359,7 +377,7 @@ " classes: list,\n", " on_label_changed: Callable,\n", " on_btn_delete_clicked: Callable,\n", - " on_btn_select_clicked: Callable,\n", + " on_btn_select_clicked: Optional[Callable],\n", " on_checkbox_object_clicked: Callable,\n", " btn_delete_enabled: bool = True\n", " ):\n", @@ -375,9 +393,6 @@ " self._btn_delete_enabled = btn_delete_enabled\n", " self._on_checkbox_object_clicked = on_checkbox_object_clicked\n", "\n", - " def __getitem__(self, index: int):\n", - " return self.children[index]\n", - "\n", " # error: Signature of \"render_btn_list\" incompatible with supertype \"BBoxList\"\n", " def render_btn_list( # type: ignore\n", " self,\n", @@ -476,8 +491,6 @@ "source": [ "@pytest.fixture\n", "def bbox_video_list_fixture():\n", - " lambda x: x\n", - "\n", " return BBoxVideoList(['A', 'B'], f, f, f, f)" ] }, @@ -489,7 +502,6 @@ "outputs": [], "source": [ "# hide\n", - "\n", "def list_to_bbox_item(bboxes: list) -> List[BBoxVideoItem]:\n", " result = []\n", " for i, bbox in enumerate(bboxes):\n", @@ -628,6 +640,55 @@ " assert len(bbox_video_list_fixture.elements) == 3" ] }, + { + "cell_type": "code", + "execution_count": null, + "id": "f37f7441", + "metadata": {}, + "outputs": [], + "source": [ + "@pytest.fixture\n", + "def readonly_fixture() -> BBoxList:\n", + " bbox_list = BBoxList(['A', 'B'], BboxCoordinate(*[5, 5, 5, 10]), f, f, f, f, readonly=True)\n", + "\n", + " classes: List[List[str]] = [[], [], []]\n", + " bbox_dict = [\n", + " {'x': 10, 'y': 10, 'width': 20, 'height': 30},\n", + " {'x': 20, 'y': 30, 'width': 10, 'height': 10},\n", + " {'x': 30, 'y': 30, 'width': 10, 'height': 10}\n", + " ]\n", + "\n", + " bbox = [BboxCoordinate(**b) for b in bbox_dict]\n", + "\n", + " bbox_list.render_btn_list(bbox, classes)\n", + "\n", + " return bbox_list" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "73aff8a4", + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_doesnt_render_btn_delete_if_readonly(readonly_fixture):\n", + " assert hasattr(readonly_fixture[0], 'btn_delete') is False" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "6c846b9c", + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_cant_change_input_if_readonly(readonly_fixture):\n", + " assert readonly_fixture[0].dropdown_classes.disabled is True" + ] + }, { "cell_type": "code", "execution_count": null, diff --git a/nbs/02b_grid_menu.ipynb b/nbs/02b_grid_menu.ipynb new file mode 100644 index 0000000..49122ea --- /dev/null +++ b/nbs/02b_grid_menu.ipynb @@ -0,0 +1,439 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": null, + "id": "ef1feeff", + "metadata": {}, + "outputs": [], + "source": [ + "# default_exp custom_widgets.grid_menu" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "8ed8645d", + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "%load_ext autoreload\n", + "%autoreload 2" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "cc299383", + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "from nbdev import *" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "1d7bd480", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "from math import ceil\n", + "from functools import partial\n", + "from typing import Callable, Iterable, Optional, Tuple\n", + "import warnings\n", + "import attr\n", + "from ipywidgets import GridBox, Output, Layout" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "6ac14bff", + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "import pytest\n", + "import ipytest\n", + "ipytest.autoconfig(raise_on_error=True)" + ] + }, + { + "cell_type": "markdown", + "id": "4def5438", + "metadata": {}, + "source": [ + "## Grid Menu\n", + "\n", + "The current notebook develop a grid menu widget that allows clickable widgets to be displayed as grid. The next cell will design the `Grid` class that contain the settings for the `GridMenu` component." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "de01fde1", + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "@attr.define(slots=False)\n", + "class Grid:\n", + " width: int\n", + " height: int\n", + " n_rows: Optional[int] = 3\n", + " n_cols: Optional[int] = 3\n", + " disp_number: int = 9\n", + " display_label: bool = False\n", + "\n", + " @property\n", + " def num_items(self) -> int:\n", + " row, col = self.area_adjusted(self.disp_number)\n", + " return row * col\n", + "\n", + " def area_adjusted(self, n_total: int) -> Tuple[int, int]:\n", + " \"\"\"Returns the row and col automatic arranged\"\"\"\n", + " if self.n_cols is None:\n", + " if self.n_rows is None: # automatic arrange\n", + " label_cols = 3\n", + " label_rows = ceil(n_total / label_cols)\n", + " else: # calc cols to show all labels\n", + " label_rows = self.n_rows\n", + " label_cols = ceil(n_total / label_rows)\n", + " else:\n", + " if self.n_rows is None: # calc rows to show all labels\n", + " label_cols = self.n_cols\n", + " label_rows = ceil(n_total / label_cols)\n", + " else: # user defined\n", + " label_cols = self.n_cols\n", + " label_rows = self.n_rows\n", + "\n", + " return label_rows, label_cols" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "8060b3de", + "metadata": {}, + "outputs": [], + "source": [ + "@pytest.fixture\n", + "def grid_fixture() -> Grid:\n", + " return Grid(width=300, height=300)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "2e7e4f3b", + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_return_num_items(grid_fixture):\n", + " assert grid_fixture.num_items == 9" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "075e7b8c", + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_adjusts_area_missing_args(grid_fixture):\n", + " grid_fixture.n_rows = None\n", + " assert grid_fixture.area_adjusted(12) == (4, 3)" + ] + }, + { + "cell_type": "markdown", + "id": "ec5b5eb9", + "metadata": {}, + "source": [ + "The `GridMenu` doesn't have a `on_click` event listener, but grid elements itself should implement `on_click(ev)`, `reset_callbacks()` and `update(other: SameWidgetType)` methods to register/reset onclick callback function and update its internal values, respectively. Also grid element shoudl have a field name in order user can destinguish between grid children." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "0e7a34da", + "metadata": {}, + "outputs": [], + "source": [ + "#export\n", + "class GridMenu(GridBox):\n", + " debug_output = Output(layout={'border': '1px solid black'})\n", + "\n", + " def __init__(\n", + " self,\n", + " grid: Grid,\n", + " widgets: Optional[Iterable] = None,\n", + " ):\n", + " self.callback = None\n", + " self.gap = 40 if grid.display_label else 15\n", + " self.grid = grid\n", + "\n", + " n_row, n_col = grid.area_adjusted(grid.disp_number)\n", + " column = grid.width + self.gap\n", + " row = grid.height + self.gap\n", + " centered_settings = {\n", + " 'grid_template_columns': \" \".join([f'{(column)}px' for _\n", + " in range(n_col)]),\n", + " 'grid_template_rows': \" \".join([f'{row}px' for _\n", + " in range(n_row)]),\n", + " 'justify_content': 'center',\n", + " 'align_content': 'space-around'\n", + " }\n", + "\n", + " super().__init__(\n", + " layout=Layout(**centered_settings)\n", + " )\n", + "\n", + " if widgets:\n", + " self.load(widgets)\n", + " self.widgets = widgets\n", + "\n", + " def _fill_widgets(self, widgets: Iterable):\n", + " if self.widgets is None:\n", + " self.widgets = widgets\n", + "\n", + " self.children = self.widgets\n", + "\n", + " if self.callback:\n", + " self.register_on_click()\n", + " else:\n", + " iter_state = iter(widgets)\n", + "\n", + " for widget in self.widgets:\n", + " i_widget = next(iter_state, None)\n", + " if i_widget:\n", + " widget.update(i_widget)\n", + " else:\n", + " widget.clear()\n", + "\n", + " def _filter_widgets(self, widgets: Iterable) -> Iterable:\n", + " \"\"\"Limit the number of widgets to be rendered\n", + " according to the grid's area\"\"\"\n", + " widgets_list = list(widgets) # Iterable don't have len()\n", + " num_widgets = len(widgets_list)\n", + " row, col = self.grid.area_adjusted(num_widgets)\n", + " num_items = row * col\n", + "\n", + " if num_widgets > num_items:\n", + " warnings.warn(\"!! Not all labels shown. Check n_cols, n_rows args !!\")\n", + " return widgets_list[:num_items]\n", + "\n", + " return widgets\n", + "\n", + " @debug_output.capture(clear_output=False)\n", + " def load(self, widgets: Iterable, callback: Optional[Callable] = None):\n", + " widgets_filtered = self._filter_widgets(widgets)\n", + " self._fill_widgets(widgets_filtered)\n", + "\n", + " if callback:\n", + " self.on_click(callback)\n", + "\n", + " @debug_output.capture(clear_output=False)\n", + " def on_click(self, callback: Callable):\n", + " setattr(self, 'callback', callback)\n", + " self.register_on_click()\n", + "\n", + " @debug_output.capture(clear_output=False)\n", + " def register_on_click(self):\n", + " if self.widgets:\n", + " for widget in self.widgets:\n", + " widget.reset_callbacks()\n", + "\n", + " widget.on_click(\n", + " partial(\n", + " self.callback,\n", + " value=widget.value\n", + " )\n", + " )\n", + "\n", + " def clear(self):\n", + " self.widgets = None\n", + " self.children = tuple()" + ] + }, + { + "cell_type": "markdown", + "id": "7bd92bce", + "metadata": {}, + "source": [ + "We now can instantiate the grid menu and load widgets on it. For this example we're using the custom widget `ImageButton` to be displayed using the load function. " + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "bf7b8b0b", + "metadata": {}, + "outputs": [], + "source": [ + "from ipyannotator.custom_input.buttons import ImageButton, ImageButtonSetting\n", + "from ipywidgets import HTML\n", + "from IPython.display import display" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "23cee979", + "metadata": {}, + "outputs": [], + "source": [ + "grid = Grid(width=50, height=75, n_cols=2, n_rows=2)\n", + "grid_menu = GridMenu(grid)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "521c486f", + "metadata": {}, + "outputs": [], + "source": [ + "widgets = []\n", + "setting = ImageButtonSetting(im_path='../data/projects/capture1/pics/pink25x25.png')\n", + "for i in range(4):\n", + " widgets.append(ImageButton(setting))\n", + "grid_menu.load(widgets)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "389902c1", + "metadata": {}, + "outputs": [], + "source": [ + "grid_menu" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "04f89101", + "metadata": {}, + "outputs": [], + "source": [ + "widgets = []\n", + "setting = ImageButtonSetting(im_path='../data/projects/capture1/pics/teal50x50_5.png')\n", + "for i in range(2):\n", + " widgets.append(ImageButton(setting))\n", + "grid_menu.load(widgets)" + ] + }, + { + "cell_type": "markdown", + "id": "9562916f", + "metadata": {}, + "source": [ + "\n", + "While ipyevents implementation lacks `sender` or `source` in callback args, `functools.partial` used to back element `name` into return value. You can see example of on_click event handler `test_handler` below. \n", + "name of the button is printed out on click." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "3c34d5db", + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "h = HTML('Event info')\n", + "display(h)\n", + "\n", + "\n", + "def test_handler(event, value=None):\n", + " event.update({'label_name': value})\n", + " h.value = event['label_name']\n", + "\n", + "\n", + "grid_menu.on_click(test_handler)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "ebd98ab4", + "metadata": {}, + "outputs": [], + "source": [ + "#hide\n", + "from ipyannotator.custom_input.buttons import ActionButton" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "35a4ee8e", + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_doesnt_load_more_widgets_than_the_grid_area():\n", + " with warnings.catch_warnings(record=True) as w:\n", + " grid = Grid(width=50, height=75, n_cols=1, n_rows=1)\n", + " grid_menu = GridMenu(grid)\n", + " widgets = [ActionButton() for _ in range(2)]\n", + " grid_menu.load(widgets)\n", + " assert len(grid_menu.widgets) == 1\n", + " assert bool(w) is True" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "c1fdd429", + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_doesnt_throw_warning_if_number_of_widgets_is_small():\n", + " with warnings.catch_warnings(record=True) as w:\n", + " grid = Grid(width=100, height=100, n_rows=2, n_cols=2)\n", + " grid_menu = GridMenu(grid)\n", + " grid_menu._filter_widgets([1])\n", + " assert bool(w) is False" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "a935f0d4", + "metadata": {}, + "outputs": [], + "source": [ + "#hide\n", + "from nbdev.export import notebook2script\n", + "notebook2script()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "25e3af36", + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3", + "language": "python", + "name": "python3" + } + }, + "nbformat": 4, + "nbformat_minor": 5 +} diff --git a/nbs/03_storage.ipynb b/nbs/03_storage.ipynb index a334d76..35d391e 100644 --- a/nbs/03_storage.ipynb +++ b/nbs/03_storage.ipynb @@ -45,10 +45,11 @@ "outputs": [], "source": [ "#exporti\n", - "\n", + "import warnings\n", "import copy\n", "import json\n", "import os\n", + "from typing import List, Union, Iterable\n", "from collections import defaultdict\n", "from collections.abc import MutableMapping\n", "from pathlib import Path" @@ -91,16 +92,15 @@ " if file_name is not None:\n", " annotation_file_path = Path(file_name)\n", " results_dir = annotation_file_path.parent\n", - " print(f\"WARNING: `results_dir` is deduced from `file_name` path: {results_dir}\")\n", " elif project_path is not None:\n", " results_dir = Path(\n", " project_path, 'results') if results_dir is None else Path(project_path, results_dir)\n", "\n", " annotation_file_path = Path(results_dir, 'annotations.json')\n", " if annotation_file_path.is_file():\n", - " raise ValueError(f\"Error: Annotations file already exists in {results_dir}!\"\n", - " \"\\n If you want to create annotations from scratch\"\n", - " \" - use empty dir!\")\n", + " warnings.warn(f\"Error: Annotations file already exists in {results_dir}!\"\n", + " \"\\n If you want to create annotations from scratch\"\n", + " \" - use empty dir!\")\n", " else:\n", " raise ValueError(\"You must define `project_path` or `file_name`!\")\n", "\n", @@ -202,7 +202,7 @@ "import glob\n", "\n", "\n", - "def get_image_list_from_folder(image_dir, strip_path=False):\n", + "def get_image_list_from_folder(image_dir) -> Iterable[Path]:\n", " ''' Scans to construct list of existing images as objects\n", " '''\n", " # if no files in `image_dir` assume all images are under `class_name` directories\n", @@ -212,9 +212,11 @@ " path_list = [Path(image_dir, f) for f in os.listdir(image_dir) if\n", " os.path.isfile(os.path.join(image_dir, f))]\n", "\n", - " if strip_path:\n", - " path_list = [p.name for p in path_list]\n", - " return path_list" + " return path_list\n", + "\n", + "\n", + "def strip_path(paths: Iterable[Path]) -> Iterable[str]:\n", + " return [p.name for p in paths]" ] }, { @@ -234,7 +236,7 @@ "outputs": [], "source": [ "# hide\n", - "get_image_list_from_folder('../data/mock/pics', strip_path=True)" + "strip_path(get_image_list_from_folder('../data/mock/pics'))" ] }, { @@ -481,46 +483,45 @@ "outputs": [], "source": [ "#exporti\n", - "\n", "from ipyannotator.helpers import flatten, reconstruct_class_images\n", - "import warnings\n", "\n", "\n", "class JsonLabelStorage(AnnotationStorage):\n", - " def __init__(self, im_dir: Path, label_dir: Path, annotation_file_path):\n", + " def __init__(self, im_dir: Path, label_dir: Union[Iterable[str], Path], annotation_file_path):\n", " self.annotation_file_path = annotation_file_path\n", " self.label_dir = label_dir\n", "\n", " self.has_annotation_file = True if (annotation_file_path is not None and\n", " annotation_file_path.is_file()) else False\n", "\n", - " print(f'has anno file: {self.has_annotation_file}')\n", " self.images = get_image_list_from_folder(im_dir)\n", "\n", - " # artificialy generate labels if no class images given (TODO: temorary workaround)\n", - " if 'class_autogenerated_' in str(label_dir):\n", - " print(f'autotgenerated: {label_dir}')\n", - " label_dir.mkdir(parents=True, exist_ok=True)\n", + " if isinstance(label_dir, Path):\n", + " # artificialy generate labels if no class images given (TODO: temorary workaround)\n", + " if 'class_autogenerated_' in str(label_dir):\n", + " label_dir.mkdir(parents=True, exist_ok=True)\n", "\n", - " if self.has_annotation_file:\n", - " print('reconstruct: FROM annotation file')\n", - " reconstruct_class_images(label_dir, annotation_file_path, lbl_w=50, lbl_h=50)\n", - " else:\n", - " warnings.warn(\"Annotation file should be provided\"\n", - " \" to generate labels automatically!\")\n", + " if self.has_annotation_file:\n", + " reconstruct_class_images(label_dir, annotation_file_path, lbl_w=50, lbl_h=50)\n", + " else:\n", + " warnings.warn(\"Annotation file should be provided\"\n", + " \" to generate labels automatically!\")\n", "\n", - " self.labels = get_image_list_from_folder(label_dir, strip_path=True)\n", + " self.labels = strip_path(get_image_list_from_folder(label_dir))\n", + " elif isinstance(label_dir, Iterable):\n", + " self.labels = label_dir\n", + " else:\n", + " raise ValueError(\"label_dir should have str or Path type\")\n", "\n", " if self.has_annotation_file: # init from json\n", - " print('load')\n", " self.load()\n", " else: # init storage from folder\n", - " print('save')\n", " super().__init__(self.images)\n", " self.save()\n", "\n", " def get_im_names(self, filter_files=None):\n", - " images = sorted(k for k in self.images if str(k) in self.keys())\n", + " keys = self.keys()\n", + " images = sorted([k for k in self.images if str(k) in keys])\n", "\n", " if not images:\n", " raise UserWarning(\"!! No Images to dipslay !!\")\n", @@ -532,14 +533,17 @@ " raise UserWarning(\"!! No image files to display. Check filter !!\")\n", " return images\n", "\n", - " def get_labels(self):\n", + " def get_labels(self) -> List[Union[Path, str]]:\n", " if not self.labels:\n", " warnings.warn(\"!! No labels to display !!\")\n", " return []\n", - " if self.has_annotation_file:\n", - " return sorted(v for v in self.labels if str(v) in set(flatten(self.values())))\n", - " else: # create mod -> display all labels from folder, not json\n", - " return sorted(self.labels)\n", + "\n", + " if self.has_annotation_file and isinstance(self.label_dir, Path):\n", + " values = set(flatten(self.values()))\n", + " return sorted([v for v in self.labels if str(v) in values])\n", + "\n", + " # create mod -> display all labels from folder, not json\n", + " return sorted(self.labels)\n", "\n", " def save(self):\n", " super().save(self.annotation_file_path)\n", @@ -1462,7 +1466,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/04_bbox_annotator.ipynb b/nbs/04_bbox_annotator.ipynb index dd8c64f..bc9ba12 100644 --- a/nbs/04_bbox_annotator.ipynb +++ b/nbs/04_bbox_annotator.ipynb @@ -61,7 +61,8 @@ "from ipywidgets import AppLayout, Button, HBox, VBox, Layout\n", "\n", "from ipyannotator.mltypes import BboxCoordinate\n", - "from ipyannotator.base import BaseState, AppWidgetState\n", + "from ipyannotator.base import BaseState, AppWidgetState, Annotator\n", + "from ipyannotator.mltypes import InputImage, OutputImageBbox\n", "from ipyannotator.bbox_canvas import BBoxCanvas, BBoxCanvasState\n", "from ipyannotator.navi_widget import Navi\n", "from ipyannotator.right_menu_widget import BBoxList, BBoxVideoItem\n", @@ -97,7 +98,8 @@ " coords: Optional[List[BboxCoordinate]]\n", " image: Optional[Path]\n", " classes: List[str]\n", - " labels: List[List[str]] = []" + " labels: List[List[str]] = []\n", + " drawing_enabled: bool = True" ] }, { @@ -130,15 +132,9 @@ " self._app_state = app_state\n", " self._bbox_state = bbox_state\n", " self._bbox_canvas_state = bbox_canvas_state\n", + " self.on_btn_select_clicked = on_btn_select_clicked\n", "\n", - " self._bbox_list = BBoxList(\n", - " max_coord_input_values=None,\n", - " on_coords_changed=self.on_coords_change,\n", - " on_label_changed=self.on_label_change,\n", - " on_btn_delete_clicked=self.on_btn_delete_clicked,\n", - " on_btn_select_clicked=on_btn_select_clicked,\n", - " classes=bbox_state.classes\n", - " )\n", + " self._init_bbox_list(self._bbox_state.drawing_enabled)\n", "\n", " if self._bbox_canvas_state.bbox_coords:\n", " self._bbox_list.render_btn_list(\n", @@ -147,6 +143,7 @@ " )\n", "\n", " app_state.subscribe(self._refresh_children, 'index')\n", + " bbox_state.subscribe(self._init_bbox_list, 'drawing_enabled')\n", " bbox_canvas_state.subscribe(self._sync_labels, 'bbox_coords')\n", " self._bbox_canvas_state.subscribe(self._update_max_coord_input, 'image_scale')\n", " self._update_max_coord_input(self._bbox_canvas_state.image_scale)\n", @@ -156,6 +153,19 @@ " display='block'\n", " )\n", "\n", + " def _init_bbox_list(self, drawing_enabled: bool):\n", + " self._bbox_list = BBoxList(\n", + " max_coord_input_values=None,\n", + " on_coords_changed=self.on_coords_change,\n", + " on_label_changed=self.on_label_change,\n", + " on_btn_delete_clicked=self.on_btn_delete_clicked,\n", + " on_btn_select_clicked=self.on_btn_select_clicked,\n", + " classes=self._bbox_state.classes,\n", + " readonly=not drawing_enabled\n", + " )\n", + "\n", + " self._refresh_children(0)\n", + "\n", " def __getitem__(self, index: int) -> BBoxVideoItem:\n", " return self.children[index]\n", "\n", @@ -243,13 +253,13 @@ "#hide\n", "\n", "# on bbox_canvas_state annotation change it reflects on the element list\n", - "assert len(bbox_coordinates.children) == 0\n", + "assert len(bbox_coordinates.children) == 0 # type: ignore\n", "bbox_canvas_state.bbox_coords = [BboxCoordinate(**{'x': 10, 'y': 10, 'width': 20, 'height': 30})]\n", - "assert len(bbox_coordinates.children) == 1\n", + "assert len(bbox_coordinates.children) == 1 # type: ignore\n", "\n", "# on element click it removes from state\n", - "bbox_coordinates.children[0].children[0].children[-1].click()\n", - "assert len(bbox_coordinates.children) == 0" + "bbox_coordinates.children[0].children[0].children[-1].click() # type: ignore\n", + "assert len(bbox_coordinates.children) == 0 # type: ignore" ] }, { @@ -259,9 +269,7 @@ "outputs": [], "source": [ "#hide\n", - "\n", "# it sync coords with classes\n", - "\n", "bbox_canvas_state.bbox_coords = [BboxCoordinate(**{'x': 10, 'y': 10, 'width': 20, 'height': 30})]\n", "assert len(bbox_state.labels) == 1" ] @@ -273,18 +281,21 @@ "outputs": [], "source": [ "#exporti\n", - "\n", "class BBoxAnnotatorGUI(AppLayout):\n", " def __init__(\n", " self,\n", " app_state: AppWidgetState,\n", " bbox_state: BBoxState,\n", - " on_save_btn_clicked: Callable = None\n", + " fit_canvas: bool,\n", + " on_save_btn_clicked: Callable = None,\n", + " has_border: bool = False\n", " ):\n", " self._app_state = app_state\n", " self._bbox_state = bbox_state\n", " self._on_save_btn_clicked = on_save_btn_clicked\n", " self._label_history: List[List[str]] = []\n", + " self.fit_canvas = fit_canvas\n", + " self.has_border = has_border\n", "\n", " self._navi = Navi()\n", "\n", @@ -308,7 +319,7 @@ " )\n", " )\n", "\n", - " self._image_box = BBoxCanvas(*self._app_state.size)\n", + " self._init_canvas(self._bbox_state.drawing_enabled)\n", "\n", " self.right_menu = BBoxCoordinates(\n", " app_state=self._app_state,\n", @@ -339,6 +350,7 @@ " self._redo_btn.on_click(self._redo_clicked)\n", "\n", " bbox_state.subscribe(self._set_image_path, 'image')\n", + " bbox_state.subscribe(self._init_canvas, 'drawing_enabled')\n", " bbox_state.subscribe(self._set_coords, 'coords')\n", " app_state.subscribe(self._set_max_im_number, 'max_im_number')\n", "\n", @@ -351,6 +363,14 @@ " pane_widths=(2, 8, 0),\n", " pane_heights=(1, 4, 1))\n", "\n", + " def _init_canvas(self, drawing_enabled: bool):\n", + " self._image_box = BBoxCanvas(\n", + " *self._app_state.size,\n", + " drawing_enabled=drawing_enabled,\n", + " fit_canvas=self.fit_canvas,\n", + " has_border=self.has_border\n", + " )\n", + "\n", " def _highlight_bbox(self, btn: ActionButton):\n", " self._image_box.highlight = btn.value\n", "\n", @@ -404,13 +424,13 @@ "outputs": [], "source": [ "#hide\n", - "\n", "app_state = AppWidgetState()\n", "bbox_state = BBoxState(classes=['test'])\n", - "\n", + "# TODO::check why this 'test' str it's been used on the actual annotator.\n", "BBoxAnnotatorGUI(\n", " app_state=app_state,\n", " bbox_state=bbox_state,\n", + " fit_canvas=False\n", ")" ] }, @@ -428,7 +448,6 @@ "outputs": [], "source": [ "#exporti\n", - "\n", "class BBoxAnnotatorController:\n", " def __init__(\n", " self,\n", @@ -450,7 +469,7 @@ " if render_previous_coords:\n", " self._update_coords(self._last_index)\n", "\n", - " def save_current_annotations(self, coords: dict):\n", + " def save_current_annotations(self, coords: List[BboxCoordinate]):\n", " self._bbox_state.set_quietly('coords', coords)\n", " self._save_annotations(self._app_state.index)\n", "\n", @@ -507,7 +526,7 @@ "metadata": {}, "outputs": [], "source": [ - "# hide\n", + "#hide\n", "# new index -> save *old* annotations -> update image -> update coordinates from annotation\n", "# |\n", "# |-> _update_annotations -> get current bbox values -> save to self.annotations" @@ -521,7 +540,7 @@ "source": [ "#export\n", "\n", - "class BBoxAnnotator:\n", + "class BBoxAnnotator(Annotator):\n", " \"\"\"\n", " Represents bounding box annotator.\n", "\n", @@ -534,24 +553,28 @@ " def __init__(\n", " self,\n", " project_path: Path,\n", - " input_item,\n", - " output_item,\n", + " input_item: InputImage,\n", + " output_item: OutputImageBbox,\n", " annotation_file_path: Path,\n", + " has_border: bool = False,\n", " *args, **kwargs\n", " ):\n", - " self.app_state = AppWidgetState(\n", + " app_state = AppWidgetState(\n", " uuid=str(id(self)),\n", " **{\n", " 'size': (input_item.width, input_item.height),\n", " }\n", " )\n", "\n", + " super().__init__(app_state)\n", + "\n", " self._input_item = input_item\n", " self._output_item = output_item\n", "\n", " self.bbox_state = BBoxState(\n", " uuid=str(id(self)),\n", - " classes=output_item.classes\n", + " classes=output_item.classes,\n", + " drawing_enabled=self._output_item.drawing_enabled\n", " )\n", "\n", " self.storage = JsonCaptureStorage(\n", @@ -569,7 +592,9 @@ " self.view = BBoxAnnotatorGUI(\n", " app_state=self.app_state,\n", " bbox_state=self.bbox_state,\n", - " on_save_btn_clicked=self.controller.save_current_annotations\n", + " fit_canvas=self._input_item.fit_canvas,\n", + " on_save_btn_clicked=self.controller.save_current_annotations,\n", + " has_border=has_border\n", " )\n", "\n", " self.view.on_client_ready(self.controller.handle_client_ready)\n", @@ -589,9 +614,7 @@ "outputs": [], "source": [ "#hide\n", - "from ipyannotator.mltypes import InputImage, OutputImageBbox\n", - "\n", - "in_p = InputImage(image_dir='pics', image_width=640, image_height=400)\n", + "in_p = InputImage(image_dir='pics', image_width=640, image_height=400, fit_canvas=True)\n", "out_p = OutputImageBbox(classes=['Label 01', 'Label 02'])" ] }, @@ -612,7 +635,7 @@ "metadata": {}, "outputs": [], "source": [ - "# hide\n", + "#hide\n", "from ipyannotator.storage import construct_annotation_path\n", "\n", "project_path = Path('../data/projects/bbox')\n", @@ -626,7 +649,7 @@ "metadata": {}, "outputs": [], "source": [ - "# hide\n", + "#hide\n", "bb = BBoxAnnotator(\n", " project_path=Path(project_path),\n", " input_item=in_p,\n", @@ -882,6 +905,71 @@ " assert new_max_coord == BboxCoordinate(*result)" ] }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_can_fit_canvas(bbox_fixture):\n", + " bbox_fixture.view._image_box._state.fit_canvas = True\n", + " bbox_fixture.view._navi._next_btn.click()\n", + " state = bbox_fixture.view._image_box._state\n", + " assert state.height == 400 \n", + " assert state.width == 640" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_can_fit_canvas_on_init():\n", + " in_p.image_width = None\n", + " in_p.image_height = None\n", + " in_p.fit_canvas = True\n", + " \n", + " bbox_fixture = BBoxAnnotator(\n", + " project_path=Path(project_path),\n", + " input_item=in_p,\n", + " output_item=out_p,\n", + " annotation_file_path=anno_file_path\n", + " )\n", + " \n", + " state = bbox_fixture.view._image_box._state\n", + "\n", + " assert bbox_fixture.view._image_box.state.fit_canvas == True\n", + " assert state.height == 400\n", + " assert state.width == 640" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_can_disable_drawing(bbox_fixture):\n", + " bbox_fixture.bbox_state.drawing_enabled = False\n", + " assert bbox_fixture.view._image_box.drawing_enabled is False" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_cant_delete_annotation_when_drawing_enable(bbox_fixture):\n", + " bbox_fixture.bbox_state.drawing_enabled = False\n", + " assert hasattr(bbox_fixture.view.right_menu[0], 'btn_delete') is False" + ] + }, { "cell_type": "code", "execution_count": null, @@ -898,7 +986,7 @@ "metadata": {}, "outputs": [], "source": [ - "# hide\n", + "#hide\n", "bb.to_dict()" ] }, diff --git a/nbs/05_image_button.ipynb b/nbs/05_image_button.ipynb index fe2467e..3860774 100644 --- a/nbs/05_image_button.ipynb +++ b/nbs/05_image_button.ipynb @@ -6,7 +6,7 @@ "metadata": {}, "outputs": [], "source": [ - "# default_exp image_button" + "# default_exp custom_input.buttons" ] }, { @@ -39,9 +39,11 @@ "#exporti\n", "from pathlib import Path\n", "\n", + "import attr\n", "from ipyevents import Event\n", "from ipywidgets import Image, VBox, Layout, Output, HTML\n", - "from traitlets import Bool, Unicode, HasTraits, observe" + "from traitlets import Bool, Unicode, HasTraits, observe\n", + "from typing import Optional, Union, Any" ] }, { @@ -62,6 +64,24 @@ "# Image button" ] }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#exporti\n", + "@attr.define(slots=False)\n", + "class ImageButtonSetting:\n", + " im_path: Optional[str] = None\n", + " label: Optional[Union[HTML, str]] = None\n", + " im_name: Optional[str] = None\n", + " im_index: Optional[Any] = None\n", + " display_label: bool = True\n", + " image_width: str = '50px'\n", + " image_height: Optional[str] = None" + ] + }, { "cell_type": "code", "execution_count": null, @@ -93,23 +113,21 @@ " image_path = Unicode()\n", " label_value = Unicode()\n", "\n", - " def __init__(self, im_path=None, label=None,\n", - " im_name=None, im_index=None,\n", - " display_label=True, image_width='50px', image_height=None):\n", + " def __init__(self, setting: ImageButtonSetting):\n", "\n", - " self.display_label = display_label\n", - " self.label = 'None'\n", + " self.setting = setting\n", " self.image = Image(\n", " layout=Layout(display='flex',\n", " justify_content='center',\n", " align_items='center',\n", " align_content='center',\n", - " width=image_width,\n", - " height=image_height),\n", + " width=setting.image_width,\n", + " margin='0 0 0 0',\n", + " height=setting.image_height),\n", " )\n", "\n", - " if self.display_label: # both image and label\n", - " self.label = HTML(\n", + " if self.setting.display_label: # both image and label\n", + " self.setting.label = HTML(\n", " value='?',\n", " layout=Layout(display='flex',\n", " justify_content='center',\n", @@ -117,15 +135,18 @@ " align_content='center'),\n", " )\n", " else: # no label (capture image case)\n", - " self.im_name = im_name\n", - " self.im_index = im_index\n", + " self.im_name = self.setting.im_name\n", + " self.im_index = self.setting.im_index\n", " self.image.layout.border = 'solid 1px gray'\n", " self.image.layout.object_fit = 'contain'\n", + " self.image.margin = '0 0 0 0'\n", + " self.image.layout.overflow = 'hidden'\n", "\n", " super().__init__(layout=Layout(align_items='center',\n", " margin='3px',\n", + " overflow='hidden',\n", " padding='2px'))\n", - " if not im_path:\n", + " if not setting.im_path:\n", " self.clear()\n", "\n", " self.d = Event(source=self, watched_events=['click'])\n", @@ -137,26 +158,26 @@ " self.image.value = open(new_path, \"rb\").read()\n", " if not self.children:\n", " self.children = (self.image,)\n", - " if self.display_label:\n", - " self.children += (self.label,)\n", + " if self.setting.display_label:\n", + " self.children += (self.setting.label,)\n", " else:\n", " #do not display image widget\n", - " self.children = []\n", + " self.children = tuple()\n", "\n", " @observe('label_value')\n", " def _read_label(self, change=None):\n", " new_label = change['new']\n", "\n", - " if isinstance(self.label, HTML):\n", - " self.label.value = new_label\n", + " if isinstance(self.setting.label, HTML):\n", + " self.setting.label.value = new_label\n", " else:\n", - " self.label = new_label\n", + " self.setting.label = new_label\n", "\n", " def clear(self):\n", - " if isinstance(self.label, HTML):\n", - " self.label.value = ''\n", + " if isinstance(self.setting.label, HTML):\n", + " self.setting.label.value = ''\n", " else:\n", - " self.label = ''\n", + " self.setting.label = ''\n", " self.image_path = ''\n", " self.active = False\n", "\n", @@ -164,21 +185,36 @@ " def mark(self, ev):\n", " # pad to compensate self size with border\n", " if self.active:\n", - " if self.display_label:\n", + " if self.setting.display_label:\n", " self.layout.border = 'solid 2px #1B8CF3'\n", " self.layout.padding = '0px'\n", " else:\n", " self.image.layout.border = 'solid 3px #1B8CF3'\n", " self.image.layout.padding = '0px'\n", " else:\n", - " if self.display_label:\n", + " if self.setting.display_label:\n", " self.layout.border = 'none'\n", " self.layout.padding = '2px'\n", " else:\n", " self.image.layout.border = 'solid 1px gray'\n", "\n", + " def __eq__(self, other):\n", + " equals = [\n", + " other.image_path == self.image_path,\n", + " other.label_value == self.label_value,\n", + " other.active == self.active,\n", + " ]\n", + "\n", + " return all(equals)\n", + "\n", + " def update(self, other):\n", + " if self != other:\n", + " self.image_path = other.image_path\n", + " self.label_value = other.label_value\n", + " self.active = other.active\n", + "\n", " @property\n", - " def name(self):\n", + " def value(self):\n", " return Path(self.image_path).name\n", "\n", " @debug_output.capture(clear_output=False)\n", @@ -208,7 +244,8 @@ "outputs": [], "source": [ "# hide\n", - "imb = ImageButton()\n", + "setting = ImageButtonSetting()\n", + "imb = ImageButton(setting)\n", "display(imb), display(imb.debug_output)" ] }, @@ -220,7 +257,7 @@ "source": [ "# hide\n", "assert not imb.active\n", - "imb.name" + "imb.value" ] }, { @@ -276,7 +313,7 @@ "imb.image_path = '../data/mock/pics/test200x200.png'\n", "imb.label_value = 'new_label'\n", "imb.active = True\n", - "assert imb.name == 'test200x200.png'" + "assert imb.value == 'test200x200.png'" ] }, { @@ -296,8 +333,12 @@ "outputs": [], "source": [ "# hide\n", - "im_button = ImageButton(im_path='../data/mock/pics/test200x200.png', label='hm',\n", - " display_label=False)\n", + "button_setting = ImageButtonSetting(\n", + " im_path='../data/mock/pics/test200x200.png',\n", + " label='hm',\n", + " display_label=False\n", + ")\n", + "im_button = ImageButton(button_setting)\n", "\n", "\n", "def handle_event_(event, name=None):\n", @@ -321,20 +362,6 @@ "notebook2script()" ] }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, { "cell_type": "code", "execution_count": null, @@ -345,7 +372,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/06_capture_annotator.ipynb b/nbs/06_capture_annotator.ipynb index c9d1e80..92f05b9 100644 --- a/nbs/06_capture_annotator.ipynb +++ b/nbs/06_capture_annotator.ipynb @@ -40,20 +40,33 @@ "\n", "import math\n", "import warnings\n", - "from functools import partial\n", + "from copy import deepcopy\n", "from pathlib import Path\n", - "from typing import Dict, Optional, List, Iterable, Callable\n", + "from typing import Dict, Optional, Callable, List\n", "\n", - "from IPython.core.display import display\n", - "from ipywidgets import (AppLayout, VBox, HBox, Button, GridBox,\n", - " Layout, Checkbox, HTML, IntText, Output)\n", + "from IPython.display import display\n", + "from ipywidgets import (AppLayout, HBox, Button, HTML, VBox,\n", + " Layout, Checkbox, Output)\n", "\n", - "from ipyannotator.base import BaseState, AppWidgetState\n", - "from ipyannotator.image_button import ImageButton\n", + "from ipyannotator.custom_widgets.grid_menu import GridMenu, Grid\n", + "from ipyannotator.base import BaseState, AppWidgetState, Annotator\n", "from ipyannotator.navi_widget import Navi\n", + "from ipyannotator.ipytyping.annotations import LabelStore, _label_store_to_image_button\n", "from ipyannotator.storage import JsonCaptureStorage" ] }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "import ipytest\n", + "import pytest\n", + "ipytest.autoconfig(raise_on_error=True)" + ] + }, { "cell_type": "markdown", "metadata": {}, @@ -85,190 +98,25 @@ "outputs": [], "source": [ "#exporti\n", - "\n", "class CaptureState(BaseState):\n", - " annotations: Dict[str, Optional[Dict[str, bool]]] = {}\n", - " disp_number: int = 9\n", + " annotations: LabelStore = LabelStore()\n", + " grid: Grid\n", " question_value: str = ''\n", - " all_none: bool = False\n", - " n_rows: int = 3\n", - " n_cols: int = 3" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "## View\n", - "\n", - "For the view an internal component called ` CaptureGrid ` it's developed, this component allows us to display the options on screen. " - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "#export\n", - "\n", - "class CaptureGrid(GridBox):\n", - " \"\"\"\n", - " Represents grid of `ImageButtons` with state.\n", - "\n", - " \"\"\"\n", - " debug_output = Output(layout={'border': '1px solid black'})\n", - "\n", - " def __init__(self, grid_item=ImageButton, image_width=150, image_height=150,\n", - " n_rows=3, n_cols=3, display_label=False):\n", - "\n", - " self.image_width = image_width\n", - " self.image_height = image_height\n", - " self.n_rows = n_rows\n", - " self.n_cols = n_cols\n", - " self._screen_im_number = IntText(value=n_rows * n_cols,\n", - " description='screen_image_number',\n", - " disabled=False)\n", - "\n", - " self._labels = [grid_item(\n", - " display_label=display_label, image_width='%dpx' % self.image_width,\n", - " image_height='%dpx' % self.image_height) for _ in range(self._screen_im_number.value)]\n", - "\n", - " self.callback = None\n", - "\n", - " gap = 40 if display_label else 15\n", - "\n", - " centered_settings = {\n", - " 'grid_template_columns': \" \".join([\"%dpx\" % (self.image_width + gap) for i\n", - " in range(self.n_cols)]),\n", - " 'grid_template_rows': \" \".join([\"%dpx\" % (self.image_height + gap) for i\n", - " in range(self.n_rows)]),\n", - " 'justify_content': 'center',\n", - " 'align_content': 'space-around'\n", - " }\n", - "\n", - " super().__init__(children=self._labels, layout=Layout(**centered_settings))\n", - "\n", - " @debug_output.capture(clear_output=True)\n", - " def load_annotations_labels(self, annotations: Optional[Iterable[Dict]] = None):\n", - " # error: Argument 1 to \"iter\" has incompatible type\n", - " # \"Optional[Iterable[Dict[Any, Any]]]\"; expected \"Iterable[Dict[Any, Any]]\"\n", - " iter_state = iter(annotations) # type: ignore\n", - "\n", - " for label in self._labels:\n", - " p = next(iter_state, None)\n", - " if p:\n", - " label.image_path = str(p) # type: ignore\n", - " label.label_value = Path(p).stem # type: ignore\n", - " label.active = annotations[p].get('answer', False) # type: ignore\n", - " else:\n", - " label.clear()\n", - "\n", - " if self.callback:\n", - " self.register_on_click()\n", - "\n", - " def on_click(self, cb: Callable):\n", - " self.callback = cb\n", - " self.register_on_click()\n", - "\n", - " @debug_output.capture(clear_output=True)\n", - " def register_on_click(self):\n", - " for label in self._labels:\n", - " label.reset_callbacks()\n", - " label.on_click(partial(self.callback, name=label.name))" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "ibg = CaptureGrid(grid_item=ImageButton, image_width=50, image_height=75)\n", - "ibg" - ] - }, - { - "cell_type": "markdown", - "metadata": {}, - "source": [ - "You should not see anything at this step, until you set a correct visual state (see next step below)" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "ibg.debug_output" - ] - }, - { - "attachments": { - "06_capture_annotator1.png": { - "image/png": "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" - } - }, - "cell_type": "markdown", - "metadata": {}, - "source": [ - "The result will be similar to:\n", - "\n", - "![06_capture_annotator1.png](attachment:06_capture_annotator1.png)" + " all_none: bool = False" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "Actually `CaptureGrid` does not have own `on_click` event listener, but grid elements itself should implement `on_click(ev)` and `reset_callbacks()` methods to register/reset onclick callback function respectively. Also grid element shoudl have a field `name` in order user can destinguish between grid children.\n", - "\n", - "In current implementation `ImageButton` is default grid element.\n", - "\n", - "While ipyevents implementation lacks `sender` or `source` in callback args, `functools.partial` used to back element `name` into return value. You can see example of on_click event handler `test_handler` below. \n", - "`name` of the button is printed out on click." - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "# hide\n", - "annotations_on_state = {\n", - " '../data/projects/capture1/pics/pink25x25.png': {'answer': True},\n", - " '../data/mock/pics/test200x200.png': {'answer': True}, '': {'answer': False}\n", - "}\n", - "\n", - "ibg.load_annotations_labels(annotations_on_state)" - ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [ - "# hide\n", - "h = HTML('Event info')\n", - "display(h)\n", - "\n", - "\n", - "def test_handler(event, name=None):\n", - " event.update({'label_name': name})\n", - " h.value = event['label_name']\n", - "\n", - "\n", - "ibg.on_click(test_handler)" + "## View" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "The ` CaptureAnnotatorGUI ` joins the internal component (`CaptureGrid`) with the navi component and its interaction." + "The ` CaptureAnnotatorGUI ` joins the internal component (`GridMenu`) with the navi component and its interaction." ] }, { @@ -289,6 +137,8 @@ " activated when the user navigates through the annotator\n", " \"\"\"\n", "\n", + " debug_output = Output(layout={'border': '1px solid black'})\n", + "\n", " def __init__(\n", " self,\n", " app_state: AppWidgetState,\n", @@ -304,12 +154,6 @@ " self._grid_box_clicked = grid_box_clicked\n", " self._select_none_changed = select_none_changed\n", "\n", - " self._screen_im_number = IntText(\n", - " value=self._capture_state.n_rows * self._capture_state.n_cols,\n", - " description='screen_image_number',\n", - " disabled=False\n", - " )\n", - "\n", " self._navi = Navi()\n", "\n", " self._save_btn = Button(description=\"Save\",\n", @@ -333,13 +177,7 @@ " )\n", " )\n", "\n", - " self._grid_box = CaptureGrid(\n", - " image_width=self._app_state.size[0],\n", - " image_height=self._app_state.size[1],\n", - " n_rows=self._capture_state.n_rows,\n", - " n_cols=self._capture_state.n_cols,\n", - " display_label=False\n", - " )\n", + " self._grid_box = GridMenu(capture_state.grid)\n", "\n", " self._grid_label = HTML()\n", " self._labels_box = VBox(\n", @@ -355,9 +193,10 @@ " )\n", " )\n", "\n", - " self._navi.on_navi_clicked = on_navi_clicked\n", + " self.on_navi_clicked = on_navi_clicked\n", + " self._navi.on_navi_clicked = self._on_navi_clicked\n", " self._save_btn.on_click(self._btn_clicked)\n", - " self._grid_box.on_click(self._grid_clicked)\n", + " self._grid_box.on_click(self.on_grid_clicked)\n", " self._none_checkbox.observe(self._none_checkbox_changed, 'value')\n", "\n", " if self._capture_state.question_value:\n", @@ -367,12 +206,12 @@ " self._set_navi_max_im_number(self._app_state.max_im_number)\n", "\n", " if self._capture_state.annotations:\n", - " self._grid_box.load_annotations_labels(self._capture_state.annotations)\n", + " self._load_menu(self._capture_state.annotations)\n", "\n", " self._capture_state.subscribe(self._set_none_checkbox, 'all_none')\n", " self._capture_state.subscribe(self._set_label, 'question_value')\n", " self._app_state.subscribe(self._set_navi_max_im_number, 'max_im_number')\n", - " self._capture_state.subscribe(self._grid_box.load_annotations_labels, 'annotations')\n", + " self._capture_state.subscribe(self._load_menu, 'annotations')\n", "\n", " super().__init__(\n", " header=None,\n", @@ -383,6 +222,20 @@ " pane_widths=(2, 8, 0),\n", " pane_heights=(1, 4, 1))\n", "\n", + " def _on_navi_clicked(self, index: int):\n", + " if self.on_navi_clicked:\n", + " self.on_navi_clicked(index)\n", + "\n", + " self._grid_box.load(\n", + " _label_store_to_image_button(self._capture_state.annotations)\n", + " )\n", + "\n", + " @debug_output.capture(clear_output=True)\n", + " def _load_menu(self, annotations: LabelStore):\n", + " self._grid_box.load(\n", + " _label_store_to_image_button(annotations)\n", + " )\n", + "\n", " def _set_none_checkbox(self, all_none: bool):\n", " self._none_checkbox.value = all_none\n", "\n", @@ -403,13 +256,42 @@ " if self._select_none_changed:\n", " self._select_none_changed(change)\n", "\n", - " def _grid_clicked(self, event, name=None):\n", + " def on_grid_clicked(self, event, name=None):\n", " if self._grid_box_clicked:\n", " self._grid_box_clicked(event, name)\n", " else:\n", " warnings.warn(\"Grid box click didn't triggered any event.\")" ] }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "#hide\n", + "from ipyannotator.custom_input.buttons import ImageButton" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_gui_loads_image_button_from_menu():\n", + " annotations = LabelStore()\n", + " annotations['../data/projects/capture1/pics/pink25x25.png'] = {'answer': True}\n", + " grid = Grid(width=200, height=200)\n", + " gui = CaptureAnnotatorGUI(\n", + " app_state=AppWidgetState(),\n", + " capture_state=CaptureState(annotations=annotations, grid=grid)\n", + " )\n", + " gui._load_menu(gui._capture_state.annotations)\n", + " assert isinstance(gui._grid_box.widgets[0], ImageButton)" + ] + }, { "cell_type": "code", "execution_count": null, @@ -421,9 +303,12 @@ "# \"**Dict[str, Tuple[int, int]]\"; expected \"Optional[str]\"\n", "app_state = AppWidgetState(**{'size': (50, 50)}) # type: ignore\n", "\n", + "grid = Grid(width=100, height=100, n_rows=5, n_cols=5)\n", + "\n", "# error: Argument 1 to \"CaptureState\" has incompatible type\n", "# \"**Dict[str, int]\"; expected \"Optional[str]\"\n", - "capture_state = CaptureState(**{'n_rows': 5, 'n_cols': 5}) # type: ignore\n", + "capture_state = CaptureState(\n", + " **{'grid': grid, 'annotations': LabelStore()}) # type: ignore\n", "\n", "ca = CaptureAnnotatorGUI(\n", " capture_state=capture_state,\n", @@ -442,9 +327,10 @@ "outputs": [], "source": [ "# hide\n", - "ca._capture_state.annotations = {\n", + "data = {\n", " '../data/projects/capture1/pics/pink25x25.png': {'answer': True}\n", - "}" + "}\n", + "ca._capture_state.annotations.update(data)" ] }, { @@ -484,8 +370,8 @@ "pub.sendMessage('CaptureState.annotations', annotations={\n", " '../data/projects/capture1/pics/pink25x25.png': {'answer': False}\n", "})\n", - "\n", - "assert list(filter(lambda l: l.active, ca._grid_box._labels)) == []\n", + "assert ca._grid_box.widgets is not None\n", + "assert list(filter(lambda l: l.active, ca._grid_box.widgets)) == []\n", "\n", "# it throw warning if no btn_clicked callable is provided\n", "\n", @@ -593,9 +479,6 @@ " self.output_item = output_item\n", " self._last_index = 0\n", "\n", - " self._capture_state.subscribe(self.update_state, 'disp_number')\n", - " self._capture_state.subscribe(self._calc_screens_num, 'disp_number')\n", - "\n", " self.images = self._storage.get_im_names(filter_files)\n", " self.current_im_number = len(self.images)\n", "\n", @@ -603,11 +486,12 @@ " self._capture_state.question_value = ('

{question}

')\n", "\n", - " self.update_state(self._capture_state.disp_number)\n", - " self._calc_screens_num(self._capture_state.disp_number)\n", + " self.update_state()\n", + " self._calc_screens_num()\n", "\n", - " def update_state(self, disp_number: int):\n", + " def update_state(self):\n", " state_images = self._get_state_names(self._app_state.index)\n", + " tmp_annotations = deepcopy(self._capture_state.annotations)\n", " current_state = {}\n", "\n", " for im_path in state_images:\n", @@ -618,7 +502,9 @@ " # error: Incompatible types in assignment (expression has type\n", " # \"Dict[str, Dict[Any, Any]]\", variable has type\n", " # \"Dict[str, Optional[Dict[str, bool]]]\")\n", - " self._capture_state.annotations = current_state # type: ignore\n", + " tmp_annotations.clear()\n", + " tmp_annotations.update(current_state)\n", + " self._capture_state.annotations = tmp_annotations # type: ignore\n", "\n", " def _update_all_none_state(self, state_images: dict):\n", " self._capture_state.all_none = all(\n", @@ -626,13 +512,13 @@ " )\n", "\n", " def save_annotations(self, index: int): # to disk\n", - " state_images = self._capture_state.annotations\n", + " state_images = dict(self._capture_state.annotations)\n", "\n", " self._storage.update_annotations(state_images)\n", "\n", " def _get_state_names(self, index: int) -> List[str]:\n", - " start = index * self._capture_state.disp_number\n", - " end = start + self._capture_state.disp_number\n", + " start = index * self._capture_state.grid.disp_number\n", + " end = start + self._capture_state.grid.disp_number\n", " im_names = self.images[start:end]\n", " return im_names\n", "\n", @@ -642,24 +528,21 @@ " '''\n", " self._app_state.set_quietly('index', index)\n", " self.save_annotations(self._last_index)\n", - " self.update_state(self._capture_state.disp_number)\n", + " self.update_state()\n", " self._last_index = index\n", "\n", - " def _calc_screens_num(self, disp_number: int):\n", + " def _calc_screens_num(self):\n", " self._app_state.max_im_number = math.ceil(\n", - " self.current_im_number / self._capture_state.disp_number\n", + " self.current_im_number / self._capture_state.grid.disp_number\n", " )\n", "\n", " @debug_output.capture(clear_output=False)\n", " def handle_grid_click(self, event: dict, name=None):\n", " p = self._storage.input_item_path / name\n", - " current_state = self._capture_state.annotations.copy()\n", - "\n", + " current_state = deepcopy(self._capture_state.annotations)\n", " if not p.is_dir():\n", - " # error: Item \"None\" of \"Optional[Dict[str, bool]]\"\n", - " # has no attribute \"get\"\n", " state_answer = self._capture_state.annotations[\n", - " str(p)].get('answer', False) # type: ignore\n", + " str(p)].get('answer', False)\n", " current_state[str(p)] = {'answer': not state_answer}\n", "\n", " for k, v in current_state.items():\n", @@ -669,7 +552,7 @@ " if self._capture_state.all_none:\n", " self._capture_state.all_none = False\n", " else:\n", - " self._update_all_none_state(current_state)\n", + " self._update_all_none_state(dict(current_state))\n", " else:\n", " return\n", "\n", @@ -677,8 +560,13 @@ "\n", " def select_none(self, change: dict):\n", " if self._capture_state.all_none:\n", - " self._capture_state.annotations = {p: {\n", - " 'answer': False} for p in self._capture_state.annotations}" + " tmp_annotations = deepcopy(self._capture_state.annotations)\n", + " tmp_annotations.clear()\n", + " tmp_annotations.update(\n", + " {p: {\n", + " 'answer': False} for p in self._capture_state.annotations}\n", + " )\n", + " self._capture_state.annotations = tmp_annotations" ] }, { @@ -708,7 +596,7 @@ ")\n", "\n", "app_state = AppWidgetState()\n", - "capture_state = CaptureState()\n", + "capture_state = CaptureState(grid=grid)\n", "\n", "caController = CaptureAnnotatorController(\n", " app_state=app_state,\n", @@ -893,7 +781,7 @@ "source": [ "#export\n", "\n", - "class CaptureAnnotator:\n", + "class CaptureAnnotator(Annotator):\n", " debug_output = Output(layout={'border': '1px solid black'})\n", " \"\"\"\n", " Represents capture annotator.\n", @@ -913,7 +801,7 @@ " annotation_file_path,\n", " n_rows=3,\n", " n_cols=3,\n", - " disp_number: int = 9,\n", + " disp_number=9,\n", " question=None,\n", " filter_files=None\n", " ):\n", @@ -927,21 +815,29 @@ " self._annotation_file_path = annotation_file_path\n", " self._n_rows = n_rows\n", " self._n_cols = n_cols\n", - " self._disp_number = disp_number\n", " self._question = question\n", " self._filter_files = filter_files\n", "\n", - " self.app_state = AppWidgetState(\n", + " app_state = AppWidgetState(\n", " uuid=str(id(self)),\n", " **{'size': (input_item.width, input_item.height)}\n", " )\n", + "\n", + " super().__init__(app_state)\n", + "\n", + " grid = Grid(\n", + " width=input_item.width,\n", + " height=input_item.height,\n", + " n_rows=n_rows,\n", + " n_cols=n_cols,\n", + " display_label=False,\n", + " disp_number=disp_number\n", + " )\n", + "\n", " self.capture_state = CaptureState(\n", " uuid=str(id(self)),\n", - " **{\n", - " 'n_cols': n_cols,\n", - " 'n_rows': n_rows,\n", - " 'disp_number': disp_number\n", - " }\n", + " annotations=LabelStore(),\n", + " grid=grid\n", " )\n", "\n", " self.storage = CaptureAnnotationStorage(\n", @@ -999,7 +895,6 @@ " input_item=in_p,\n", " output_item=out_p,\n", " annotation_file_path=anno_file_path,\n", - " n_cols=3,\n", " question=\"Select pink squares\"\n", ")" ] @@ -1065,7 +960,7 @@ "# it save annotations status when user navigates\n", "\n", "ca_annotator.controller.handle_grid_click(\n", - " event={},\n", + " event=None,\n", " name='pink50x125.png'\n", ")\n", "\n", @@ -1145,7 +1040,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/07_im2im_annotator.ipynb b/nbs/07_im2im_annotator.ipynb index d55b087..b43af3a 100644 --- a/nbs/07_im2im_annotator.ipynb +++ b/nbs/07_im2im_annotator.ipynb @@ -37,22 +37,37 @@ "outputs": [], "source": [ "#exporti\n", - "\n", + "import io\n", "import warnings\n", - "from math import ceil\n", "from pathlib import Path\n", - "from typing import Optional, Dict, List, Callable\n", + "from copy import deepcopy\n", + "from typing import Optional, Callable, Union, Iterable\n", "\n", "from ipycanvas import Canvas\n", - "from ipywidgets import (AppLayout, VBox, HBox, Button, Layout, HTML, Output)\n", + "from ipywidgets import (AppLayout, VBox, HBox, Button, Layout, HTML, Output, Image)\n", "\n", - "from ipyannotator.base import BaseState, AppWidgetState\n", - "from ipyannotator.bbox_canvas import draw_img\n", - "from ipyannotator.capture_annotator import CaptureGrid\n", - "from ipyannotator.image_button import ImageButton\n", + "from ipyannotator.base import BaseState, AppWidgetState, Annotator, AnnotatorStep\n", + "from ipyannotator.bbox_canvas import ImageRenderer\n", + "from ipyannotator.mltypes import OutputImageLabel, OutputLabel, InputImage\n", + "from ipyannotator.ipytyping.annotations import LabelStore, LabelStoreCaster\n", + "from ipyannotator.custom_widgets.grid_menu import GridMenu, Grid\n", "from ipyannotator.navi_widget import Navi\n", "from ipyannotator.storage import JsonLabelStorage\n", - "from IPython.display import display" + "from IPython.display import display\n", + "from ipyannotator.doc_utils import is_building_docs\n", + "from PIL import Image as PILImage" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "# hide\n", + "import ipytest\n", + "import pytest\n", + "ipytest.autoconfig(raise_on_error=True)" ] }, { @@ -90,16 +105,12 @@ "# exporti\n", "\n", "class Im2ImState(BaseState):\n", - " annotations: Dict[str, Optional[List[str]]] = {}\n", - " disp_number: int = 9\n", + " annotations: LabelStore = LabelStore()\n", " question_value: str = ''\n", - " n_rows: Optional[int] = 3\n", - " n_cols: Optional[int] = 3\n", + " grid: Grid\n", " image_path: Optional[str]\n", " im_width: int = 300\n", - " im_height: int = 300\n", - " label_width: int = 150\n", - " label_height: int = 150" + " im_height: int = 300" ] }, { @@ -118,28 +129,62 @@ "outputs": [], "source": [ "# export\n", + "if is_building_docs():\n", + " class ImCanvas(Image):\n", + " def __init__(\n", + " self,\n", + " width: int = 150,\n", + " height: int = 150,\n", + " has_border: bool = False,\n", + " fit_canvas: bool = False\n", + " ):\n", + " super().__init__(width=width, height=height)\n", + " image = PILImage.new('RGB', (100, 100), (255, 255, 255))\n", + " b = io.BytesIO()\n", + " image.save(b, format='PNG')\n", + " self.value = b.getvalue()\n", + "\n", + " def _draw_image(self, image_path: str):\n", + " self.value = Image.from_file(image_path).value\n", + "\n", + " def _clear_image(self):\n", + " pass\n", + "\n", + " def observe_client_ready(self, cb=None):\n", + " pass\n", + "else:\n", + " class ImCanvas(HBox): # type: ignore\n", + " def __init__(\n", + " self,\n", + " width: int = 150,\n", + " height: int = 150,\n", + " has_border: bool = False,\n", + " fit_canvas: bool = False\n", + " ):\n", + " self.has_border = has_border\n", + " self.fit_canvas = fit_canvas\n", + " self._canvas = Canvas(width=width, height=height)\n", + " super().__init__([self._canvas])\n", + "\n", + " def _draw_image(self, image_path: str):\n", + " img_render_strategy = ImageRenderer(\n", + " clear=True,\n", + " has_border=self.has_border,\n", + " fit_canvas=self.fit_canvas\n", + " )\n", "\n", - "class ImCanvas(HBox):\n", - " def __init__(self, width=150, height=150, has_border=False):\n", - " self.has_border = has_border\n", - " self._canvas = Canvas(width=width, height=height)\n", - " super().__init__([self._canvas])\n", - "\n", - " def _draw_image(self, image_path: str):\n", - " self._image_scale = draw_img(\n", - " self._canvas,\n", - " image_path,\n", - " clear=True,\n", - " has_border=self.has_border\n", - " )\n", + " self._image_scale = img_render_strategy.render(\n", + " self._canvas,\n", + " image_path\n", + " )\n", "\n", - " def _clear_image(self):\n", - " self._canvas.clear()\n", + " def _clear_image(self):\n", + " self._canvas.clear()\n", "\n", - " # needed to support voila\n", - " # https://ipycanvas.readthedocs.io/en/latest/advanced.html#ipycanvas-in-voila\n", - " def observe_client_ready(self, cb=None):\n", - " self._canvas.on_client_ready(cb)" + " # needed to support voila\n", + " # https://ipycanvas.readthedocs.io/en/latest/advanced.html#ipycanvas-in-voila\n", + " def observe_client_ready(self, cb=None):\n", + " self._canvas.on_client_ready(cb)" ] }, { @@ -180,40 +225,48 @@ "#exporti\n", "\n", "class Im2ImAnnotatorGUI(AppLayout):\n", + " debug_output = Output(layout={'border': '1px solid black'})\n", + "\n", " def __init__(\n", " self,\n", " app_state: AppWidgetState,\n", " im2im_state: Im2ImState,\n", + " state_to_widget: LabelStoreCaster,\n", " label_autosize=False,\n", - " save_btn_clicked: Callable = None,\n", - " grid_box_clicked: Callable = None,\n", - " has_border: bool = False\n", + " on_save_btn_clicked: Callable = None,\n", + " on_grid_box_clicked: Callable = None,\n", + " on_navi_clicked: Callable = None,\n", + " has_border: bool = False,\n", + " fit_canvas: bool = False\n", " ):\n", " self._app_state = app_state\n", " self._im2im_state = im2im_state\n", - " self.save_btn_clicked = save_btn_clicked\n", - " self.grid_box_clicked = grid_box_clicked\n", + " self._on_save_btn_clicked = on_save_btn_clicked\n", + " self._on_navi_clicked = on_navi_clicked\n", + " self._on_grid_box_clicked = on_grid_box_clicked\n", + " self.state_to_widget = state_to_widget\n", "\n", " if label_autosize:\n", " if self._im2im_state.im_width < 100 or self._im2im_state.im_height < 100:\n", - " self._im2im_state.set_quietly('label_width', 10)\n", - " self._im2im_state.set_quietly('label_height', 10)\n", + " self._im2im_state.grid.width = 10\n", + " self._im2im_state.grid.height = 10\n", " elif self._im2im_state.im_width > 1000 or self._im2im_state.im_height > 1000:\n", - " self._im2im_state.set_quietly('label_width', 50)\n", - " self._im2im_state.set_quietly('label_height', 10)\n", + " self._im2im_state.grid.width = 50\n", + " self._im2im_state.grid.height = 10\n", " else:\n", - " label_width = min(self._im2im_state.im_width, self._im2im_state.im_height) / 10\n", - " self._im2im_state.set_quietly('label_width', label_width)\n", - " self._im2im_state.set_quietly('label_height', label_width)\n", + " label_width = min(self._im2im_state.im_width, self._im2im_state.im_height) // 10\n", + " self._im2im_state.grid.width = label_width\n", + " self._im2im_state.grid.height = label_width\n", "\n", " self._image = ImCanvas(\n", " width=self._im2im_state.im_width,\n", " height=self._im2im_state.im_height,\n", - " has_border=has_border\n", + " has_border=has_border,\n", + " fit_canvas=fit_canvas\n", " )\n", "\n", " self._navi = Navi()\n", - "\n", + " self._navi.on_navi_clicked = self.on_navi_clicked\n", " self._save_btn = Button(description=\"Save\",\n", " layout=Layout(width='auto'))\n", "\n", @@ -227,13 +280,7 @@ " )\n", " )\n", "\n", - " self._grid_box = CaptureGrid(\n", - " grid_item=ImageButton,\n", - " image_width=self._im2im_state.label_width,\n", - " image_height=self._im2im_state.label_height,\n", - " n_rows=self._im2im_state.n_rows,\n", - " n_cols=self._im2im_state.n_cols\n", - " )\n", + " self._grid_box = GridMenu(self._im2im_state.grid)\n", "\n", " self._grid_label = HTML(value=\"LABEL\",)\n", " self._labels_box = VBox(\n", @@ -241,51 +288,74 @@ " layout=Layout(\n", " display='flex',\n", " justify_content='center',\n", - " flex_wrap='wrap',\n", " align_items='center')\n", " )\n", "\n", + " self._save_btn.on_click(self._on_btn_clicked)\n", + " self._grid_box.on_click(self.on_grid_clicked)\n", + "\n", " if self._app_state.max_im_number:\n", " self._set_navi_max_im_number(self._app_state.max_im_number)\n", "\n", " if self._im2im_state.annotations:\n", - " self._grid_box.load_annotations_labels(self._im2im_state.annotations)\n", + " self._grid_box.load(\n", + " self.state_to_widget(self._im2im_state.annotations)\n", + " )\n", "\n", " if self._im2im_state.question_value:\n", " self._set_label(self._im2im_state.question_value)\n", "\n", " self._im2im_state.subscribe(self._set_label, 'question_value')\n", " self._im2im_state.subscribe(self._image._draw_image, 'image_path')\n", - " self._im2im_state.subscribe(self._grid_box.load_annotations_labels, 'annotations')\n", - " self._save_btn.on_click(self._btn_clicked)\n", - " self._grid_box.on_click(self._grid_clicked)\n", + " self._im2im_state.subscribe(self.load_menu, 'annotations')\n", + "\n", + " layout = Layout(\n", + " display='flex',\n", + " justify_content='center',\n", + " align_items='center'\n", + " )\n", + "\n", + " im2im_display = HBox([\n", + " VBox([self._image, self._controls_box]),\n", + " self._labels_box\n", + " ], layout=layout)\n", "\n", " super().__init__(\n", " header=None,\n", - " left_sidebar=VBox([self._image, self._controls_box],\n", - " layout=Layout(display='flex', justify_content='center',\n", - " flex_wrap='wrap', align_items='center')),\n", - " center=self._labels_box,\n", + " left_sidebar=None,\n", + " center=im2im_display,\n", " right_sidebar=None,\n", " footer=None,\n", " pane_widths=(6, 4, 0),\n", " pane_heights=(1, 1, 1))\n", "\n", + " @debug_output.capture(clear_output=False)\n", + " def load_menu(self, annotations: LabelStore):\n", + " self._grid_box.load(\n", + " self.state_to_widget(annotations)\n", + " )\n", + "\n", + " @debug_output.capture(clear_output=False)\n", + " def on_navi_clicked(self, index: int):\n", + " if self._on_navi_clicked:\n", + " self._on_navi_clicked(index)\n", + "\n", " def _set_navi_max_im_number(self, max_im_number: int):\n", " self._navi.max_im_num = max_im_number\n", "\n", " def _set_label(self, question_value: str):\n", " self._grid_label.value = question_value\n", "\n", - " def _btn_clicked(self, *args):\n", - " if self.save_btn_clicked:\n", - " self.save_btn_clicked(*args)\n", + " def _on_btn_clicked(self, *args):\n", + " if self._on_save_btn_clicked:\n", + " self._on_save_btn_clicked(*args)\n", " else:\n", " warnings.warn(\"Save button click didn't triggered any event.\")\n", "\n", - " def _grid_clicked(self, event, name=None):\n", - " if self.grid_box_clicked:\n", - " self.grid_box_clicked(event, name)\n", + " @debug_output.capture(clear_output=False)\n", + " def on_grid_clicked(self, event, value=None):\n", + " if self._on_grid_box_clicked:\n", + " self._on_grid_box_clicked(event, value)\n", " else:\n", " warnings.warn(\"Grid box click didn't triggered any event.\")\n", "\n", @@ -316,21 +386,26 @@ "metadata": {}, "outputs": [], "source": [ + "grid = Grid(\n", + " width=50,\n", + " height=50,\n", + " n_rows=2,\n", + " n_cols=3\n", + ")\n", "im2im_state_dict = {\n", - " 'im_height': 500,\n", - " 'im_width': 500,\n", - " 'label_width': 50,\n", - " 'label_height': 50,\n", - " 'n_rows': 2,\n", - " 'n_cols': 3\n", + " 'im_height': 200,\n", + " 'im_width': 200,\n", + " 'grid': grid\n", "}\n", "\n", + "output = OutputImageLabel()\n", + "state_to_widget = LabelStoreCaster(output)\n", + "\n", "app_state = AppWidgetState()\n", - "# Argument 1 to \"Im2ImState\" has incompatible type \"**Dict[str, int]\";\n", - "# expected \"Optional[str]\"\n", "im2im_state = Im2ImState(**im2im_state_dict) # type: ignore\n", "\n", "im2im_ = Im2ImAnnotatorGUI(\n", + " state_to_widget=state_to_widget,\n", " app_state=app_state,\n", " im2im_state=im2im_state\n", ")\n", @@ -346,7 +421,7 @@ "outputs": [], "source": [ "# hide\n", - "im2im_._grid_box.load_annotations_labels(label_state)" + "im2im_._grid_box.load(state_to_widget(label_state)) # type: ignore" ] }, { @@ -378,9 +453,17 @@ "outputs": [], "source": [ "#exporti\n", - "def _storage_format_to_label_state(storage_format, label_names, label_dir):\n", - " return {str(Path(label_dir) / label): {\n", - " 'answer': label in storage_format} for label in label_names}" + "def _storage_format_to_label_state(\n", + " storage_format,\n", + " label_names,\n", + " label_dir: str\n", + "):\n", + " try:\n", + " path = Path(label_dir)\n", + " return {str(path / label): {\n", + " 'answer': label in storage_format} for label in label_names}\n", + " except Exception:\n", + " return {label: {'answer': label in storage_format} for label in label_names}" ] }, { @@ -489,8 +572,6 @@ "outputs": [], "source": [ "# hide\n", - "from ipyannotator.mltypes import InputImage, OutputImageLabel\n", - "\n", "imz = InputImage('pics')\n", "\n", "lblz = OutputImageLabel(label_dir='class_images')" @@ -539,39 +620,10 @@ " # Tracks the app_state.index history\n", " self._last_index = 0\n", "\n", - " self._im2im_state.n_rows, self._im2im_state.n_cols = self._calc_num_labels(\n", - " self.labels_num,\n", - " # Argument 2 to \"_calc_num_labels\" of \"Im2ImAnnotatorController\"\n", - " # has incompatible type \"Optional[int]\"; expected \"int\"\n", - " self._im2im_state.n_rows, # type: ignore\n", - " self._im2im_state.n_cols # type: ignore\n", - " )\n", - "\n", " if question:\n", " self._im2im_state.question_value = (f'

'\n", " f'{question}

')\n", "\n", - " def _calc_num_labels(self, n_total: int, n_rows: int, n_cols: int) -> tuple:\n", - " if n_cols is None:\n", - " if n_rows is None: # automatic arrange\n", - " label_cols = 3\n", - " label_rows = ceil(n_total / label_cols)\n", - " else: # calc cols to show all labels\n", - " label_rows = n_rows\n", - " label_cols = ceil(n_total / label_rows)\n", - " else:\n", - " if n_rows is None: # calc rows to show all labels\n", - " label_cols = n_cols\n", - " label_rows = ceil(n_total / label_cols)\n", - " else: # user defined\n", - " label_cols = n_cols\n", - " label_rows = n_rows\n", - "\n", - " if label_cols * label_rows < n_total:\n", - " warnings.warn(\"!! Not all labels shown. Check n_cols, n_rows args !!\")\n", - "\n", - " return label_rows, label_cols\n", - "\n", " def _update_im(self):\n", " # print('_update_im')\n", " index = self._app_state.index\n", @@ -583,14 +635,17 @@ "\n", " if not image_path:\n", " return\n", - "\n", + " tmp_annotations = LabelStore()\n", " if image_path in self._storage:\n", " current_annotation = self._storage.get(str(image_path)) or {}\n", - " self._im2im_state.annotations = _storage_format_to_label_state(\n", - " storage_format=current_annotation or [],\n", - " label_names=self.labels,\n", - " label_dir=self._storage.label_dir\n", + " tmp_annotations.update(\n", + " _storage_format_to_label_state(\n", + " storage_format=current_annotation or [],\n", + " label_names=self.labels,\n", + " label_dir=self._storage.label_dir\n", + " )\n", " )\n", + " self._im2im_state.annotations = tmp_annotations\n", "\n", " def _update_annotations(self, index: int): # from screen\n", " # print('_update_annotations')\n", @@ -620,14 +675,19 @@ " @debug_output.capture(clear_output=False)\n", " def handle_grid_click(self, event, name):\n", " # print('_handle_grid_click')\n", - " label_changed = self._storage.label_dir / name\n", + " label_changed = name\n", "\n", - " if label_changed.is_dir():\n", - " # button without image - invalid\n", - " return\n", + " # check if the im2im is using the label as path\n", + " # otherwise it uses the iterable of labels\n", + " if isinstance(self._storage.label_dir, Path):\n", + " label_changed = self._storage.label_dir / name\n", + "\n", + " if label_changed.is_dir():\n", + " # button without image - invalid\n", + " return\n", "\n", - " label_changed = str(label_changed)\n", - " current_label_state = self._im2im_state.annotations.copy()\n", + " label_changed = str(label_changed)\n", + " current_label_state = deepcopy(self._im2im_state.annotations)\n", "\n", " # inverse state\n", " current_label_state[label_changed] = {\n", @@ -668,7 +728,7 @@ "anno_file_path = construct_annotation_path(project_path)\n", "\n", "app_state = AppWidgetState()\n", - "im2im_state = Im2ImState()\n", + "im2im_state = Im2ImState(grid=grid)\n", "\n", "storage = JsonLabelStorage(\n", " im_dir=project_path / imz.dir,\n", @@ -702,11 +762,11 @@ "metadata": {}, "outputs": [], "source": [ - "# hide\n", + "# # hide\n", "\n", - "# \"Im2ImAnnotatorController\" has no attribute \"index\"\n", + "# # \"Im2ImAnnotatorController\" has no attribute \"index\"\n", "i_.index = 2 # type: ignore\n", - "i_._im2im_state.annotations" + "dict(i_._im2im_state.annotations)" ] }, { @@ -717,7 +777,7 @@ "source": [ "# export\n", "\n", - "class Im2ImAnnotator:\n", + "class Im2ImAnnotator(Annotator):\n", " \"\"\"\n", " Represents image-to-image annotator.\n", "\n", @@ -730,8 +790,8 @@ " def __init__(\n", " self,\n", " project_path: Path,\n", - " input_item,\n", - " output_item,\n", + " input_item: InputImage,\n", + " output_item: Union[OutputImageLabel, OutputLabel],\n", " annotation_file_path,\n", " n_rows=None,\n", " n_cols=None,\n", @@ -742,23 +802,31 @@ " assert input_item, \"WARNING: Provide valid Input\"\n", " assert output_item, \"WARNING: Provide valid Output\"\n", "\n", - " self.app_state = AppWidgetState(uuid=str(id(self)))\n", + " self.project_path = project_path\n", + " self.input_item = input_item\n", + " self.output_item = output_item\n", + " app_state = AppWidgetState(uuid=str(id(self)))\n", + "\n", + " super().__init__(app_state)\n", + "\n", + " grid = Grid(\n", + " width=output_item.width,\n", + " height=output_item.height,\n", + " n_rows=n_rows,\n", + " n_cols=n_cols\n", + " )\n", "\n", " self.im2im_state = Im2ImState(\n", " uuid=str(id(self)),\n", - " **{\n", - " \"im_height\": input_item.height,\n", - " \"im_width\": input_item.width,\n", - " \"label_width\": output_item.width,\n", - " \"label_height\": output_item.height,\n", - " \"n_rows\": n_rows,\n", - " \"n_cols\": n_cols,\n", - " }\n", + " grid=grid,\n", + " annotations=LabelStore(),\n", + " im_height=input_item.height,\n", + " im_width=input_item.width\n", " )\n", "\n", " self.storage = JsonLabelStorage(\n", " im_dir=project_path / input_item.dir,\n", - " label_dir=project_path / output_item.dir,\n", + " label_dir=self._get_label_dir(),\n", " annotation_file_path=annotation_file_path\n", " )\n", "\n", @@ -771,22 +839,47 @@ " question=question,\n", " )\n", "\n", + " self.state_to_widget = LabelStoreCaster(output_item)\n", + "\n", " self.view = Im2ImAnnotatorGUI(\n", " app_state=self.app_state,\n", " im2im_state=self.im2im_state,\n", + " state_to_widget=self.state_to_widget,\n", " label_autosize=label_autosize,\n", - " has_border=has_border\n", + " on_navi_clicked=self.controller.idx_changed,\n", + " on_save_btn_clicked=self.controller.save_annotations,\n", + " on_grid_box_clicked=self.controller.handle_grid_click,\n", + " has_border=has_border,\n", + " fit_canvas=input_item.fit_canvas\n", " )\n", "\n", - " self.view.save_btn_clicked = self.controller.save_annotations\n", - " self.view.grid_box_clicked = self.controller.handle_grid_click\n", - "\n", - " # link current image index from controls to annotator model\n", - " self.view._navi.on_navi_clicked = self.controller.idx_changed\n", + " self.app_state.subscribe(self._on_annotation_step_change, 'annotation_step')\n", "\n", " # draw current image and bbox only when client is ready\n", " self.view.on_client_ready(self.controller.handle_client_ready)\n", "\n", + " def _on_annotation_step_change(self, annotation_step: AnnotatorStep):\n", + " if annotation_step == AnnotatorStep.EXPLORE:\n", + " self.state_to_widget.widgets_disabled = True\n", + " self.view._grid_box.clear()\n", + " elif self.state_to_widget.widgets_disabled:\n", + " self.state_to_widget.widgets_disabled = False\n", + "\n", + " # forces annotator to have img loaded\n", + " self.controller._update_im()\n", + " self.controller._update_state()\n", + " self.view.load_menu(self.im2im_state.annotations)\n", + "\n", + " def _get_label_dir(self) -> Union[Iterable[str], Path]:\n", + " if isinstance(self.output_item, OutputImageLabel):\n", + " return self.project_path / self.output_item.dir\n", + " elif isinstance(self.output_item, OutputLabel):\n", + " return self.output_item.class_labels\n", + " else:\n", + " raise ValueError(\n", + " \"output_item should have type OutputLabel or OutputImageLabel\"\n", + " )\n", + "\n", " def __repr__(self):\n", " display(self.view)\n", " return \"\"\n", @@ -795,6 +888,15 @@ " return self.controller.to_dict(only_annotated)" ] }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "! rm -rf ..data/projects/im2im1/results" + ] + }, { "cell_type": "code", "execution_count": null, @@ -827,7 +929,7 @@ "metadata": {}, "outputs": [], "source": [ - "# hide\n", + "#hide\n", "im2im.to_dict()" ] }, @@ -837,8 +939,32 @@ "metadata": {}, "outputs": [], "source": [ - "# hide\n", - "im2im.controller.debug_output" + "@pytest.fixture\n", + "def im2im_class_labels_fixture():\n", + " ! rm -rf ../data/projects/im2im1/results/annotation.json\n", + "\n", + " proj_path = validate_project_path('../data/projects/im2im1')\n", + " anno_file_path = construct_annotation_path(\n", + " file_name='../data/projects/im2im1/results/annotation.json')\n", + "\n", + " in_p = InputImage(image_dir='pics', image_width=300, image_height=300)\n", + "\n", + " out_p = OutputLabel(class_labels=('horse', 'airplane', 'dog'))\n", + "\n", + " im2im = Im2ImAnnotator(\n", + " project_path=proj_path,\n", + " input_item=in_p,\n", + " output_item=out_p,\n", + " annotation_file_path=anno_file_path,\n", + " n_cols=2,\n", + " question=\"Testing classes\"\n", + " )\n", + "\n", + " # force fixture to already load its children\n", + " im2im.controller.idx_changed(0)\n", + " assert len(im2im.view._grid_box.children) > 0\n", + "\n", + " return im2im" ] }, { @@ -847,7 +973,19 @@ "metadata": {}, "outputs": [], "source": [ - "im2im.view._grid_box.debug_output" + "%%ipytest\n", + "def test_it_doesnt_share_state_with_other_annotators(im2im_class_labels_fixture):\n", + " other_im2im = Im2ImAnnotator(\n", + " project_path=proj_path,\n", + " input_item=in_p,\n", + " output_item=out_p,\n", + " annotation_file_path=anno_file_path,\n", + " n_cols=2,\n", + " question=\"Hello World\"\n", + " )\n", + " assert other_im2im.app_state.index == 0\n", + " other_im2im.app_state.index = 1\n", + " assert other_im2im.app_state.index != im2im_class_labels_fixture.app_state.index" ] }, { @@ -856,21 +994,46 @@ "metadata": {}, "outputs": [], "source": [ - "# it doesn't share state with other annotators\n", - "im2im.app_state.index = 0\n", - "\n", - "other_im2im = Im2ImAnnotator(\n", - " project_path=proj_path,\n", - " input_item=in_p,\n", - " output_item=out_p,\n", - " annotation_file_path=anno_file_path,\n", - " n_cols=2,\n", - " question=\"Hello World\"\n", - ")\n", - "\n", - "assert other_im2im.app_state.index == 0\n", - "other_im2im.app_state.index = 1\n", - "assert other_im2im.app_state.index != im2im.app_state.index" + "%%ipytest\n", + "def test_it_activate_button_on_user_click(im2im_class_labels_fixture):\n", + " im2im_class_labels_fixture.controller._update_state()\n", + " buttons = im2im_class_labels_fixture.view._grid_box.children\n", + " airplane_btn = buttons[0]\n", + " airplane_btn.click()\n", + " assert im2im_class_labels_fixture.im2im_state.annotations[airplane_btn.value] == {'answer': True}\n", + " assert im2im_class_labels_fixture.view._grid_box.children[0].layout.border is not None\n", + " for button in buttons[1:]:\n", + " assert button.layout.border is None" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_disables_grid_menu_when_app_state_step_is_explore(im2im_class_labels_fixture):\n", + " assert len(im2im_class_labels_fixture.view._grid_box.children) > 0\n", + " im2im_class_labels_fixture.app_state.annotation_step = AnnotatorStep.EXPLORE\n", + " assert len(im2im_class_labels_fixture.view._grid_box.children) == 3\n", + " im2im_class_labels_fixture.view._navi._next_btn.click()\n", + " assert len(im2im_class_labels_fixture.view._grid_box.children) == 3\n", + " for button in im2im_class_labels_fixture.view._grid_box.children:\n", + " assert button.disabled is True" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_loads_grid_menu_when_app_state_step_is_not_explore(im2im_class_labels_fixture):\n", + " im2im_class_labels_fixture.app_state.annotation_step = AnnotatorStep.EXPLORE\n", + " im2im_class_labels_fixture.app_state.annotation_step = AnnotatorStep.CREATE\n", + " assert len(im2im_class_labels_fixture.view._grid_box.children) == 3" ] }, { diff --git a/nbs/08_tutorial_road_damage.ipynb b/nbs/08_tutorial_road_damage.ipynb index 309fecf..50934a0 100644 --- a/nbs/08_tutorial_road_damage.ipynb +++ b/nbs/08_tutorial_road_damage.ipynb @@ -4,24 +4,22 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "# Tutorial: Road damage" + "# Road damage - Iterative annotations on road damage images" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "## Iterative Annotation Process with ipyannotator\n", - "\n", "This tutorial demonstrates how you can build an annotated dataset for road damage classification without ever leaving the\n", "jupyter notebook / lab. We do this in three steps:\n", "\n", - "1. Use bounding box annotation to crop the orignal images.\n", - "2. Group the damage type in groups using classification labels.\n", + "1. Use bounding box annotation to crop the original images.\n", + "2. Group the road damage types in categories using classification labels.\n", "3. Refine the inital class labels in a supervision step.\n", "\n", - "This steps can be applied iteratively for practical applications and significantly speed up by integrating the predictions of imperfect machine learning models.\n", - "For example we might train an image classification model on the first annotations and refine it's prediction on new data to increase the training data and repead the process again." + "These steps can be applied iteratively for practical applications. By integrating the predictions of imperfect machine learning models, the process can be accelerated significantly.\n", + "For example we might train an image classification model on the first annotations, which then refines the prediction on new data. Therewith, the training data size is increased, then repeat the process." ] }, { @@ -56,8 +54,7 @@ "source": [ "## Get Road Damage Images from BigData Cup 2020\n", "\n", - "First we need to retrieve some images from which we can build or data set. Fortunately the [Global Road Damage Detection Challenge 2020](https://rdd2020.sekilab.global/data/) provides\n", - "a freely usable images that we can download (Please cite [the paper](https://github.com/sekilab/RoadDamageDetector#citation) if you use this for your own work)." + "First we need to retrieve some images from which we can build our dataset. The [Global Road Damage Detection Challenge 2020](https://rdd2020.sekilab.global/data/) provides images that are free to use and can be download. (Please cite [the paper](https://github.com/sekilab/RoadDamageDetector#citation) if you use this for your own work)." ] }, { @@ -66,7 +63,7 @@ "source": [ "### Installing via `pooch`\n", "\n", - "For this, we provide a Github repository with a subset of the Global Road Damage Detection Challenge, containing only the images from Japan." + "For this tutorial, we provide a Github repository with a subset of the Global Road Damage Detection Challenge, containing only the images from Japan." ] }, { @@ -81,7 +78,7 @@ "\n", "github_repo = pooch.create(\n", " path=pooch.os_cache(\"tutorial_road_damage\"),\n", - " base_url=\"https://github.com/palaimon/ipyannotator-data/raw/master/\",\n", + " base_url=\"https://github.com/palaimon/ipyannotator-data/raw/main/\",\n", " registry={\n", " \"road_damage.zip\": \"sha256:639b3aec3f067a79b02dd12cae4bdd7235c886988c7e733ee3554a8fc74bc069\"\n", " }\n", @@ -102,14 +99,15 @@ "with zipfile.ZipFile(road_damage_zip, 'r') as zip_ref:\n", " zip_ref.extractall(path)\n", "\n", - "path_japan = path / \"road_damage\"" + "path_japan = path / \"road_damage\"\n", + "path_japan" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "## 1) Use bounding box annotation to crop the orignal imges.\n", + "## 1. Use bounding box annotation to crop the orignal imges.\n", "\n", "We can now use the BBoxAnnotator to quickly inspect the available images." ] @@ -158,42 +156,32 @@ "bb" ] }, - { - "attachments": { - "road_damage_one.png": { - "image/png": "iVBORw0KGgoAAAANSUhEUgAAAgoAAAKACAIAAAByiMe8AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAgAElEQVR4nOy9SZMcSXIuqGqbrxGRmQAKqEK/arJJCmdIHuYw78L/fyRFSAr5DtPN19VLNYDMRGbG4ptt+g6fuyGqpg8z3RSZOZRLSUkgY3FzMzXVTz9djEWEfrp+un66frp+un66fnip/68H8NP10/XT9dP10/X/x+sn8/DT9dP10/XT9dP1Ry5z/Y8/fPjtPM+n0ymEwMy3t7cxxmVZ/BK11jHGnPPd3d3xeLTWns/nrunHcbxcLv/2b//2hz/84XK5EFHTtq9evxERERnHET84DENKSSnFzKCzmJmIrl/jBb5IRCQKfylfuf48PqbUat5yzsyccyYiY4xzbr/fv3r1qt93fd9baz9//vzdd9/lnG9ubqx1VbNj1iEE/KYxRms9DINzzjn313/91//0T//U973WelkWpVRKyTmHO8YYtdbMrLVOKS3LYq0dx1FrPY7jbrfDb47jmFLquq7rusfHR611TImIuq7TWhtj+r7/8OFDCEErZa0lorZtY4wpJWPMNE3TNL19+7Zpmufn53Ec7+7uQgg55/P53HXdzc0NbppSenx8xINXVfX27duqql5eXh4fH8dx/Prdz7TW8zzfvrpr+/7p6clWbp5niimlVNf13d3tzc3+w4cPzy+ftdZGO++9tXae55yzc05rfTwenXPPz8//8A//8OnTp7JGwftlWd6/fy8i5/P59vb25eUFYlDXNR5Ea11V1eVyEZHzcHLOVVW1LIuIaK3xyDf720+fPuWcrbXOOaWUUkpETuczfmSe57qu53l+9+5dzllr+/LyYowxxnz+/NlaW9f13/3d//bP//zP1tqUUt/3VVUdj8cQQozRGAPZg+Q0TZNz9t5DACDYKSUs3GF3o7W2Th8Oh5ub/d3dXd+31toQwjRNr1+//vjxI/bF6XS6u3uz62/meY4pfv2z97/81a9cVSmtHv7w8e///u/7vn96emqa5nK5/Pu///vf/d3/Ljn+j//xHzHGb75517btze3h9vY25zhfht1uh5mpqoqZ53kehiFJ9N5j/JB2a+1hf8ui5nlmZry72+1+//vf//a3vz2dTufzWUS++eablBKk0SgjIssS6rrW2mqtmbSIxBzK3mdmzHlKyRBrrcsWw7wREWX8Ta43b6bESsomTSnhW/jnjza7iJAY7B3cF5tLa2YlSimjXUpJa5uziFASEc74OkaSc845UybNqmmaGKNzbp5n/OASAyuFEaaUyn2dNnqbQFxa63lZlDXlk1pryJ611hjz/PwcQvDe49YikiUL5Wma8OHyyMMwQIyhJLF9mNk51zTNMAwiYq3NOWutc05KCzOllOZ5hvh1XTfPs2TVdZ1zblmW3W4XQqjrum6cUlzXdVVVSqllWYhoWZa6rr/++usYYwhBKTUMA2bAe19VFYYHxYjl+Ie//9s/1zxgz+AxqqrKOY/jGGMMPvV9H2MUEb9dwYdgwjAMT09Pnz9/nudZROq67voemy2lNE3TPM/FMEB9Y+jYjcaYH4nRlYVYhbIYj2IYyl+K2YBw40Vd1/v93jmHZTbGHI/H+/v7GKO1VimVc04pxxggECmlqqowfjzCOI6wNymlEAIGbIzx3scYMcicMyQbCw89WMbPzNbaGGPf99M07Xa7cRyJCKOCQRrHEXJDIiGEqqogplVVWWuHYcCv1XWN2YPlgDTgQSAx+GU8V13XXddh8n8g2TnnnCFbWB2JUSlljNnv91gF2L8QAh4KTwdhgD6F7fTeY9gppcV7aCjMPEwahgqEsaobYzBFIgK50loXi461w6ZqmqasMkY+TVPf98YYSE6MEU8K64IXInI4HMZxyjljfjADePCy+SEh+LxzDis4z/M4jnVdL8vSdR3e+tnPfnY6v9R1jQmBGTufz03TYAUhS7vdru/7pmlOp5MyOoQgIsM4eO+bpnl4eHh+fsYjf/r0aZ7nZfGKxblaKb+qOSJmjjHnnKuqcs7hAZdlwTyXTV5GjnnOUcZxxFS3beu911rf3NycTidM1LIsIYQ3b97EGHPM+OI1oiIiiG4BVbRZysicMzEzMWmjscEkC/G6SzN+bd16ipQIkRCJCGvNRDlnfIaVKrtVMTOrnJm3X8AAmElbHaNnZjZaa4WvMikmSXndcUUbOOcokwiRUqx1zFmYSSmllGEqy11ADIaaf6guJGe5UjsQNswPpgLCf23nRHLKCVq4YA7sC8g5JhOIChMeY5znGUYaeoZIp7zgl/u+H8cRqK6qKqOrvu/xxbquiaiqqqq2zISb4mexswAocccycmjanHNRU3i6Yoz/314/MA8QKeecMQbGbRgGqFcgUGY+n8+n08k51zYtMy/Lcn9/D12GXQQ1uiwLpmaaJixYEfGis34EQ671ftH+18ajXD/6Oj6AaYKahnnruq5pahE5Ho+YL+gpkVQrJRKh44Dl8bNENAwDEBlteqpMOnS39x4rgZviNyFhVVXVdY1HxlM3TXM+n/u+H4ahrmtjDH4Ezor33jknORNR13XjOFpr27at6/rh4QHK0TkHHIHHXJYFo4oxlkmDCsaPE9G1SSYSIlZKhRCqpun7PuY0zzNMsff++++/v709eO9Xs8dZKYU5DCFgPvECIAXgoO97Zs4pAbs1TQOzdw2soKyNMUWojHZN3UG+gQ9ImIkA04CaY4x1XWO98Bfa9COeSCmtlC7bGMNr23YchzIbWCAigiqnVQtHtTlqUL4wANM0tW1LRH3fhxBubm72+70Pc1VVWvO7d+/6vk0ptW3rnMN4rLV43sPhYE1tjFFaPz8/p5RCDMfj8evXXz0/P2Ma4Ui9efNmt9st83g4HC6XyzBM2P9938/zHH2A4Snjr6qKmIilODcFdkgW7/3lcsGi3N/ff/jwATOMO64fE4HTQxtsstYqpUWEhFYFvJnMogfLhiqQa0P9pLbdyhucJyJRUgDT1XfFOV12JW9AW6lVzQFrQ98ZY0QyEMl2O2Jat3ZBXbQ5+kopUpJzZiWKKOdkrAJcbqsaAIVg0XLCOLVhZmLeTKxSOSdWYgzAk2jNcB2MUcaoEAK2Pm4bY8AQmEUpSikRZaLMzN7DcclEOcaUc27bfYxxmkateyLVNFXTVEqprmvqulZaxbgUKLbb7eBPxxirqoaL/EWGc4iRlVpnDz4rXocQwE8UVQMhgcHAomOPAHb8PzcJ19cPzMPpdMo5wzEnoo8fP57P53fv3kFQlFK4DTaJc84v4dOnT99//z3wI7zjeVpypst5TCmFEKdxgXTmvDqqKUoMAUuYYjRWFRvIPyCX5I+aB0zEl49tErnKgdbQwnDHcqbn5+eXl5Mxtm1hBhQzWWunaS4qhoigxYhoWRYQSgACxhjAdiBNIN9iLTASYAHMUkF/cAiapqnrGhYIzElKCWYArkBVVawU7ASAKnA0BgandZomYHygbCICQAAEhrbFqIA68Ud8oKqqYRi8Xzrudvt+HMdpmFxllxCnaarr+nI5T9OAzeacA4NQxA6YGls05/zw8IAfx1AhjHDDMfmQe7hZmBk8AvTUfr/HbMC5gVBBg8OiE9E4jvAwoN/ruh6GAW8REZQIGBLMFUCZc+7h4RlKHA5+0zRw3TCStm2vBayqqhDCPM/zPLdta609HA77/T6ltN/vsXBYXCIqo23bdpom2Dyo8t1uN08hxmi1clXd9/3L8QhqblmWtm3hI+K+l8vFOdf1ex/COA1JyMcoxMZao1RMEpMn1kobH2YiVVVNzolkCj4U9JMtLYvPOcMkhxCWZdHbBUJ4GAbsFLCRlauKyTTGFUsDtVKUO16sOhqYQmtWqijsLEIMD4LwUSIiysTbv+Fxpqy0Yq2YSPMKs7AKSmvObK3lTZkQs1IqxCWTMJO2hlIC95UzkZDWKyuAtYDwQFVnEQzpmuqBBhAR7DVMGpATlh7LB9CATYfZKMxS0zRL8KyV8Mpa28qFELxflOJlWbE/EUFQ8eTGGKw4CKWmaV69egUUAvAB31RpZl7NNt7FsCFOgLB4wKJmYR6gJHnjtKEWsM0htMuyACYW6Fw85j/5+oF5gJIq6Bg3U0phi2KiAbe11lny73//+0+fPp3PZ6UU/O4QArMKwRNxSlmEqqrGShhjCwQobjUzCeUCOcvzi4hkLg+JZaYN5hScgvFAaYKEgWLFIjVNcxnH4/FExE1TaH0bQsh5NS140sIYYjanaSpIFvctXAQgLfz6Qj1576F9oCILoQE1ARwH4YNuXfd5zsVDtNYWWbHWLstye3sLpwG8QREOKCnMCSgRrTVGjqeYpul4PBYCSijNy0gMoJSG4TyOA56aN7YBdkIpnuc5+JUuAxooqrzQU5BC2Qix4D2ksERuYKVgR/EZ2rivqm7n2VdVNc+XqmowLdbaZR4LzHl4eMAMT9PUdh0UPVYE01XX9fk8xBhfXl6w+ptbrWF0MWnLsiAKhb9AnguXCCHE6sAywSHYnM7GDc57b4ye51kpAlaIMULFQ7WtMwzGRsRoLSJ1XWXJWhgqCQTUOI7v379vmmaYJ+0saa2tU8ZNiz+PQ1s3lGQYYBclxiRCINaJVh4POxGK4OnzM8D1tcteVZUxBq6Jc26apsPh8PT0dDgcCqiCv6iUIuKcM2vebMaqU+Bl8kZHFDcduwB4/trjV0plYSJiYmbKKSs2xjnso5SSVivNYvQaw9BOA7hgLXBHow2rdQAxRpGMCARtKLGwkatmZ+n6NqUEygraGRv55naPbQg3F+8CgkDgjVUhBKUp56w0aeEQkxBniSRKSKUcUgrOmZxXxzel5FxjDIfoW1MDA2Had/sOso1bgOgTSs456zSxSSkpTc45oWSd1VqLZDwFJvnaUSMS56xSX8AQZBUgAH+ULSAUYwQ3JSLTNEGksdBQ0W3bqity7881Dxj0NSlcJMM5d7lcmqYpEQW/+F/+X7/6/vvvIbJN0wCOxUjTvIzjCGSH3UtEELvyy0XCWJFIgtgVjJNzJlGwE7Au5fN4t7gOUHMxxhgjdohzbrfbHQ4HZp6nqdgzuHIikmLCcmJmCxEBIYMgQixk847hDi/LUpi3wrqUKBD8d4ykeHMfPnw4HA4wCfhMiWEYY7quU0qNwwDoAf0F/DsMQ9/3zjlQedDgJVgCC1SAzzVFME1ToWJyzjAVGGHxhEo0BXO77p/kp2li0iCycs5+Cy1cLhcE6otdLwITvIfsgh2iLU2ANlWCwePDdd1M07QsHpMBp3Ecx/Fyho8yTROMPeJ1MaWClaDiAYcL1BURELsfP34MYSmguLAx8kPOPW8BGIglb6EdY8zt7S3WdL/fY/zLstT1DpIft+tyuUD7wAUhIoADH3wWyTmfL5emac5PLwjI/83f/A0G3HXd+XwmpbUySumqamCSU5Tn5+Oh70JK56enGOPNzQ0RCXNMSbFm1tZWIuJcrbVm1n3fxxCJCEzmzc1N27ZwKBF4CyF8/vwZ9imlxIpFRGszz3NVaRGBqhbKZWb4KpKZSZgUkSTJRERMQl/YnrIZEW8gZlar/CvFRCRESXJVVdlLJtFGp5RYKxFRWiXJpHj9T5iESHGW7KxJKbFS2piUxGhLpEhEkQB/AElgvYRy3TRXLpEhIt6QE5YVE7IaJ2t9CCSitdZKCVGMURuDgB9gEFRtSinEqKxpK5eZqrZRa/BcG6dTrEQEGwRICHgOYnA4HKqqgsMKxx3G79q3gHkgIjwLPHVZgy0F9atNaIlZiLhs86JqRAQkR9FOtMEF6BbotxIj+VOMw4/MQ2EtETiFokkpgSFBUkQI4cOHD9M0LcsCVqGqqpubG2h/pVRMISXR2ta1KXpcRLT+gjuu1AcpJVlW7VwgrYjktHLNRUdfOxD4Ol5gjuq6PhwOdV03TQNzNQzD+TyQqK7dWVMtFPwSU0rM62wWhwDRAiD6aZqAx7FyRc9KYX43640wAEAl9AVsJxFBC5fQH6LTMSUouHme+74vIy+eb95C+lhyYPZ5nhE5L7xTgQyYxiJeeJzii2xzRSFEZo7JpxRyjsYoZso5eR+IyBgtksZxjGn9StqyBmAj8VOQM8wbUE/OeZlnJG7xlviRr4LGeF3iCvBC8Auw5ZiKYbikGDrd4tH2+z2MUM4ZJgHIHbiMiEKImBnaeE5MGnINsBCY+QIskA5U1zUzj+MIMw9YA0KJmREDUEq9vLx472NavaLvv//+dHqp6xrpcFjWEALcDgjt5XJp2oZEcAs4cD//+c/btt3v95fLBZI5jMNlHJ+en1MWYrqM027XZZJxmpw2p9MFvPnd3WvgYmOdUVQca9wR02vNmmRRnLyiEEEOn89nOAoicrlc6rpm1l84FhKllGxs/rYZV/RG6kuiR7HEBMX/Ry7e8uN/gFO9jyJIOojwV0KIWNhimyEnWmsiVko757Syy7IQKa21UkZrregHlBEWXWm2TlPIWitXGSKa55mY6tptAxYEpJnJGKUN97sWrgkR2axztk3TKGKiDAiSko0xphRIiTLGWlvXDksMHCZZkcBZYWOMUmyMLpMPVrmqqhC8c7Zp4GTkEDyiF/M87fc7ayvgDEwC1CY2/uGwQyoQoNsWSTIiVNYIuBMyrLZ8H2wl/BPyWYzTlV/yp1w/9h4ghTnnT58+IQIZY7y5uYsxNk1jjHl4eBiG4Xw6nY6n4/HY9z3QFnQBgN5mFYSZoNiLb4gbQay11lorkUibo6quctd4w6HQSsU8XD+t2i69hZedczc3N7vdLsZ4PB69X6BZsJ+xl5i1a2rgUCKCg6K1bpoG0BIWGIoJ81u0HpIsQfsUAgf6tLA9y7KI5GLnC00XYkReUHHeMeamrh8eHpxz8A8wMKVUztH7WSQTyeynftfdP4zWOCJq2/ZyuSDcXVUVgkZQ0Gv+zDCA8U8pE6mu65nU+TyklBXr6EFPr0RZ01QiEmNKMTGvnjhMXZFgGDbgMhEB9cHMfvEprpOjlEKu3grZjHHOjeMYQqiqZp58lvF8PoOgQyqetfb52bd1XXITc86Pj49wq1PObdtiCTDhKaUQfIgRc4hMASjNeU7WGtmoS9CDl8sFzC/sHFZQbUGat2/fdl2HHAoiAlMX4phS1lrJlqd3Pg/LEvq+77o8z8vpdAE609rmzAgSfvX2q6fjizGmqer7+3tI2t3dXdM09/efvv76m2WZiKSqqzdfvYYIEVHf903TXs7DME0xhA8fP/zVX/3VZRyUUjLSNIxNZZ2rYO8BKYhonibAfygR2EillDJ6J2sSHeIuiMMN6cLceD87V2nNOecsiUTgD6gtKF0csqRKrkeG+t+g7R+xDyKSsLWvsoAAm5ZlSSLKGKxFEmHFxq5jBgwNwQ/DmJJoa7KwUqydda5OMeVMldGKCI4a8AG2hnXaOIecpSSSc15C6LoupMhCILrLWjdNw0YjRGE259t7b6xt3JoueD6fsfVERBmdSZJkZbTS2lmXZQ07l52rt4A5FhoLURA9WJ1lWaDHrbXTNIHw3O/3106t1noYhnWQ/IVOL0kQOWeEgfAuDDw0JL5SuMGCXHe7XaE3iAi44b/APGD2CxRFluTd3Z1z9nicXGUU6ftPn37/3e+maRrHEd47+G647eM4pgxZiteQP6UMxEo/Si8TIWLJrJQiYSIKIW6u1pdELtowMm25sMWTwFuUs1FkNMfkUw7aqss4HYeTdlZrnUjY6KaqRUTyaqLgEtJGbhYXjK+yMvKWvQppA4I+nU5aa8QeiQiRrpwzLAcRGaNijMyaWZjV+XyWzDFkoIYSTYUAKaWctYfDASlAAMjAHdbqaRr6vjldzj75elfTM81pqtsq5jDOAykRzotfYg4UlVbGWKj7ZpyWm5sb7711jfdeaXu5TJfLBN8WYmStrSpHJHXdDMNgbc0U67oOIeUsIYQYcwjJGCfCMSZsae+9MZZZ5xzbpknpxVpbuVYrhOw4BK+MXkJoqy7G+PR8rOsmCw3TpK3RGjmsJsZwOBzmeTbGTMtirZ2XQKz73e7h8cmE1HY7YxyRWpaQUmbWInw6XVhrZfXldGnqRhZyxkwjzFhkVk3TIfLhXD2OYwjJGDMMw5s3b72PMeJ3SGvbNLbrdinlaVqWxQ/DRKSOx9PiLymlqmq899ZUKeUQkqvsze2beQnD6HOWLCIkIVLKy/H8cvf6tqrdPnUikvyyjMObN2+7rnPOpBSM5bZz8xx3u9rWVZL88cM9c3Kufvv27bKEtunFx6eX5xyFRfl5IaJ5HIho37ewptALK1qsayIk/PwAV4li19SZZLwM/X6HLDjvfaYUxU/LZCudSDLlLKJFczRGG4opp6SNZuKUMzNRSClGY0wKQWkNSsBoU7YbhEdWroZLeKmE3GBdsmbnKl4T7ULtjDHaVaaAbmPMJSyUQ982ObEIJZ+UKIlJUjJaK0U5RWa2TrOStqtB5deNq+qKctoYfC3JNZVLKSutjLOkeFrmmCJnrqjWko02khMTOaO9T5U1WnHM0TizLIuwCEsSpCOm3jYkopSqEEdxLrBXUmPDaq1LzI+IalcVeJRCrKwTIykl4/TEY992RukooWtazSosPsQ1MQdgDnGjrYzJwn+bZ18iE7BnpUgLnkHlqrqqcs6ucta5kFOMUYhq88XFhJkBOP4vMA94bMx427YI8MIADMPw+fPn08v5u+++wz93u10mKomehR1LG+QvyvdagiFeoOyvOSKz5RrzliEH0ka2lLtrAyhbEAbWQmstKfZdr40hIYR2n56eiAgRkXIjmPtrE1V+7UfeScH+xQjLlsgML0GuguQrONpIHniOKSG9QYZhqFyjlAIYx+3gjeWcX716pZgvlwszhxBKwkyMCVEKpEIbY16eX4ZxqFxVWCZgNHyXac2XxVSDTD+fzykloAneUo9AsyAZVCmFFFVYR6ih4n8Uq7zb7UCaaa0Ph0PaUsIhqVpr5+wwnFjxMFyMMc5UYEKg/eGSgnWRq6ARSDO8xqqVYHhJl4KnDK8LhKeyxpLDt5qqylngOVWuqptaXaWcgfkUERQ3FL5LbQlmMcbz+RxjRPikqirv52maU4opife+qQHoFJ4ipTzPC34Eo/J+gTCP49j3PaCiMeb9+/fMjHwShNyVorZtTV3FFI12oFCMMfPsL8PFijoeT/Myeu+Voufn592ue3x81MyQYcwzkkcgV3hSaAEsemZac96yvH79eh6nlNLLy0vR10Du8zynFIMPLFapqKMuLNxqhLaoTN7y4qyxRNR1HZLWCrGmlDJurWRk5nmeEbErkKtgaoDiuq6q2pR8ZWNMygmuGG2BMfgW8Bi05uA90n9hkyCcbdtqowuHCUeZmUPwOeeqctM0PT4+QD8ao+F44cMlA6VQkZfLpeSPYCOYLUccUpRiTHGNj+Kp4+a/GmOU1mkrlaWNgwH4g/ghdFEk2QffNA22D/QYYrolcxKhixI6gkGirSplJUu06fs+pWScJWSI1XUWMcSIkiIoUDTzf4F5wGYGsL1cLkgHats2pZhzHobhN7/5zfPzMyWCXrBVpTb2vBAsxFkpna6KU4ry1VtWDxZgK4rh8km9VcxhIQtxVFyHcvFV8JOZbV03baO0Ns4qpZ6fn4/HIxuttOK4BgmQKrPSRObLg2P6Sli1OI/Xwy4DwOCLcNAW7y0+h1rjlutshODned7vblJKKP1FHgviVNh+RmsAXjw7qnmnafrqq9da22UJKUnf7f0S66plYsnrf8Nlapt+GhfFBpunTGDZY8aYvu+PxyN2EUh/7M8SucJalIA5EUEdQwVUVbXb7XizYUopJFnlnJHBZa01xpo1Zz8bQ5v/xyEE8MjPz88YIega7HOo+3mesb1Pp9PNzc2yLCAHpmmCPirU4hq9b2qo+a7rjDbTMGitu65bvFcbow0HH6lNKaW3b98Wc474LcgKJLYi0ui93+/33i+Xy5mIcqYQQt/tQwgprYFxzDAYGxTrMVOJeTjnhmGACEEtIo0d7Gtd17vdLkqunHNvquPxmJJYa8fL1LXd8/0jTMvj46MIEjrodDrFGD9//lzX9evXr2mr/MALKAvYAwwgArJrDb9Bs4Kiickjl7rve+zrddMp3sihtcqk7/tMK3w53OyICKoT27muK5py3ThMHf6uNIskvaYtZGZxzupNpFMKIqIU1TWK4bno5ZIg13WdtdYvUSnS2qDS8Hw+W6uVUpUz1tqqQsJhFElar0KIRHPeooDQYESE8swSfwJAhOAVLkiuCsL1lhQOSy9bpilMMiI3eFjskRIEwvYnESA57BqzpVMWIqjUVxbDiQ8Mw4CK6xLQhp0uZfPwTgDQC1+Cx7G1wbMoo5nZp5hz1kpRykgLQv4etvM4jkR3f6552HLpfqB/c87YRd778/mMt96+fauUCgkoXotwzmSMaxrtnPCanSaFTVNb9B8cFMKtbmtTgUUqSVNQx7QBeXA7ZVSypQATEch6pVTfNlorJFGcz+dxmpZlaWw3DgMRh5CwwDnnKGvW6bV5KLaKtkB0ceTzlsFVZoOvwiHQd2nLmMS3sJzGWGtt03QlCxZZ82YrtkSFizHmdDwW6qCke+Wcn56eCmEFdRljHC6XfbsvcAw6+u7uTmtTHLIStnLOgVuHZsdIyubE3JbcR9pSbLHiAKrGGJQFYIFKJq7eqlKROOT9Ms9zCGu1YInrpJSYCWXDh8NhC4ck3A5ZAG3bDsOA2u9Xr15dLheEPUC/grEVkdPpBFOklPIxSJYY4zxNYfZ1VVlr27apmwauLRxqWBogGGwweH4ofyt2HasAf0JrhSi6bPXS5/PZGPfu3bvN3gfEt96/fz/P8ziuBaFohwAE2nUdBoBoh3U6xth1u77vE0mMsa60tXaaFqXUx4/3ihiSfHNzYOZ5XrTWv/3t49qzIIS2bVNKqP9QSoUQUA0A26C2yyoNchIgT9La/mA6DqfT6fXr13qLcsO7dU1jjAWGhXyu8Xkk/zTN6XSyzoH2dM7VTcOIh8Ly/8cAACAASURBVGutjWnatuyjUvMPNUJbWDtfJW075/q+14aVUsgCgJxgv5fEB0gg0mFi8iklbY0ymnMCjNfWCAn8GIgllKD33vvZObcsU85Ra4RLklJU6oG6roNkYunhmcWt8BB0ergq+Mffb25uoMox4VBHIIpLtP/l5QVZ0UopP40QdeOsylpbE0JgrUixjwFgy1pbEtALCYG5wryByy25kQjrQkHVdb3f7bHQSXKM0ccgInXTDJcBnAFwTCHA/wTb8GPzAD0OGgHaAZ7O/f39L3/5y9/+9rcfPnzc7/evbl7RmrNon56ekHkpX5qrqGWDJ9dmBr9cADIWtRiDomdLSIA3JgqR0rxlGRWTgzvudjuttVGKKBPz7JeU0ul8JiKXaqVUihE8z2rAUZWzXRAFNKIozAN4D75qxlLMUlEoRcJQM5y2/hO8ZdnudvsYE9QHtG1IvuidruswCeWLmHBEsdKWyqmUCWGq69YYNy/LOE3MJkZRyqYET9wdDrdN02GogDBqi5nbrWgWEAZVPEQ0zzMGQFsMBtMLjAzwhXQjIGsoaDjgshXN7Xa7GCITxxhhHrquXTNcsiCHzxgTY1ZK3d7eLstSVa6AILs1Fykri3XBkG5vbxF7x5Sez2dgIvBReWUEs7M2h4RkJx+C2gq84aLt93sk8PR9//nzZ8C0kuMRQgDxArPadV3f987Z3/7u1ymlum6rqrq7u2uaVus1jeTl5QUhqNvbWwgzImS4kJWAHl8vLy8wQsh/3+12SJPXIpWxIiwpcebz+ey0ub25qa2r67rt6pzz588Pz8/PmPPn52fshb7vT6dT27a73Q4xnqJMi0sRrwrpmRlG5d27d0JpmqamaeA4wpbM82ytcdZiUVD2wSzGrSmGxiqhpI0Wztrw4WYnIikrYysmatq9MQYEVMnmUpq01sbonDOxJlpdBKGUc267Wps1SwcKHQAZlEuWmHJo21ZpIspVbYWQx5hTiilFYzQrypK00lrrsPWWAFMKcrJUtBQWLsQAo152btwKNuG5AvBhzwLm85ZOgmwIKLey69XWmQoeSdd1RIQ+YyGsHZ/wrbR1J0JFoaucMUYbXShByD8sFvZmKcgFMwHciUkuCHK/30No9dYHz3u/BE9El2H43a+/Q2EpRovMjuPx+F9gHkreETYM7P+vfvWr3/zmu1/+8peXy6XrOuQC5pQWH8JWLSCCfGWTcxYiEDVlQqF2gS+KasATmq3YB5Dcbk0Fip1QW26W2uIHWGb8BRSBUkqRGK3HeRqHARlRdV07a1NabR78RK21j1FEUsrw/a+jHdfSg9gAb0wXb2kGhcgLW28fQImE7MamKYwZ8vrneSl7OOYAIUBCJ9b1fD4jdQE2A/wJvmK0Aepf2S34TKLIgkUVY+wwjFqbZfEIOOetqxXMGywrMA5MIMQL+fIYA9hV3LeqqpeXF1huBELw4M654/FY8kGxGzdmzIAKg+TgqUOMi1+Mnrz3IlxQD2AH/O4iZs45BADAyyGFCfYM0AlXcdhdU5Pi8+ksIjGmkkvqrN3tdvCrlFIoVy5960piLqAZyhcgnCjou7u76/v+fE5KcVW1mEARqSrXtj2WDE+KLQdtQsQlEoPFwhdfv/6qMJNItl7lXKiqa2YdFh85Ga1v9ocYIwvd3d1lQdavPxwOISzLsiCVHA7E6XQCOK2qNSwHo6Wu+gKVuDERhWU18ON0OZ/Pd3d32AiwbdZaYm2tAwIIIZBipbXTKuZsrbWVs1XVri6C3NzdzfOsrUVdjtqKhOJVQRbRl/R8KC/o1lLVWFA5bbXocDVoy3XGjoaizzllyT4tmXOitMyLUkpYMufZz6V8DzPst35oPsyLn5xzWishJtas1v5UZksHL8HeEmouVDZdxdix02EkCmunrvq1wCtaYjheztba2tas1TzPshaCqGqjlUhxt9vh7k6bQjYgplJ2Lu6F/ZK2rgEwDxAnlBMyszbaL/7h4YGYs+TJL8aYz09P9/f3v/vd75xzX3/99d3d3bt370Tk4eHhv/+f/8efax7gZoKvGMcRDTi/++67T58+grx+dfv65z//eVu1OWe/pM8vL6A+MJVa6ZhSzglETJn3vDVIAHLBFoU3TVfpdLh74W3xd+hNuirbqbZ+hF3XvX379ubmhpm7phFJD4+PSbL3PkTUGCQihJjWnNSyyY1VAHT4i9ryxnjLueQttQBBeESAodAh5XnLeS2VAbQRUyEExB5kC2CGrXjYbYX+YWtnhKmGjkNSAETBWkukJCPLrT5eziIkwlqt0XIRAaGxhbV9gZPYmYiEI1u3JL2huhs0Fz6P7o/Ma7eAQvdjr6JTIQCLtV9y7UG2ppRIVi+bt7h9SoqYc84+ra36in2F8sI8l5QBAEmM4by1aIXCQp9XhAfAy2mt9/t9zCmG2LZtjunV7W3wIYRgtC3EIOr8ryEI4nXg30ERYG+rqwB7XdefPn0AnkVIhpmV0t775+dnmDQYVFpLeTSKgZRSp9MJOgiFaSWWU2ZgGOaUUu2qGCP69uDuCLfknPu+X/zEzJt/oO7u7mirfzwej8fj8Re/+EUIwRh9faO8XUqpmBJtUDdv5SMQdageVIyid1YSVsxKqxhT17fOOmstG426a2La7fvSKMzVbl4mbTVrNqw3s5R98EyE8CQ0qlKMCuuUUFUjdb3GlqdpSgkV0avxwD5lZmvXrUokiKekRE67kAIxLcsCk4ZVCEsAe1OcIWwo52pjFToZV1UFVTOOY4oMOwefTLbMb964KfCceuv5CKiKMUMJ4F2z1eEjpATwMU3Ty/F4czgABGQRcMJ939dNE7wnZrNFIJiZ89o0CfX5RQdChxQCULYeIfqHKUwxxrZtgw/TND09PfkYlFJz8CLy+enp4eHh06dPu90O8gOPBITQn2setFLO2eA99uQ0jp8+fXp+fv74h09hTn3fffX63b47aKXAP5yO5/P5sixeKUb1fEopyUpBYurpqiGi94F5LZ+Geio7Tba4NG01ZWZruOa35jwikkWcdcjdhDHvdx0z1VUVQ6zrpqqmGLIIxyi1siSEBFP50hmJtNHMhDJgunLGU0rEjLUHX0REp9PJWkuUQ0AvvOR9UkrnHJkVEc3THIJ3VSUiYNWIWLHWSjvnlmUAva6NXsKMzWytLSlhwC/Ik8O0wzZUde2XhYjmea6aCmaJiLQ2dssphDjSWgaRSulASfDQWs/zRETOWWN0XVeIAMcYzuczFG7hxMq2gRVE0AKCAbcybXXaBc7nnOd5zikRSalQy9lkSUoZpbRSjIwz6C8iqaoK2pA2/Ki1TimgccU4DdM0OOe6rsk5xRiz5JTS4eYQYvCLN9YcDodpnqdxCiGSSN20p9OnyzA0bXNdqA/HrmAU2BW8NU0zE8PtM0bnnJD2Pk3T4+Oj98kYZjKVQ8pQa4zRek1jL5GAcRydc+M4pURam5TkfD43TbsFJMGdyuVyVkraplWsYoyBtdapqoy1VilDRFqtRfjOOeI8DEPX9c7Ztt2JZK2Ns5JSvlwuy7K4qgohaLPmdBVaCWsnOacQiJlSTimprQ4mxojeHvgWegv6EGIWbYzWKqWc0D9Grw2XssiyIiGFErWVEjFGK81aw0lYyQalII28MbFxa1MGsFXYkpzRJ+lLf34ku+ecrF3bgJqtmHn11BUzUcwJbcuYOYYAI11VVQmXboVpzMzBJ8nsl5iSWLNGazYMNBtjEDx3zsaYgBJijF3XATQAzuotugbOcOU50MlWKVdVrqpCjCJ59p6Zq6bR1qacWamqrlPO/W7nrJ2I9NaTFbZQKWqamhWaWpmcktJaIW0nxvV1SiiWds4xqxgXWPSVAvFhHIf/+Z//+a//+q/7w6GqqpBTXdeSJef81VdfvX//MzQmSSlba+/u/pS49I/Ng7VqWcKnTx+W2b88v/h5+Z+/+vX9p4e8qF1zeH3z+pu375UoypyTGsfJxzBOU2F7MJtY97L80LwgQ3hllpHOpbQ2mAIREqEYv5RMEzGnnJlCWvPJMomQJN7CnkzGqrarlBJjTBZZfHx+OVpbGZv6/a33UalKKaVXVkohVBiiV5qZDWSCrzqBr/9nRvM8772rKiFiTUqRUjRN42bSskgyWiviLNlok3KOKVVa+xisdZpYkarr1ocUYmz2LfIfRARqBRAAWEZpLSJt18WUwJ4rrauqCn5mlYSS9zgdwRq9tnUsgsJfWok57wMzgt7eOZezzPPil6lt63me/TL2fa+UHF8+d211Pp/hqOLEjpubG/g9UN926xOAjY28l1KZhXdjjEozK5Gck0SnXRbp+t3hcLi/f6gq1M0ppaC1vFJfOrBCj4PWUIqt1UqTkCglbVcTUZYY0xIlGmdqrneH3TiP2uq6rhXzeBlTSJp10zSklE/Jp9RsrWoRPoFDo7VGuLLrurqunavuPz0SUdPU2lCMASaKmZdlEkkxxq49oLkkbMwwDIfDgYiNsS8vRxFxrhrHSYSWxY/jIlmYUvDRudpa9OdxOYWUPDO/PD/VlTufTitblRSRJlJEqmkc3JqmrZx7gy3jTey7fUrpF3/5Nx8+fFiWpdu1PszPx1NVuZiydSqlRPLFHhc3CC0YjNbCSm3ZkEp00/bEuqpbay0cjnEcs3BaJkoxJdJaC4kPXpEVIgmRldZaG2LDrLK0TR1C4CzMZLZME8mkhCrn4N/D3PoVIa0NP7CnkIfit2aRWmuRbK2xFm3tMyw3UmDgu8iWf2zJ5Jy1KK2UFhV88IsvLiz84xIVB7hOaY2Dam3D1s3BOZNSEsnMEoI3RtFVqujhcCicEm3BHsxhyZ0j5kxiKxdjjDmFFEnx5TyeTmdTVeM8V03DxtRaE1HVtl3bQr9jchDhTyKkVBLJaC8oGY3/hISVJiWKaM1jVEorZavaKHW5XJbFn8/nEOLd3V3O8vxy/N0fPjw+v8whWmvP5/Pf/u3faqXP58vd3d1+vyci5yoYPGPsf4F5iDE+Pn5+fnoBMP/8+elyuQzDuO8Ot7d3f/GLv3RVFWOc5vkyXE7DgDR2s7UYpNJl749VVyKnDZe+6rB4bVfkuh/9lrzEpdUwsyJSzNZWNzc379696/sd+IRlXubZXy6Xv/zFL46ni1baWnbW+hAScvtihLMJaqjve4hv3Bqjr1GHLcpf4E/XdXXVEiXnqstl2JoFSUrZaEo5iUiSHMOKleq2SYunlHPOmlltpa0559quWdWIkSK+ApcLxCj0cikK4e1alkVrE0Pouk4rhTQ4xMTAXIkI+ijAWsDvyVuPz3lerHXYt4BFy+LTlj4UQri7u+u6Dkc2AWDCOQP/gwub/Ee0LDxx5yqAlGXxbdullNVV8PA6mgJcWWi0tPZQIa2NkOSUc8rIuldKbwEm3bbWuarvd+fz2Vqnzeo8ISyJggYk59JWAA+0mNZyivX8lrZtY0xoSdJ2zflyjjEyg37ZhxCned7tb8w4oV/pNM+KWSl1PB73h8M8TWpr/QaPdlmWcRjqurlcLtO8Z1aFMwl+zS88n8+vXr0+n8/zvLRtm5IUYpC3jiwxRsVaKaW1Ohz2VV2nmKZpRNJHCL617efPj99++9+GYTRGxxCcNXlrilVy2wrWwRazVw0f4ZK6rS8kXWUWYfeFrcUvom4lLbIkR9FGA/KWWQtOtWkauK3Fp+Stxh7Qu2y9eNVpZkt7TeUcG4wfmVq0Icu8ZRnB0pS+GmWD4DVoPSLqug4Vu3FLLISPBRmIW4dNvKW1nqZF//CYFkAWZkaxfeG+iKiqKxVXxQWmDh8+ns5aKWuMVmqe567t5nlyripx3ILkCodBxCgzlCx566+OZDxmDldnWI3jaLX+/vvvcbSa9/50Ohlj7h8fn56e6rq+u3s1jsPt7d3t7V1K8et3X/+3b7+VnKuq6vo+hIgg5Z9kHX58HJAssyei0+mcUn5+fvn8+Lnrupvb22/ef3O4vclZpnl+eHh4evr88PQUwtrf7YtOpyI/RF96cq8Hh9Af+9C1p1k85fJpEcmoOUC3L8VE3HXdN19/c3d3p/VKHVwul3n+cgyDta6qjauq0kSPiArbzlvuKb5rtx4SOWdtTJFgpKag8ytJmqY5BIRPMWkp56xIhRDQPrAyxlg7TmOYl9pYSM80z/stIgcRgUc8z/OrV6+GYcD5LWBIh2F4/fo1gEbe2r4aY1LOKabKmZyz26LNdV0jBxSfge98Pb3r/skqxhhCPF8GV9XB+67bDcNIG3uOhBZeq5bWUyWKd5+3Pk45Z9iMYmbM1r5QRFLKXdfHeKzrpuQsFmIHsl4Ku6BhS0SqaZplmawxJMysjeJlWaZxnqeFSDGrtu3neWnbLsbUdT2oDKgtrN1utzNbDVG86r9CRHZrol7gCP5IQl+9+zqleDweYb+J6PPzS9XUKQuSO+dlcc5V1i7LQswI78PgIU96WZaUvwCjqlq5rKenJ8l5mubj8QhcM02L99G6KqVcJlBtrSoRiVXM8zIZYzrTEuW6qYhRPtkufvr662+++uqraRrbtrbmS1msyJfy6ZKlRj88BgqTXF1VKW0BM1YKnZXFGAvvUymllEF9olLqOp5cyNgyw2HraGCuzvUqeLG8KNnMBfHEre89HgQCfx08o80G6K3XRdoqh+AuuK3tRKEcivYod0wpbaHvKsbkfWiaBtjFba189dbcLG8567wlROGhgCdCCFDiGBU+BmEDTa21NkprYs2sWbV1ra5a3Oet6oiZg4+KNRHltFKseNh5ni6XC5LrMG9oX6SY/+Vf/kUpdT6fX79+vVZRpPzmzdtvv/0LxBHbtkUu4rff/sXbt29hQtqmSymhDc+fdv3wOKDZK6X6fnd//yDC9/cP1tbM+s3bt2/evs0i1tnlxSfJIaVlWUC+Y6XLIhUvYbOTP2jsdW0MfvSCr1jU67eyrO3DCOZXUIK4ApkQwvl8fnp6Apvx8ePH4L0Q93WNQhhUAoNAVNtxjDirz2wVksVXUFvJW9rq4PC6rlxak00bpXRKkVmllKd58t5b52prldYiGWQKb1m5sFjn8zlf5XenlG5vb//xH//x17/+9X/8x3+gWXqpM8CQZOvvtiwhe59T0tvBnHHrCIvggVxVlgDTYcyQyyUsWcRaVzetde48DFVVL+FLT0coC3NVHA5fx28Nxsu6FAiJvQrNW5Ip67q+XAasdUnCg5ahq0qosB31gendSrSQSY9SjBxjco6IWGslQnVd55SdcymmZV7r4Wlzj+J2nCQTQ2XTVkvf933btr/5za+RamWMOR6f4Eb0/a7ve6G8LL7t2uB91/dKaWOtDyciUlrFFI0YYnZV1TSNahpE+5kZrX9Pp5OzrizB5XLJOb16dffy8nIZx7vb28/Pz7vd/vHpyTlXVfU8LykiguqgtUGC5YyzC9HAgkHE7/e729u7vou7Xf/7739nrfHeXy6XqnKVs1LXcByhdzDJ4NaICAlOafO/oewKCECUPsbYNF0IgSgjN2aeF2MsERdJQCTAbCXE2AIQUdy6KHrIDA4Qg9FSW6IaEYHDxJCQkMbFEBkDWwv0llJCip3eMjtLJKDAFLee27N2/IS3jb1QNqxs1ayQumVRzMHaUFUtMztXa61ylhLGK3gIo0LJKhLqIOoppSV4VIE8Pz8jiQOuyWF/QIQMyWw5ZWuM2iBUGU9J3I/BA1jHGLbUc0Z4XK4OUMLsPdzfvzw/Pz4+gjJaT4c1hrWxtkKmSdO0Nzc3KaW27dwr03Wd1ibG2HU9ZAP+5Z9rHuZ5eXk55iwvL6fT6Uyi+r59//7927dvgXyHcRyG4XQ+x5Tapk1XjS6KA6HUF5Oet9NyUkqo25KrI1nMlvNePklbJiszG6JEwswZRoKEmUVxa+t3795t2WYSYzidTt57re00TeM811VDtLYCNNsBli8vLyCRgZFxnnbRXCDWi1uHv18ul/1+j4hQ5RySKo2xyCASoRjDMAxd1xlrEXswzvHWig5maU3CSQkqGBl+2LG4Cn8KDz1sp4rKFhppmoaV0lmctaR48Z5Tttai6QUcWBRMpJQAPYqnvCxLyjJOi3F1THkJydrq9tXrmHLwq56yW9kt8qCQtZW3rFMoVkBI2FHk5mIXNU3jfdRak4AF6pi11taYhCTd6uochZIOhDsys9965IEoExm1NiLJGIRtVdf2ivXnx+fb21tmbYxLSZCbqrYCwBBC3/e73e72cBjHoXTMZeb379/3fX8+H+d5jjEDCWIkrNXdq1fjNLF6CiGlTMy6btqqqog/G2cRhFRaz8vctq2rXA5rJoVz7ttvv53n+ePHj2mzo3nbCy8vz58+fbpMc1W3l3HeHW6O56HraFo8KaVyijGCxyu2wRjTNDUz4/8imVkrxXVtK1ejaj0EPwwDs5qmabicv3rzBm3G83bUYGFagDNow22YYb2d35C3rsDqqnUEqsOsddY6ZjTBW890Ctt5f7KdY1pavOmt9X2pp7lcLqX5yrXHgDZl2OOluhjjQWlI3JpS4OQrLB+MgdoqCVDaVhxchBIBmeF8w/7hK9fVLSBLl0XhGCWol2XxgC9ha9aZcwbrCyhTHCO1teHJaRX7nDOS0WF0FavL+aJZMZHTRlIionmcXO0A19J2ogxWYVnCsoTNBCbvAbAIjjW2M0yFtXaa58vl8ubNm6ZpYCHwI8SG1epdEVHbdpg0TWyM3e8PGzIA8vuBnv8TzcPDw+PLyzGlfHw5DcO42x1ub+/evPnKWCtE4zj+/ve/RyZfjFFptZ1M/gXsC6LKV7aBNz40bGcelGDDtam45jFW9zNnUZxzJoBNJmZ2VX17e3s4HKr10LE4juPxeFzmYIyLMVq0McHvbTQooPRxK06WrawXsVYkyOer5EvAK2wzvR5Qrq11fbcTkpSyc9VhfxO8Px9Pu90uxKiMXrxXioXWpLR5nkOKW2ZOapoGqetogrvf79u2PR6Pa8ek7dBguC+FWkWuXhahlI0xPoZlXoDMkQ6L1E9mLkUM03bEBRJDjXM4UapuWmaum3a/3z8+PpZsTtlYL3gDpewAeyxuVSAYM+o8rwCBYlI5CREHH61xkomEnav6vgdBd03Wwerc3Nx0XXd/f5+2qkNAh2XxXdejxwZoB+QveO/3+wMza21SekHyCcx5ycMxxrx69cr75e7urvhqiDciyXU9RJPIWjvPs6ucsAoxxZRTDsbopm2JeZxmbfQ8z9ZZV7mvv/n6/v4+xLA/HF4eP8MRRB9WcI83NzcoT4kxxhjquj6dXkIIwcfj8fThw8ef//wvXl5emqZNKY7jlJa5nGiNWQVhopQiEjT6zzkRGWtN3+9ikKpySHX99a//83A4jOP4/PTUbCeQq61DAW2O3caqT3TVOldfnU3PX/rBWMlETEbb4CMJx7hGpPLWA46+ZJ2tRHzB9dd4DnYaTgnULjwGRIAg28BG2AVYCFgLrAhcfHw3b80wYJniVkILWSpKX0SQtRm32nJEtpG8rrbD2FNKOROTYlbTOBOjWS03bQNbWWi07TyGLwk1OPcN5cfY1/Ag0eAde8FpO5VDp0sF1VZCe+2FM/M0Tt77GHFTnbMgZweeHxizEhBKKR2Px5TS+/fvi68PgGtdK8JCwsQpJaMtNlGYl5xFa1NwtlLqTz1L9Ifm4Xy+KKUf7j/HmJqmvbu7+/bbn+/3ByHCeevTNKF+lXCehv7SP1lv7ZWY0chF/m9sYyoGo7yrt9YlxcB8GU2MpFUIwaLJF5NSqlYKLCq47NNpOZ/Py7LElPu+f3l5SVvlBIQJ03w6ndAVAAuA8A5Eh4gglFBVeQsUI0D3JbIirFgjwyrFLIaY15NA1Nbeq21bY2xpMZZSijjXaOPNwHsgand7e4sIM+oSLpfLdVlcUcFy1fgvrbAvohECrFdpgvb09FSCwCiG0luH7aqusuT3X79HkWpV123XjZvjjAGUWS2irLVGTRa0DzhfNB4vimaeJmsrqA8w8hgwkvZk48eLDcYeLoJe2Cec5QfMm7dE277vWVvsOoSa9dbXbBjOshWFoTFZ0zR1XcHZgg3D4aDgHLz34B7rus4JR79VVeW00VVdQRkJ0eK99z6m9Pzycnt3t9N6t99/+PiRiLTSZflgdYDvnLU4pQ4t3Ites9YSU9M2775+93J8sc7qrGNKqCJEuid4QqwzbEOMIQTUJHoR6bo2J7XlU1BVVW3b1HV1Ph0fHx+hW1+/fj2OIxpmFGyRUiqNVYr5L9uzOK+SFc76gRmuqlorkyVeqZX1MF2sC0xCIWGKx89b4WG5F1A2gAV+Z9kOSS5MptmO1QIagN+DkLjaDu5F43oQAMvWbB8NFrE95ergW5QyzNv59sXr1Vr7JYmI0TaqlHOOIRljnLVary5XYbHwrUILF6NYVFnaSsSR7WqtRRmI2g6Jgswzc8wpXvWKd841TTMvy7J49D8W8YBxhb0oWBkKNoRAIjhTuiSzoN6F2Mb4BdTqtXKWqgrVFaqc+Zqz4IjvP9c8LJfZsrv/cJ987G9279+/b9tGJGVh5AMopZumQZuElBKyPqC8imaR9RSpEopQak1Cp0J6FHXMVwHqInO8pslmyWKtM9bGGFxVaaVv97c3N4eqqpx1KeaXl9Pz03Eal7ruAlLGrFHGKq2ttZfLYLTOIryxDSHGqq6T9wn2mSjEqLSWnInZbxEwPA5KE6Af61aR4cySU1z8XNXOWD1OyTa1a+rMNIegFIcxLcvSuspYy1qP04itn1KKeW3XAykHeHx8fKzrmoSXaTHoLs+amJytpnGY5rHazvQwWqeYQgjeB0oZXZKKcyYiWulXd3eK1xbE1tiUkrN28X7f9ZfLpaubZZys1jlGJWtnG9raCKKKuLDJmIeiELGqwzB2XXd7e4dWdzFGEdLGxBS9X7yfd7s+5eRDbupGa5OzhOCttcAyOWdSylh7uLlVSpFSPkZSyhIv3k/TVI4LSOhlr9R+YlvsKwAAIABJREFUvzNGE3H6X6y9R5NkR3YsHDquSF2yG40eDAFy+B6X/P9LmnHxjEbjCA4xwAANoLuqMqtSXhnyW/iN6Bouv0GvIKqzMm9GHOHHj7t3jFFrxxgD59w6g3haVSVWx7UuuJSEMUYJZUwVerFaUs4iIcYFyqQPoe16XRTGWKWUkPx4OB5f9nY0SqnR+aEfuqallNZ1RQKpq1lZlkM/kEjrumaMny8tF5ILuVxtjPP9MBLKIqVSK2FNP4zD0AkpRmsGY7SuovNvbu9oiHVRRuedtUopQmnX9sktA2PJaMwwjtMMFk97HMf7+3vOpXPxdDqhn1suV1qX4zgqXRob2m5smwtjfBx69c0/Bu85ZwhOxhgEUDYtLRBIU9DJvCRi1yFExwXljFEW8c8+WGstFhoy+Ri3dRzHvus4FyhXyWShSkKI8bPNGc+T5KwPgWI/plE2bgF6GnRgaKDhPRwjOZ/PkDJkjGPZEKIpoGJD+8h7B0wBgYIxorWSkjtnnbOcUym1UopSEoIjRDFOnZ20TvMAv237yAIlVOnJSoRQQjl3ITjvrXc+hno+8z4EEiMlQkrGuOBCST1yw7kIPnoWjLWUMZH2eSMhkHAijHLG+25wzgF1IISeTidrJtkPdEVqEu3gjBIaJ0jNGUcpJYEoqc7n89PT03q9hgv6JLEsRYiWJldR7y1ljFESSXTeMcI4Y5SyQD5PMv7e9PCyfTHGDP1QFuWb+3ssJDvnun68XC4PDw9N0ypVcC4pCLoksL/hLBEyVcpTbpBJSTj/X5bW1nPeyzmTJDM8NLZTVT4JfYflYqmUevPmDWTLCCF9P5xP56bpnAtSa+sjk5pz7nzgghBCnXNCyhDCbDZr21YqhdVloFXOe8U59h5pcvvLhbBSCp5C3vtqVgcabXCBhNEZuETVdbV73hV1NV+vht3Oj4N10/AZwrGCkP1hHxiVUkZKyCs7VeccvB4vl8tsNvPGkhBLXUQXtFSBB+KDMyaDYEqpwTpjTd/3ggt8ChwUmXakQ/BKKu89JdRZN7EwufDOkRgZKLacSy7GfjCplCNpAgkm1fF4TCH6s6hiJuqg6AbsBnrSbKY5Z8NgIwlCivliNgxD33dAllAQASsoy6ptW60LXZRiWnbldT1Dgc84h2/otPrEWFXXhJDNZk0pRRA5HPaoSJy31o4hhBArIWTbtsMw3t2V0FeOMZRVFSMpyopSrotyGEaeFqfni4Xz5xADocQZOw5j9IFGIigL1vV9P5vNxn7klFe65JS/HI/RBTfax4dHYy2lVEhVlKUxtusHxkWM0YfgQyAkGmcpZ/0whBhWi4Ux5p9/97vg/fXVlZTSWUtiNNZ672fzmnOOFoExEkIEh4IlJZt0ETzuE/qM+XwBCG653Pz0008+RCXEhx8/XF1tGKVcStCogCz5RC1VSjEmYwzGWFzJxPjKyB4PwTNGQfcIYdJLfs26JoQIzqWQRVFg7hrSphuh1Lop6KtkF8gYG0fDGAO5QAgxjiYED/ZBTEsbub8pyxIPgRASAgDqyYYPOQBvgzHWdd3xeIwxaD1d58SDmHCzotBC1GQSCgve+3EcrI1mtOkMTz7hWmsbbIyRUMqFSCrONMbY9V1RFFhVc971w0SYjgkLElwoqTjn3nkTCVJkSDaFwLhw2W3ygxn6oW1aa20IE6MHACAMgiilLAbHJiEDn5hL59O56zos5759+1YkYRjnIyQKYhriRj8JAmmtrbMmZYVUwb/7e9PD8/MzmAPYKshd1el0ulwux+PxcmkAJlDCjR0mX8JX8kr5+35NAyCJbEdeO1KlsX7KKIQkjtfrF8Rn44lPDTIyskvTNC8vL0AtfXLxJIlml/tBPPo8FwV3GIADDgqAAuCVuSEFVxLJIxLirO373lrrvF8vlujvvPeFKmkkwbnog7OWSkJCAJmKMdb1fVlXwzAIJTnnjEyNVNM0//7v/z6fz4dhwDorwBPUXMBbvQ9CSMZo2ubzQz9glR/BtKoqXRQ41u6V6SkyR0xq0nggWPT13oM9nXs4mqRk8dCAd6EIyjzOEAJeEMUXoEWfpBIBNtLECKSU1nVd13Um3bLEVwFSB14v/i/m4agogRrFGJFagI9dXV2dTidotQLzVUo5b0MIRVEKIYQUXdcPw7DdbsdxSPvYlJD48vyyf9nf398D3MeTwVwUUDK8KjOxDc8KPwkEzFqLzFQUBSSnABoA7vDe4wyEiTlaveyfwZsUQgDq/Oqrrz59+gRRFuzZvJxO+Fvkb8ee4Gv5zwJfxDnPuRjHISYfApxqa62Qsn55EYLHEJq2/c37L4EX9UNP0wYPS4wPnsiv+HZ4UtJ2znEu89XLQ+z8zebbNMEmYRL8YRNfeer1Q/xc/PHEFJ/P59YeOBMk0hhj8IQSxjn3zgE1Qv/EGAMrAUszIDJgjJE7Y+8dHjgCAqgiABJFIs6yJJNnrV0ulzASZmlfyjnHmGCMME4jCcvFgk5FKplVs2wEkhlZMUbv/Pl0wXCOUU4i8T6I9LgyOJbH5v7VFlEOyiSxN1Fl4uwBKc0/gwzXtu0wDNGNXAgcPNTH1phhHPp+PJ+b+WzhfXTWM8qNt8Z1XIi+h8iCQNKy1kopGOOQeuMcVo/s1+ke2rZdr9eV4G/evCmKwgZPCIGQAE5JURRCKO88IYzQwNjnZWOXLFJ9Uijjid/GElv0dTIIiXySQ1VODyx5PGTg3icd5sx76bru5eXldDpNDQEhQgjQvHDaEO+QrhC/UM/m34KxAfItQj8aT8xssWd7PB4ppafTiXLWtm1wPobAGEP4kELEENqmIZFUZcnqehzHcRicdJjC4bp770tZMca89Rmf5ZxDwi/G2HadD8GFYL0fjbHeU0IiJYxO+4NKqVNzwOBEKaWLQipVlKXznsIxGCRQzjgVGNIwxu7e3L+8vFzahnNelqUxJomMUq21tSY/cwRfgAk5Z9BEP0U0yZfBJKFZnlQH8ImgSDqbzZRSEFulycYDL0jSsuF+v8f3ixKSJiNltGshqVduNhtM7yGhA2oKWiWUfhLCc9YOwwBNMEKitY4x4b3vumG7fdK6lFIul0u4JmBTHZkYokzIfFCnQRVGCLm5ucHoBeQF+cqFpm3bn376CeNu5B4w6723yF6LxaIsC5QXMUbM4UMIKHtRXMMFVkqeK5gYiTGTiWyMUQgZwlQVUkpz2QtMz4egi4JR0reNc261WoGxJpNmGobeJDF8OC+U0tY6xih4d+NonfNKFSItr7C0TRL/dtE1JJqydRY/QQmVSoYQSIzDOBprXAxcClVoF7wLnjAaCBmNV0oIyYMPXdeHEJRU1jrWtlAEQtMgPjOzuU6LSngzzjnOWV0vwNNDFreTIDHGVCRGKgT2JGwIlnNZ1zMhZNN0aE4YE4TEEIJUMsbAOSOUpOGXiKkMxUXDf+y6TgjZNMdMNA8heOcdmzgj/hVjFcEnPyuePOdJ2tuIyecGPHvvvTEhxx9CSNM02+32fD4GN21i4ZPCEWu73bZdV1YVFWIYzfF8UUp67yNhhFIs0k5CW5P7ZNW2HdR6UOzWdWF/FUHvOf6slsvlchxHJgXqskvToeJmjIdAGOMkUql4jJ+XA15TI3IlknO+TJJwMVmf+8R8yK/AEkkL8QIwPSLCfD6/vb1FAEKUOZ1Oz8/PJFU31lqdSBooRVGbkGTe5F9RoUViTMdEus1VAE0L1dfX18fjMasCYANWMC6lAu3s97//PUSbG0qxrEtzC0xpFhnONCTnnGLTgn6OwsgTQzTlrA7ej9ZwNdncBxIVF8CRoUEbBWGCa62lVpESwmjbdy5MaviaFROZR2tCCGesKEsuJbwHCCEAHDI5FSEYlxDzQMidIr7k3iukpTacMJha0+SeZJP6DaUUiBMytHMOm6vL5fJyueB35TnnDz/88ObNGxxfJC3UU7kRxJ0B0IxfgekcS8q74zhtOKIiQ8sFzcQ8m12v13U9894vl0vkA9QWAOVijNZOFQ9OLHRbUdgul0tjDOinIinabjabm5ubDx8+YGYYk3C0lHKxWOx2W9wardXNzbU3/u3bt+j2MjGUEIKBKig96UUw2u0YmwYDSimtS2s9LrlJBss+ac6bcRz6frNZf/jhB0DzU5gI07lKt5Wl9M8YEyHYGEF24MZYiHcBnMQ7gWwRTbwm+koYDkfC2cAoBNCEc34cR+9dN/b9OHRd50JAhyeF6Idhu30pq1Ir7YP3zg/D4LyXQtxerbHb5V/tQGgN09ZpHQxnsus6DFohDckmowuwjKZEGGMkhIVApFSEUM4FISyEaIwdRwNnBWthfqCd80WhhRCMISUHYz1OCEIHaoWmaWKkWpecyxgJY8LagRCGd4hQObVrfY+CErVLjnK5P8DLojgDN0lrbe0AnABLDLvd7ttvv23bi7fDOBqt1Ww2G4bx+vrKOffjjz9GJq0LlIndy4tSWinVXC5lNRutiTEqrcqiRDHa9z0JBLjc4XgSnN/c3lrjL037K6SHxWKxXC5Xmw0S4PFyPhwOLy8vTdsn4SovhPLeexcj8YyR3CW4JKce0mw5TxdiIrnKtIGZ+77/9c+vOkqPXglKkyDG4RFordq2fX5+Rss5pRY62Q0RQlBJIWT4V2Y+7JU5VG76QlJSBLMC0AeiJ/KQUupyuVzaxhijqrosCq112zR933PGGKHEB0qIn6ygCPalgT/GGBV08Blzzgk+dcGAKXjS4COEgLNEE6TrnCNkAv3btucABwgLJKKzIYTAujJrAEg1uWJgU7aYzYSU8/nc26mqRbxA4FBKOTfVqnnWwhLhGG8AjY5NSvpK6dwFoolBKI9JuiDXWXikuZLCJ82QBZCitm3HcayqCr0pfhIfB+gfTyrNp9Mpf4OIUyF6IEKEkOPxeD6fcTFCAGEjhuCwa6FUMQwGdHj2yjJeTlrxISRaFMoIxhh+oCzLw+GAu313dwe3ouziIpNGRS5xCCEheKXU+XyuqkJrfekbZGW8MrLgMAyCMe+9SSLEr8qUif8W06oz8EkoiQIYxKEax1FICXAGSR3pE550+UZAT9SnXa38yjGRi3IjnmvbfPzAi8sJaToMhAZCt8+71XotlPTBexKt98770+mEminfqe12SwgJ0Z/PR52cDGKM3vPzmRHyBXpEn1zcMXJC4aKSd2ycyO6TMzGf7EOc1roodNt23k9NLdo774NSXAgBMQJUk4yxup5JKZwzhMS6rtFC4U9ezsj3CN9s204C9YypGOMwjMYYITilk/MxTxY91lqUEeiA/cTj+OwPgQoPVQtiC+c61zFd1/3lL3/57rvvCAmCIbvP2hb0eiuEmC8XZbX48v1759x+f8BbfX5+Xl9dN03nnGOcbTabWQ2HK66YAE/pfLpore4ot9Z17a+h2MoYQ2FFXjkws7SOy6AvGIjggjNKKLwD6eszF5NoM38lioLjTpIPRIYLSULoXv9F8mqAgbIUvGOcZlyV8/mM26uShgygvcy5xDiIEGKTRDZJHQxuGksrMy55DpMkwIBTCLUc7/18Pj+3jfdeQiGyqlDQTYBYjNiR5owTQowxXAiSlicYY1opYNnW2u50ycO0PMLSWm+ub6WSo7OUMU54iZvGqWDTZ5FKcS60EsZZpVQ1q1E4nZsLYXTaEHQeLQXnXFJW1RUUxGATjVBVVdXhcEBAATyC2icHdKi0ZpwQ3E1w/KWUIKdigspfibzi8IBli/SDUWp8RXrG08ACPJ4tJDogRJrFlvFSsIo7HA77/f7h4QF/BU/MGDOaQUoukIa9AS0yhKAU5uSDc+76uuacA43UumjbC3AJhKrcvblXks4yGSjhqrdtC0YDGoj5fL5arVCv4IsGgYSmBXUpZSShqqrr62utdZxNfgAIBLgCSilvrTHG+UkUWQhEME4IG/oeGrfej5wPqGoxpZTyc81xOp3g5oZPNJ/Xh+MRp5ckq06kTxQoIQRKqLMhBlxV4l3kXCqp8x13SYkkU/WRobPaxPF47IdRqvLh4WE0tpsvsOLQti0X5MPPP63XayFlXVWEUsnkw/aJ8WnoGpkfjTkc9sMwmMH842+/wZwPVU5MMy1KGUxk83/EXV6tVkDVtNYYVODiD7313sVAYiCCq6bpunaoytk42MeH7el4udpcL+YrQoiUgrIYguOcAUegaSp+PJ7wQdAE4yForauqNsZ5j4IjMCbKciKA8bTx91oSP2MneGg5oqI2cklICsvYhEQpJVTOdrvdzz//PAxDXZfB26IotFYIlavVknO+XK/evP2qqmvG2Pbl0+3trdSF8aHrBnjnVaqijBlrh3GYq7n13nYThlwUZVEUh8MhX9K/Kz2APY0PPI7j+XwGlTg7yYQQOZe5e+B8Gkji89Nk24BcQl/ZkZIEH+WWQiQj1pxaELhJsuPAoSmK4u3bt68lAYahf35+zgrsU1kUQ4YOM11sGIZhhAr3NMTPOJJIcv/oylFH500xSmnf93CA8t7j6xecK6XKojDJjD4DU7kgnVgKaUeJC6GUJpNu6ORSi/OHMSwCRFVXCNCYhCNuns9HYKb4gFKK1WoTScQObdM0/dBjwgEh4ubSiOSgyxgjlCqt6rpuT2dEIEwXEOAOhwO6NZJqW558eADmAMbBIcPLAugA8ILpYkwEhIy34nUQ0YqiyL4FyBbe+0vbAjPE8Tifz0gJuYmEHVsIwVr76dMn5Hh8KWg+Yox1XReFgjatsTakZZpcGbBkwuEn0cNJAA6PHWfJJJV4CHNiIwQoB+AFXAS0/zHG2Ww2m80eHh7wY/g2oYaCPhVElHwdFotF0zRIqPhdGM+QEM7ns9KCUtq2XUJ3xDD4vuvRvKImtdaGEJFIchgyySq1KIrDYc8YWy6XTdNopcqyXKyWuXcHAMsmfw6BesW98pmx3Co9QXkgwgJEDSHAj5oQAlGA0+n08vLStF0/+pfn5/OlWSwWDw8PqLh70356+Hh1Pgshrq+vvffDMHz33Xdv3t5Y2yulhqHZ7/cvLy9SSsnl5XwGbQ+ECER/xJbc4NJkHI2KGx8cqq5lWYYQrEGCZ5njRCIJITw8PFBKd7td0zSMfYPzU1WVsT1LFooZosj1AbICShA3uQnIBJN6fHHGmBgh+JqWfobBJZN5lEf/q+TFXSbJso0QgqUcOHtxzp+env7rv/7rcrkkgb+apC725uZmtVoZY642V/qVnNr9/X3f91dXVyGS91/9Bl8cWA8Z7wpu2nrJY9pfJz0IVVgfJWOH/el8vuyPh9PxQmL0LCSCqX/dEec7GV6xVGma2CAm5oDFmKCEE0pijD6t++PnGQ9DPwgprEFN4ShjwXVlVUrBlRTr1VIK7Gq6c9OcLpfRutV8SQgRUlnrqvmcUqpBUKHUh4CFBpL6x4z4p8zEs9gvbAuFpJF475x3zlE69D1nTHDedx0NUUtFQ9wsV5TSffPSD4MDZ78AvZpOUc8KKrg3puA8hKCFYJzN9CwDawBMfHLNQyCTjBDOomfeEqWkEmRWzS7HA+GSMlHMBIl0uVwIIYyznHE3mMvh5GOgkWip+q774t270/Gopeq6ThWSUjp0vR0Np8wnwWfcW1T3EOhOUED0PngfLPiJ3uN7YUIqxo/nC+VCKB1ioJwzISIhMcToQ9/3wQfGGZ5AxuiQElar1XK5/OWXX2ADMBWz3gvGPGN2HLXWzloHskDa78vXA0Ov9+/f48ZihjnJV/jAmRRCcia882VZDENf16UQjFIiJVdK1XVpjOGcem8vl3EcbYxUSj0MhnNJCKOUY4ubEHK5XAAZOeeQsHOE1VrDzhNiBphmM8ZQciL3DMNYlto5x5lglDsbQ7B984y4drlQEBAIiVorO45lpYuiwONyzilV9n1v+q5v2+BMJH5/OBFChVJN1zFGi6LwwQ7DKKU01nOhuVDO2e3ueXBW6uJxt+uNYUp1w4A1haen7RdffGGMRbjf7h6c89vtNoTw/v2XnAhjTFkp98rHCYUg1pu22+3bt2+fn5+llJ8+ffr06VPbtrv9S9t2lNA//em/1qvV+XSasikLXd+eXrZSyuZ4M47j4XDgnD8/bwPxdV1vNptA/GiHotK6ULf311zxYRzOu+bh0+O//uu/qqKOMWr5WYAPs1ncWW/d7mlrhpEEwrmInkqhx9AEFiKJVNDIgiOOCko4Kcvy8fHx1Jy01vWi9sSHEIy3nhDjgxAiUEoZDTFaN4mGIXxl0gpyfwhBSuhKGe9jCLGqiq7rOVeEcOdijCwExrkehl5KZa3Xuk4wXSBEKCX6/gIg1BhDaJjNS+cc40RR0ffj0Jtv/+e7H3/4iXFyc7OmjLAYOOebzcZae3V1hRXu66sbKeejNYzx29u31zf3Ly8vTTsoqdfzq8eHB9t7STx1DLl5ubg6HA+X9myt7Udv/Wffrb83PRBGrXf9OPTjOBp3ObfGuBAcEdQmcQWSQEBCCFBjHDJCJlQut6ivk0eM0ZqBJzYY59w5DwFUltYvoY+PIEVI8H6MJGKuGBLNaRjH5+fn/cseIlaUCS4EV4WQirNJLQNfUlYMxj1nae6N9zmOhsBpR2vQHKUUjFEA+qjiSZqfU0LqosxT0JxmkJPzr8D4lHOeW+DgPTh2ozVA25Dt8x3AZQ7RCy6qqjRmnM9ndV1jQb0sOSHex8gYK4qiaduu7+a3teSchAC3CcYY8dRaW1d1cB7DYTQKoMfg8SIUAkxDK+2Syg1JyoZ938/nc/ZKMg+iY3hQTHBdTEQX5+w4TPU4ocQnV148Ooy48+gVWI0xJvt3Ajapqmq73U708/QzqOl8kvm01sLCFkLNoK9wzkOY8M+sApLxdFyqxWKx2+3wZcXkN6e1RkeYUdNc/SFnpwWlqQsJaQMfPd/j4yOCKWbgMca2bfu+a9sm4zA4P9ZayZm1U3SGjAqmBYJTXajlcqmSLBLnwnvsrEVK6fPz8zBa70OkdDabBeeEkM6NXde9efMmxrharQBY9/1ACTucTvv98/54oJxpLtbrNeccO9Xff/+9c+7m5gZo3m635Zzf39+N4+V8Ph8O+6xZDabZarUahuGXX355fn7GUY8x/vWvf4X8iYthv9+HEOxolOBlobXSUspLe5zVNVB1yK6UZblardRMvuz3Xdfd3t56729ubrTW2+32rz/8dRiGYTC//PLpdDz/8z//36oaV6uVt2PuR2US/Racb7fb//7v/57NZre3d5xLzgSjgpBYFNoYE0kw1vDgnbdcTCojbdu+f//eGINGnL7aK0ZbGRPhFX2tMQaPIkx2EeAsaEBJ3nspeQi+aRrGRFGUAEvQbzVNI6VEzwecE7+u69qu60IELE+MMXVdjuNISHDOXy6XDz/+/O23f+GcX10vb26uQ/Tdpbm9vb26ukKPq1ThXCgL7jzbvxz7cfjtP/wDo+Jy6U7n5v/+0/tCl4yJWT0/n8/z+Xy9vhrHsem6/eE8GkcIbbvh8Q9/evv2LSLS35secgrF4SjLsqpm3dgJ9XlilsFZdNxoq3E3yCs9pdxn5T4O+7k08bKVqtMMlmQsOI09NGNMcnr/5h68QBSPwzAcj8fn52fK2KKeCSmVKpz3PpIQAqMsI90hLRahdY+vRI/zOBGdFwiIJNl8OuegKDWbzYAtymSqDIuFjHIopWgS4kZIyngLDiigrWoxjzGyxNPliVqOrWkhxGq1Cp5Qwb33ZnRlUXsX26bXWnPOWCJfAccQQlR1LRkvioIw1nYdMgfnfD6bZ0MCBNbj8QjKRAa+UO5hUwHqhOgF8UmnAbUQgIPw6DabDUQYl9USsFgIYRhGzlhZllJJk+a6Mf0BHAd0JadPmvQz0PNuYFwMvIWQkIgJmRoAMkL+4jA2eE1tYIkyCKr7bDaLkaIFAellu92KV7x4gBLIB0iWQ7KhzmxmzjmoWciIyH9oJrD3cLlcMPLFG8DsFzkMRAls+VVVFYMvCkVpvV6v63rqn47HY1UuMJuBhxoeMk0WArPZ/Nw2jLGyrEZrtVKtMagkpJSYS+OdW2uFkH3fD30PWOZwOBAbC11RSp8ed8fDef9ytNaeTxdkqf1+r5Tqup5S+tNPP61WK6Wmg+Gc2+12iOO//PLL5dKcTufT6bTZbH766Sf8ukipKrSW6mD3MRKtC0qoEFKpQpUKzwrIp9Z6Np/LUha66vve29i3Q1EUjPD1cjMM4x//+N+n02m7fZ7PFsfjHsNzwQhqF5Y2NhhjSuu+78/n82Kx4FwURUEZxeCtLCrOJijPWQ/NKABcx+NxtVohZ6N4H8dxGEbvPSXMuzCVAoFkvBdL5jTND2KMp9MJxykk52NjDMwfHx4eMmYISHyz2aCHBjqqtT4eD33fHY7ee6+UmM1mWsvD4dB1fYy8uXRP2wdC4nw+//rrfxSCbLfb8/kyny+apl2tZNt2RVHGSHwI1toQA+dcK9U0DSLxfDEv63p9dXW5XOr53Fo7GFNWFRMC/se4Ahjc/jrgUjv0UEt+etwSgmjrKWWZlkfT/BAHGtUceWWa8XrkkOMFSbsLEyxOiHi1HhF9UEJaa7ngIYYYYwyR0iiEnNT3tMbvxfbg5XJhlEohE2JFvXVKqeCnLV//alMG3ysAZZpsTPB+pJTABFE5ildO1Aj0eKyY1pZVRQjRWqP2R0PTd51N+gEAIgkh2C1AmIZWRFEUAzZ9Qsh1K+bAiI8gfmC0rpKEZFEUMTKaDIXCZwNqAn83yvloxuVqGUkcx9FbF0KYz+dQ3wTd63w+08lSdACcjZU0BFaS0htLVDFjjExbCAiaIIzGNCK6ubl5enzEd4tYj0+ao0xM2gl1Xe/3+9xpITlhVlkUBeyOgbEEiG0IgfCaBzMhrZhkcgG+DpYYaHhQlFIMBvp+7PseJMVxHHFJDofD9fX13d0d2hQ4JuGtjuOIaTN+FygoKHQwikBBgE8HiV+X1lZCYm8PQw+Uf/qihwFlvmB0s1mhrFPJvaMiTT6aAAAgAElEQVSqqjLtXgAdznN+xCbI8oRu4JxXyVA211toROhEmHHeu/l80fdj3w3z+fzpacf8tAu52+0yFCaEfHx8tNbCPaZt277vP3z4oJRaLhfoaUIIu93ueDyCjkwIBdaENgJxvxuG+7u7qqy8c8vl8vbmptAFAM96USPrECK0rheLzXq9rmblw/ZRypIQUtcLiFErpamLu91z03RlWRo7vry8cM6b5qwT7p+HXmVZeu8+PT6UdaUKPVpTeG+sY4xTJrAkwRPLLoQAoH82m33xxReMsaZpzufzzc0NLj78YmNa3EHxgYoB2D0eMjTcsNWIb5wQsl6vGWMfP35smu50On333XdQzzTGYMYGJl5O4ff398ZY/NIY4/lsy7LEKzNGhZA//7Tv+36z2ZRltVwu27ZhTAzGbZ/3uqxLLEjtj+fzOUYegricz6vNJhJyOp2GYcQhR7V3PB5l8oCB2iNqDp5cp0jiBP296QE0Elwta+E4RpyzhAWaRse58c9xP88h4is6VwaXXieJPG8gr9Yd6KvExpKleFEUVT3pu4mk+/3w8OBDmNh4gsObqaxrSai1VsnJthrLCiyNmlEUo1bFN8o5p5SgrJ7P5z5YQghWbwY7gDYzn8+PxyNYrcBkEbhRJfGk3II44pLjBSFTBQTBNa11VU1ic5RSISf9yLqukd4ZY+fzuSgqCDP3/Xi5tMYYa33XDXU9I4S0XbdcLruh7/v+zRdvu67zoxFSOu+Xy6VzDhVWqQrjR0op+owQAlAm75wQ4nQ6IT0g2+UWCj3y/f09/rUoCp+sC1KTpzLNtO965LeiLIeuL4qi6zulNW4m4jt+O2o64Cc+WdfJ5I2MJIfxbw58TdNgGRNkITxJlOc8rRchepKEDuW1mMViURTF8XgyxiyXS+89iJXohimlWP9+fn7OZY2UElkZEAFjDINln5YE8bicc03T4Enm3jf/LQxyIL6LBIA+rK7rqpBc0LLSQjKp+Dg6LmhRlloWmVKBlRTso7iklK615kKFECnnWmu6WKCVx4HBisPhcMAFvL6+PhwO8zkVXG1Pu0oVf/rTn0+n03K5PB7P33zzzdPTUwikawdKqTV+MS+aS/fx48e+G8+nS9Ncbm9v8aHgJhtjnM/n1gQpdFSUc+7reHV1RQgRWv3DN1/TSDbrq6v1ZlbXU7XHokqjUYygJsTGGUHk+7d3Mcbr1TUysZSSEaZkUehyu3s6n88v++e7+1vnbXAOQM1mszFJjhdPmFLqfcBdns3m1jqttRAqRupcMMbhy1ou19ZO5nRa6w8fPhyPx6Zplstl03TPz8+LxWKxWOA8MybO5/NqtchqC4jgLy8vqLvBOIdNG5au/t//+w8pdVEUHz58AKB0Op1++9vffv3111rr//iP/8hzPvCtQ7CX5vTmzZv9/rksS+cMNPOXy/nxdJBSLBcbrQtnw273HCOJhFX13FhPKKeMtN1gXei6frO+ZYy/e/fOO/f09MQF50L85fvvvnj7Rdd1hNHletV0rSq0cTbRMeI4jtfX14g/Lu3A/13pQSoFp/B6PhsGwzlnVBAa+r7NMf1z1Z8G0Tk9AGjKlXtMgi38lTUm/nsuhwkhnHze+I4xkqJAobdcLnF/sMeAb9o6Z63VulBSGWMZFzFEQgjjjHOOEgCzBBR9QJkAi0FRp5g8RKdeWAgxtB2aDErpAPPnVBTjziulaCJ0A6aE2NmgVO6FSeLjovZEK6C15oxzzaG5JCIFlxTDImT7rutCmDhd8/kihNg07TAM1joEEdJ1Uspu/wIm0u5py8KUWfu+x9Y044xQYtNyu08a2tZaRkimWKBU11rf3d2RJAlZliU0MPCp0VSBOR6SqjMSNkYaMcZCF4ygGpiE9a21mJcsFgtUTA8PDySxXTHvzcBXbiVhqoOSHJuPwG1ocq0JSYslvlJcx/wmD0Xwf11Sa8jJOFXZ9vb2Vmvd930WlO66Dhjxfr+HUgD6hvl8jq4L7c44ji8vL2/fvkXIwCEBJk6T7qn3NkaWeykQw+bzuZZMSCaEQEGHx55bWzQZaf2CKaUEnew5lVIi0qIouZRSynPag8lp0jl3Op61Kr0LQsiiKBiryrL44u27q+XqL3/5i5RysVh0XffVV1/h5J9OB6XUbDZbrVb7/R5atoxTMOtZ2g2sqmq1WlVV3TY95xyXRQjx9u3bEEJk9Ob2JoZ4vblar1ZKqb7rUPdxJcuyBns7lYmcUlaoQktdVZWbuRgjjTSGSDmv6zlN8/Aff/zr+/fvXl6kG90wDO/fv8+3iTEmlarmMxt8UZdSaymVrso4jDEGTNfitMUW+SRDMFmttW37n//5n2AH7Pf777//4Xw+397eYpbpnLu9vd3v99fXGxwh8L+NMd9++y1jbLfbhRDatn15eTHGPD4+xhj//Odv1+srlBeoew6Hw1dffQXY8N/+7d9wjC+Xy5s3bwghzo1ScVwcay3MSObz+Xw+m8/roTfz+WyxWG23T1LocRzfvn23Xq9ns9l8vjyfz4yJ0+lSFGWM8f7+riiK4+kkhECss8bigiAZQGce9RbwA0wBcQB+nfRwPB5Ncp8HVhsDJZRBBick8cUcc3M9FV85uL6OlfmiUkqxFMpfLWTil/JX3QNCMwLWm7dvATUgkO33e4h6aq0JmbA0xphz1vrgjJHzSccbsSAkjR2alvUwXcCnkxK5geXaIQcmvAjwIlSFSqlICHBMFEfIcEVRDElVP0clkjzZy7JknCutKKWDGUMI1oUQgkg6xtltmKSZGIwEXl5ecnWfmZ0IXlJKmgLKaIxSSmoVYmCMhaSCkMm7OMQkdXIJsIq4M7Ab9MmnCPljGAaZugp0SABYcaOEEMBY2CsZFetsrhuAnxwOB2PMw8MDTBZBCsITzvLa8NtC/1TXNWEMClR5PJBTe95IQFaAfDd2qnPLjBSFDg8vi50VkpYAcBIyLhQS2R9fJb6C3W5HCIEWSFVVl8vldDrh9R8fH0HYzTMPrGUMwwA7W5qYTpfL5fb2dj6fkzhSGqXkhATnrDEDIcF7v5itEbwwssLfatuWk7hcLhnzkVGlyxhJIARTipB21oBihRBubm9iIMvl0phxuVwJIaqqFOJmXs222+flcn1zc8O5XC7Xq9XGOYtfhNgkhLq+vhVCaK0isShBXj/k9XrzvDtmKzSAJ4QQMkGJLIYguZBCOuEppVyKoq7SfsYkkyOEIiTECFKyhD7AfC6MMYQwISYlkv/59s/Pz9vNZvVP//TPl2NzPB7fvHmDs4FvRyo5mCE54cxjnOoGSmnTdNb6y+Wy3W5ns9nd3V3fj5yzrmvByP/zn/9MCKmq6t27d99++90wjLvdS2bB3dzcbLeP79+/g/8EY2yz2ZzP5x9++KGu6x9++AEwLCL7ZrMBr4FSDq09mXRW7u/vq6r64x//uNvtECexzyilYIzc3F4BDmnbdrlcIACG6MuyuLm5poxZaw+HI+Yly6urfjRFFS9tx6Wy1nKplqt11/RK66ZpmsulKMtqVjvnCGEu+HNzmc/nUE/wMWDxBe9NJIEJNDq/Qnp43u36vrfOEUI5l8654COhJEaXJ885f8RXxNZccYdXpqE0dQXpHz6vulBK4XPLGSfO5+4BZWOpi6qqshwpatjT6Rx84FIy5rUuGDAKSkkkRaGHJBmExp+kOQGCO94VyMuvWx/ceetGjBn7vo+RIIEBcmVJX8WHABAcwCVLZookydogjeeiVQhRlmVIaBvehgsj8itjbL1eo8wRQvT9KJPXG9pPfORxHGQSj7LWVnXV972xloUJvdFFgfdzaRtOWJU2aZEAfPIWXSwWJPHNhmQriMc1DAMUgWJivlrvh+RJieeGXYf98bBark7HI2csxOits86GELngIdn+YHoMIAtpANSRGCNCP04ONJkJIZDMwmkGdJB5I7iEORy/jvV5ey4DVtjac84Xhcwvglob3DDAZXktgyTIaxgGa816vd7tdpSS8/m0WCwJiWVRPD09kkn2uc3HICZlDiyO4ZuFgA9yNsYVcdKQ9xmecs456yglHWtDiMheeVNdSskiPONqKniIlDHuQsD7994joiUsjlZVRSLnnIbguOBVVUK6I1i3Wq3QBGBY+vXXXw/DYMzAkycV0CRCiBA8EhdjwMQrhGnOv1qt6mq5Xq992ulB+iScQRufEkIjYYw5a5331llmTK7Dcg/tvcfcJZcsYbIg5IwVmDegpPj+++/P56aQBWMcS/K5D44kGmuXy6VUKhAilQwxdn1nrX18ePz+++8IIZSyvLgzDN1+/4xhMlA4tBfb7T7GCLEvUCSen5/HsX9+3sH8FbY/yEw4UUDDgEOgMAKNAocQKo339/fffPPNbDZDvSKlPB6PGM4rpeq6uL66jjHCY5wxHkLQWpBIOefzeblaXn/46ecQadf3s8Vis7kG+HE6na+vr7fbZzpZYHmp5DAMXIiiKBhl1lrvwilJxp3P57IsrTGGC5jtoJbFUfTBoz77e9MDI4wGIpmIMdIYWAyUBEKoj8RbRykNxDNCGeMxRAEnH5TxEIiKhEbCKCWRkEhiCsSEEkppJDFg641yyphzPkbCeWAkMEbzzLZS0gVflqVknIZIQ6SEPe/2/WCELIqybIYXqYsQoxkHVLXD0EolCQ0hOh8oTir0693oIuHGjkWpKYs+WMaJc55RyRiDfCmhQUrZ95MoilLaGiN6jpXLerbgQnkz5HYHvUXf9yRGWJVlGBoZ1DqvdUGYKAo9jEYq5VwoynLoxhgpY6Is62EwIRBKuXNhMAbmB9b7puso523fS61HN9KBKs3b7jwObV3q4K0UIgagGYQz0ZyO8/nsEoKPAXMjkqbE6KCfn7bB+VIXkZBIKZfSeNcOPdosgNq5iQHbVQgB+g2amDdv3kgpt9vt+XQCdOa9986FZGpPCJkxxqWE8E6k9Nw0uiwvbcuSmYx1rigK61w/jsa5ABVS/AJKpZTr9RotPGAi2E/hUyC7gFuCLI5LCzYh0vzpdCIkeG8ZI217qaqC0giwaxg6pUSMXilhzMAYG4bOOefMEIKn0XMazNByGouiUIJaawvNteBm6BkjwzDMFvPT8RhjBH6COgPtVN/30IAjhEollJbGjvvDi9aSUmpH45zbPz8rJWmMj4+Pv/36d8MwIpqIJAlMKS0KTWjgnOpSOx9jJMFHxqhSwtrAWCwKBcUIa11ZWEYxe5w8X5WQVVUdDi93b+/A/qqqqqqq4/FY1IVILnIhhE0isXjvhfisGecnaWjOOb/hk3ALT1xwBGuSdA1CCE3fjd4CN8hVF+qJhD/TEFwIE84cJpTMh+iUlpwzygLnNMbY9+OHDz/RQDebzf/8z/+s1+sszlhV1elwMoOpisonpPRyPhpjjsfn43E/m81ubm7Gcfz06Wcp5U8/fPjw4cN6vb69vaWBBhtCCLvHnXeOc14X06OgJJ4P+xijMSIEAOakbfsYKRScshUYtvaQ7RbLpbGuZMX9/b21th+72/ub5XpR1/X19fX19TVa4bqutdar1UoI8e7db5DwvIuHfUdINCMdx/7Tx9P66oqK/tg01WKmq2qm1GKxAt7btv3NDedcKqUo44Szqq6RSmEXGH3glLFAVrNFXZQskDH00QXJuI+eMeZjkJwRTo/t+e7urm27XyE91LOZLorcB2CyJ6UEne71INElA6mYTLtYUrWlfzt5zv/gQxJtjRTCTVP7EW1OD4SQzWZzd3c3n89JmMR8jufL8XiKMRZF1XadS0IIuebF/lou20Py8kQEzOc1JjUhpVTwFKFHStl2F0xy8KEoJSgE0hWadobzPhRIroQQpXVMxrDoNFmiAFVVZYMfhiFS4rxv2harj+jl+75/fHzE0il+KeCdGGPXdcA3hBQ8mfYAvl8ul9vtdjarnaVKqdPpiAqdEMIpg88dABbQToCQ8GRzSFjKB8vF+XLBx4GGSp4rAgvCknz+OCDVUEpPpxPeLUurKqht85yDEAL0Yz6ff/nllz///HPuIFFfA0ux1oIS7dMKHhRw8Zr4yhCbXl5eMrqK44eMnnGkfErzPwM0y0Vfxu4y8kYSey0Ej4+w2Wz+8Ic/zOdzpMZCy9ub60+/fDJaYW3Fx6C0JnH6spCGEdyhxSSEkFIUhX5+Nt7bcey9t1JKZ+3EaeHs+7/+tW3bt+++ulwuhBDMvYB3F4UuCyXEdIbRuANYE0leFMwIHOBYksmTJ8k24AGWZUkZxcAApS6YTmCC8UT4zoAwor9Lxs75qmbmWP75PAdiyQPOex8Su4wk7mJM3HFCYgiEMeKSxzImiJfLRRdqHA3n/Hg8F0VVlrUQCnfqeDz+6U9/Wi6XX3zxBVzwjDFwwvn48ePhcNhsNi8vL7///e8hgQcwQgh+OLTjOPZ99/333xtjZrMZiMggPWuth2S1C1weF8cmYQ98ZMDLU/6gEWeDMfb+/Xs4Mt3dvzmeTt771Wp1Pp+vrq76vv/DH/6wWV1tt9ubmxt0sdi2A2bVNC28sIwx42j6vp/PV8HHYTBKF8YYpfXV1ZUuiwz2YoDvkzEcYgvyE4o/bBTBvJJSKoW0wgLI7boOARBzTVBQmBAAnf7e9IAIlZfaMw+MJVfbnA9YWsOjaRAdwmd5rJwkMs5LCIlk0rGKgUD2JEZKSNSy5Jzle4vCBz0gHsrlctnv97oslFKH01EINY4QmKyEkJRya+04Os6VkoXg0lIv5LR5WxQFpYxEJoU2o+NcUsEzKBGT5FxICnrFZOQQu66TQgshIEXX921IwkQAHCAwHl/ZKaO0tNYqpSmlED/wMSANTChZ8FDycTEM1pRlaQaXxiEScCdiQVmUhEalVAjkcmmVKoxx1vpZXTWX8/X1dQjBe5uzF09280BvgEs0TcMjQcFrvVsul09PT6KRlFIhJX4YUyxEARxQPApkLJucDjOYhiBFkn8qYCJETBD7gMyil8oirCAdIQfHRCi0yTgMdmAgCOJR4ElmzhL+IMHQRFB2yY4YOSmXV0KIx8fHPIgKSUsZMwlEzxxbq6oCbRFctbyeLdJOjHMhhkgT8vkamYxJzTQmmRaEy8PhsFwu+74f+xZv++PHj9vtVuuibRuQ6JxzXddhSEbTZG4KT2QCYBmlkVK8SZpEkXHLvHPWTnGNc84Fdd5IJbngADldEk/MeTdfc/QKGcTPt5UmEkcuDpD16Sv6SXj15zWenKvDhCpTpQRQR1B+GROXy+V8PsE621l/PJwoYXU145zX1axjuiiKw/4UPPni7ZdDb3bblzdv355OFyl1CKQo2tPp8uOPPz4/vyyXy74f/+Efvnl8fHx5OcRIl8v16XSqqpn3l7qehxBXqw0hDOvuqERx2MpSKaWcCz6Eej4TyQNG60n3wtoRZxg0p//zf/7l+++/f/Pmi8VyJcQTQv842Lqat237pz/+mVMRY7y9uf/06ZNWoLzo4ImzYRwgN+nH0YzjWNc1iZPDipQSivEsqWnB4RhlFigwVVVh02AYBmwRgUjpnDOjRXoA1R51QL6eaGKKsszC8r9CehjtNAHDF+99IIRQRin5bNeTF3nyUcNZEckPILySHCF/a+yTccn0/1kI3loTAs+dB14fORBFysePH0MIb9++pYQfTsfoHWWMEiKVcklaGUnbJiHMmDa3ZdJZI6kNwhuuKg18AEkR4aNOy58hTM4E0MAAGAriCqIbfh1jLFIa8C0xxjiP1jLOhZA06QNb7/K8IZKIoh7bas45bMBm/B2DEwTfsiydt+Ddt2371VdfXS6Xuq6XyxVjCou4T08PMskUKqW11qjK8VmgSh3tJBXp2gbQKkKSYNNQB9MCrET1fS+UAuSKKIYafLfbIZXmYQA+YH7D0/p3CGCCM8bAKoEdHupHlQS68zQITQA4J5gZ/vzzz0gV+HVJtG4asuGrxEgpHzyTZGcQm7766qvj8Zi7sdw+4tUADWOEmGvkbDTEGLu/v28upxDC/f19N/QuEEp9WVchBDua1zMAnGqQ64qiYIwMQ48AlOnOduw55zEEjHyklMMwohuTUmJXmTE2m1WMTkPdEEJEf+OD854T4tMSZU6NnHOaeiCe9lfCK8VQkpJZZtblO0iSpllObBkJyCWdSaJyWa5OJ9VbkvaZUA2wxOYgqacUk+cagX1bnkmM49g0jXNea4ZU4VwghL179x5HZbO5Rdv3u9/9brlcfvjw4adfPu5P56enx19++aWqqqurK1BUkCwXi8VsNkMJj42W8/l8vJzn8/nm5nocx9FZJqcva3NznU8vVgG6ceAxXt/cWGsJpUJKlA7r9do5JzillN7e3hpj7u/vt9vtarXa7Z7LsgSLWiWF2vV6bYzD4bm9v8OpI4wGEoWSuizwZC5tc24uTPDRGmcjKv2yLJdhqbVu+04I8fLygjsLcg2lFL8IxRmywm63o9PQ1K6WKzSgXdfd3d1h3KKUapqGcgYEBZOJ8/n8K6SH9XpNXuFC+I9KKbAYc4PJ0+rv6x6CJhPE/CekIQ/SiXUGQaTvBu8haRAoJYxMviVo/dbrdd6LwaXa7/eYBZnRVVVFvGdSee9H50OIgTIuuNTa+8iYGEcTI4UphbWGECalVsp5D+E/DgsR/EEbeDyNeQqXqqSpT5JSGtOPyQMdlzAT5K21kdHIKBWcMTo6O1ijlOKCM8bKogDQhHgtpWScz5cLXRac8+PHjyEE4yxhVHKJUIsv8t27d8/Pz+v1+nK5xECkUJT0i/miKmv4cQ6DPR6PCL63tzfPzzvnHGOTSToyUErwXiQ2AU3SpIDaVFGCzYn7k9fx5osFzi5OIabi4A7mBJ/nnFrr3Pa9DjTAjvGaKpkoAG3HzJwnAhtCA2ZoSN5gFiCspNnExHnL1ejUML2ycZ/NZqgB0eGh+chdQo6D4pXVB0+Mo++++w4Q036/R2Zt27auZ93Qc86kVFIrKUTXdkopgL8QYMdnTBU0E1KE5FPik1GMMcaZ0XsPLpBP0h3gL8FkYj6fhRhJCFxQ5xxlU7nnnPPBi6SZmCEg72yME7JKyOR55b0VQoZELwQTTAiRvfxs8irAAQgJmwUiQZMcGSIRfgBhZRzHq6sr/PXcLuTPjl2TPF3I4DNjrOuGGGlVzSD6OwzGGKN1SSkfBiOEms+XNzd3QhxWq1WI1Hu/WCy+/uafTqfTMNqirF/2h2HAUDB6H6E3rFRxPjf/8i+/YUwY48bRbjbzp6cdIayeLZQuHh63Simly6KsgUxmQo0Q4u7u7unp6eb2vq7r+XLR9b2LRIYgdLFarRhjwbt3E4GKLBartu2vrm6cc5TyyeheKa1L557v7uZK6WqmUXxc3dwik3nvy3oWCGVCEkIul8vp0lAuuFQ+kr7vT6dTb8bf/va3i8VCKbU/Hvq+N6PB5gceNbbzcE1Qv+KVp8UvNpkK13UNDmtmKjLGlNbDMJybi9Dq/3du+N/pYb5csFdiGBOGwDgnn/msNInCxyR4kOeZLkkbhTRnjkmKjlJKaER60KpvWzhWUu9dobj3DrU2fExFWjjAGmcIYbGYhKaNtT4QzwNjXArlnJOCxRitcVVZUkr7HnJJNITIGJdSMcYZA4k+ci5yOrtcLoyRvPCMSDGbzeq6dm4aqndd1zRYPZ1chjKwBhZToCSXVyGx8bTSlFKL1xSTAZFO0qqofwFex6TKhxk7tBVhNrler9u2RwNnrS+K6nJpGWNaF0L0TdNgkwRdjvcQsJqUU/FdYO0zEkrh9s5ZhINFoRFAEe6FEFAtzSsOqJHBROKcY6EB/Bb2SmsdC7dobKESgajEGAPOBim6fCTGyemTZPYXEhhPf5AYKKV4ZRRKSCok+XkNwwDUDoGJpPIZuQRAM3iKKPOxSQDsCygNGi+SSMPYfMQTAABdLOfIl8vlshtGSvlqtWqaBm0Nfh0qtfV6HdJINiY9FaXUfr+f5L6j8t5H796/f980DaVsHCdVxHEcd7sdcnnbtpLPQiDW+MgoZUIIwaQyxmAfwr9aNkIpQGLIgy6fXBm8nwrGkFTEc8+Hz4uPjGSfa9K8v00IQV7f7XaId1hhwcVHS5RjBeZ2OM/I8bmqI4SAvwDVptvb291uh24phLBabobeDL0xo9OqnM+WjIq7u7tz00gp7+/vi6K8XJrN5sqYScnm7u4Ob/50OuE94KPtdruHhwdjDKaV3vvVao1cBdYcIQSV+DCMSukY43K5KstK60Lrop7NCKVK6lk9P51Oy8VqPlsMwxBD5JxfX19vt1ugPYjL42gZF875ulaEEKU0ZgOaCu/jfL4sisr7SAjz3jkXtC6xPdb3I2NCaz2fLwlhKFPu7+8vlwtW+nD3BREoIHDGYIUCqTGaAFWMN7z3ZVWTRBlF64wvGlWLLgt8p4GSzJr5e9MD6q+QXBVzQ8DjhLT6tN7Jk6gcTy7QLBEQc62HEIDgRQhhnAJo69reGJhSSe+dM0FKgUUzwCwTSOV90zT7/R48llQrERc8sUYIwaOQWkkpD4dD9L5OvFWV1JNkUslnyeYF3S4KUmOM1p8XnlXSIa/r+qeffjHGaFXii6nruusapMCMtxBCfAyDMZgDzxbz0Ro4KY5mpGTqTgKJmMQSQrgQhNLRGOtcVdec8xCjSKMacEDxiNBpIvgOyS8Iu7LD0BNCjsfjarXEZBitCWV/08BxzrH9vqhqPA0ltLE249eEEEzDvvzyy/81WkckLYoCXTxYpyQ5pOfvNE/hcAZs0nTq+x4CBiQBEUgPyMEgHKOwReOMaxDSBD5HGRwGnUxD8d0BKMMbw3eNk4OI8OWXXyJMAMgCxoXc8JqIHEJYLpecTtRYtDiApNCbhxAo4Uxw5A+VzOvxWTI49vbt2+12yyc7LAuBbsgFYnPYE8I532w2s6rEu314evn48eObN2/6vsfZXiwWVVVwGqUU3kUQW6WUhI/eexOmLgSNWgJyWQwuA2g+KZdwrnEa8co/GcYAACAASURBVLZxetGBgfmK3AwU+3A4IPrgcQEzQaX5888/49GhmIUeBgbLuPU4LToJcyEHhxCS8HXs+/58anBuvYvnU6OU4kyasRdCdl0fQtS6+Jd/+cZaSwgtq6ozTkipisq4sD+eB+N8pF988eX+eXd9fY3q+Pamff/+PXKSkoWS7st3v7lcLmZ0XTvMZ0sfqNQaN3q1Wm23Wx9jIIRyroqi67q273VZgiLSdp3UygevCr3ia0II5UxqtVgsBFfeeyl0Xc2t8WVZBj/OZnOsx+LMIP30fU+5dMFD0YAJTjnTZTGY8YsvvsBtarpWSKEK7WPQZdF/6tF5d13X9l3XdZQzIUQwU6TKpVuOXajYZrMZeA3jOMZAbm/vgCzhGiJJPD09McZmi3lRFJSzl+MBtdevkB4+PT4lSAEScjHGoKTUQlpjnZuUl9quVUpRQjkXlJJhGFE0WWsI+TzBpkl7w1obIxGCWec5ZV3XORecc1JqQqLkxNpp3fz9+/dKyb7vm6Ydx/HlcDDWllXpgz+dTs5HrcvWNYxSrVRVVYxOxk+MTvy8PHD7/1h7ryXJjitL1LUfFSdkikKBANm07mvXuuf/P2JsbOy+0dpIAg2gKrNShDrS9Tys8MPk3Lcm4wFWyIqKPOFii7XXXpsLkciN+QOLBjL1PM9KSTR8AbPFZiglpdQhpNPpjP4X6DzX9QphNTxi3/eguHRdF4M3ZmKMaaXKoowxSiEFF9M4UsKkkoSQeZoTSZvNhuZBVJSQcRxjCG3blkUxDgNILHjCmKvxWmvnPPR5sDgwnUrprusvl/N+v4X/2+12zlmTVbtpFlSHveaC+9njGo+XCwwEYww6hlJKyDTB9yxYM2JzdAU659AbAUbpAkADcwPTKdfklcj68pje7D/oqGMXFsYI9CGW4Pd6vSJTBGyFNb+7u0NjOWw3CDBLiQghqszzQU+nE1RgAZphVgH50KwDc4bjsdls7DzB4QHtAdqLhwkhJkqKooiE1k0b8ziQ5W0Q1Gua5vX1FTUbSolzrq6qaZopZeM41mVVtq21lqYohGjbtq7rl7cz3IO1FlC/MdPxeGQkCiGm0RDOyqpxzhkfqqIw87Q4P0LITRMpxXke0VZJMvmNMcaYuF47OFFCyJcvX5DUCiF2u91CZoNU2vPz8wIiOeceHh7u7+8BAx6PR5prkD4P5UUyAWANM9jrrKtRFGXTNO2qZUwgy3/59tp1fVVVd3d3IYREUjf0Xd8LIShl4zhgWudqtX5/f9e6TIk5F9p2IyUGSdEQ0jjO3z0+MkLu7u7atn1+fvr3f/+P/X6Pi1BWBQKUl5cXxNRN0xR17XzUWluL+YbQlqZVVaOrSd0GgUStC+fcMlQO0mdI4gtdSCX95Ku6NtZaazHVmXHYLo8vjgCrKAq0cK9Wq9fXV0xDQRIGtAAFZ4QmRVFEH6CGCVDIOiekcN5rqbgSGOUChiHoJ1jnu7u7BVCFm6eUoUUXJo7kGhVydGxxouidtOS/+/q/Zk2vgWYwxjiPhNAYQ1kW0TsqguDqFgMKTjiLMTHCvQ+MSymh3ClSSikQlhhHFdqlGCNNLMWYpCjK1TiOzifvoxQqpcQSiZwSSqjgWstVU63Xq8vpPHTXp9fXkJLUpbF2mEwgyczOBa8YS94rzoM1LkRKaXVTQwrG2UCi9VYigQhhdqZUOgSPQNY5a8wcI4h3Xojier3GQBgTUpTexTl5M4fg6WpVG2M4vxVCP3/+hAsjs2DtNE3OW6VFcK4sq7oozDCW9UoL7ek8z2NdbbvrxfpQlJVWhWCCRmKGmVLan7v7+/tV2UQXS1VyLq31lHJCktZljCSE5FwYhqFtW/xSoOqbzaYo9PH4GqMLwVGaIIsnpZpn46NPedQtoICyLJv12oZjpCTk0WCEkKIoh66DCYb+NgSmbBahBB4Nxcol2VpYXnAP8JchCw2xPF+FZaU5IcRmswHoRAgBMgBLjV5cWHzM9IYr8t6jq2ipFfusIU+yJBdcCD52mWSHS/7t2zfvPdpWUWBHIguOoMyKabhyZpo4F0gCAGQ9Pz+37Zqrwo7T+dpxqYQQ8zQopRhJQNuQcvV9j9v7fjpud1vrHSGk78a6Wk3dLLkWNKVgJuPLskyJ9ePUbrZCiNPpfL1eGSPWzpSmVVs7a30w76eTlLLvxxij1teiKBiTtIk+WNS6vPd93yOsEYx758qy8z4gm0yJNE2TaDpfzjBPX758CSGACNe2m6Uojel7IYSff/4ZISfwqM1mQwhBbzBj7Mcff/zTn/6Ehd3v9+M4vr29My4Yl/0wXa691tqHJIQYxpELeXd3NxlXlmNVVa9vb19++8IZq5qGCNFN0+ictfa9u3DOt90lxjDb6fHx0XpzuhyFEIUrvbH95Soos9ZG572x//Lj7ykRh7vP4zRy4UOUm9UuEWXs7JNtNutK6RhCUdfXYWi3W0ppIYpdVc3zrHWhlCaES6mnydRF6Wa33xyUUlIKvS6rqppUFYhbUvb1qj2+vSulSl24mIbZFEXxejxqrevVinDOCLlcrji0RVEso6vKshhHM4+GEa5lwQhXQn96+I5xRlhyzrRtE32oqqrSKoXIucQccjOa/XpHKe36jgnGpMBAtiWWSimVZbXd7CDy70hQsmBUCM6QusPxpJTAoB3HqSwxwpZO06y0TpRIoZZm2H/IPUzTnLIKHvJWIcQ8m+hNjqpu3DXnnLOBEcg03tq4b9W/PNWW/f0MijAN165jjAXnGWPOWqVU8l4x5TLsAIzeO3e5nJ++fiWU1vVqNqbv+0ASZ39T7AhZ/epv6QLn0USA/jc2EaXzPEfnP8a8/EafH3ieVgSwaEHVl8wOmMlut3t7e0PwjkCMc/729tZ1nVS8KCRnrG3btl2djqeyLKcJ2qjWWue9T4QqqUBoUVI54lB0RaQWP3STLGBx3/chBCiKA72N8Sa1P47jX/7yl2/fvrmsccQYWyBBn6fXZgfPlVJAORIhiLgxqgWS10jS8QV5boO3OQshmREPx4DMDIWNpd0BbwN9AosGABTkXQRlAEwp4q/8BtT3AFKJTAl1zkHeDtUg1DCg1LtgGuHDINiYW9kB5SMi67oOSPEivYUsarvdpqwYhofBXv/www/QjwEqZa1VRYHzgCIHHmAYhpsKGWNIs7TW3759CyGczmfBuTWWEGKM6WLa7XbXy1Up4bMuISCa4/H022+/GWMul8v5fLZ25oEtXKPr9fr167P3XqlCKVWVzbWuhbxNTxvHESWBGKOWMoYbcosGbHw+4eDXm3Ecn56egJhhcgCWAkAE6JJgzmw2m+fnZ3h9YHovLy/r9RqzZVar1Wq1whRrSllV1YfD3S+//II2wJRICFEI7byPkby9n6z91ratc67rR5bIqhnP5+uvv/4qlHTO9cOklX5+fjns9pdz9/sf/yX4xJl8fXljVKDNG2PkERm0bdt1vZQyBI0kdb8/3Igh0c+z3e12XddRJvaHe0QDyC8RR0/TtN/v27b99u3lsN0zxlGux5gZtOVTzoVQlHLGmBDKuRBjJIR5H9fr7TRNKdGiqJQuz5eubVbtalMURaErxliKVAo9z/M8GyjfoQwDW1xVlSzEtesIYSGkeTKcS29D3/dS3sRg5nk+HA5AWfb7/fFy5lniHkQVpdRq1a7Xa0IIKAbQPI4xbjZb3DKYzff3d3h6VC5vV0xrF8My++8fdQ9vb2+LZXRZ6DilJCghWVlo4Rcyyii5Fc14nvuGq8hyWRsvfAhXEndSS8U5nzHHWKl2u27XN4EXEPjO5/Plcj2eTk3T4DEw1Nk5J/WtjYXk+jks1A2eDlEoyVmeVuT9PM+O/E35GYEScFLAIwD9YekWmwhTiOoWLM5CgEEhId3UpYIQgjFe17V3f2sZWVxLSinlWZgYCYDHwNOCLgLdoaVyu3SZYKILkJmu65AN4PkJIVBPghdCzE6yRI/3HnLZKbM2QX4PWVB90cQuyxKYJhCDOc8/WLjFIJsSQuAhsLki9xOgjwGUX3QeLHVyvMBcenl5SSmtViu4CoA8gH3gIEkWB0Sfyul0gnIwvu8wDFDp0B8GlUM7D1fRZ61TSMjha0K4G72EeAOldMrzMCClPo4j5l5A7XGpfoUYkZEwxmBG4eGu176ua4hSoFz5+vo6TRMxdLvZppRCjMYYM07jOHo/fv78qSxLKB9Dd6jrepZ1Mk6nU4yecQrgjhByPp/neXbOeR+7rvNtHMdRSIaVf319RXMWY0wKlqJH1dRai+/FOOFSOxeMsfv9HUBRzqWUOkV6OXeU0nmyRVGs2601/vHhu81m09TNYR9jjE3dpkidDc6Gu8NDWdQP95+2223btu1qM47jql0LpWZrr32/3e95VruJ1mmuV6sVptFdr1e4Fs7Fr7/+erycpZRNu+q6DnJ158ulbhpdlbu7Q9d1l75zMQzzRIW89B1jzKf4eHcAIbCoK8q5CyGkpMuyahoXApcyuAh1xa9fvyI3XWBVyIKBUrFarXa73R/+8IeqqtrthnP+/PyM0xVuE7wJRjPhRmO6CUwzpovDKCGf3m63XElVFqospJTbwz6l9Msvv6iUwA6PmS3NOVe5VmSMQfQzTdNczk3TFHXDpCCObfa73d3hdDpt1ruirvQ8IeJhjMGvgwRPMlk05fYaAAmwUbvdDsExnhwnJKRY17WQspa3QOef4B4Qt5JMqUZOIKU0HjaawlhjvYy3kkvc8MVJwFj4EFyGLOE2WB7ni88BPzelVJZl0zS73baqqt1uB+vZ9/3L64tzFs0jRVE4FxhlPkSotAMiX2iaS9kZzxBiYJEJKZNziSSYrKWmD98AS5pSQimP54kCPLNXEQuAf7KU5QkhWuv1ev3rr79qrWPyVVVhvN0wvMUYwZvy+aR678tmReiNhLDZbJDvY2ERxWPXjTGwaDD66/X6er0KgZnzyXsbo1dKEBK1vqkcAnNnmdYi8gjM3W73+9///j//8z/RjsQ5b9sWJBPUt2nmzsNJAFBOKeEPNP8cu4la5cJJ895P0wRyCOI1JCtIfeBuUfBcShGITOERse9gN0GWDt4i5O70GGPf95CGBDjmP4xYgMtBLILnJ5lqhdTkfD5/9PEIRxCR4f1w1Yj10JRUliWiotsEp3kmeezd5XKB1WuahnOBAizeY6399ddfr9erscbHUBXoxb2pSxljYpzhI7338Lt4mD/84Q8///wzPryuS+schDZd8ODFS0Y5EyFGqUGZ9fgEVPvh/pVWziefAuGUcEoFq+q6bCrvqWKCUFqUpdI6EXJ3f88Yu3u4x5NwKR4+PR4OBx+D9W673QohNrvtzz//3LSr9XYTQtjud4f7u5Di4f6uqqpEyKW7dl3HhFit1sfjkXOJQbxCCCnZ6dSllDCltSxr55y1PoTUrNvz+fzy8vbp0ydGxdvr8Yff/2iNF0JZ64dh+p//838VRXE8npVSKdEFovztt99AAoYd9O5GkcKdQtwjhWiatixrYGbzPHsfkKixLLQFW/H29oYyGCY17XY7Qm6aGcBOnQtSUqV0jPHbt6eyLENIVdVIqZ0LYMZzLodhen19N8as11shlFK6qhpjDCHMGn9/t2nb9vn5OfhECXc2dNdBKOZs2G0PnElrjFKqLOv9fs+VXJp1UHDmnG82m6GfUiSUcUZ5U7fenRkVnIklKIQdgxgdiOxAHRBXYa5iCIFzbo0dx/HTd9+5GJZ49B91D/v9HoQQ732R51+nlKK76WEs7L0bcBRvpLe//eRDwwTJ3Qz4s6eJxpRSYoRyzklKQoi2ru8/3Vd1AXKkUmoc+5eXl9PxREB7CDffQ9it6A2rikAV5wnt7M45QknMDVM4bUEpQRkhJGchN2EJl/UJEHiyTLuCF4G/AWsItnsBJTjnLy8vCAequpB5MD0mTzGWYkyUkKIovI9I5J11MaW2bZ11KPMiGEFz3DzP02QIIS8vL1BshUU7HPbWzj54H5xUYpom5/V608Z4ax7G9AJQ4pbcC4xSgNQfnQfJ7ZQyt9FhGYui+Otf/4otQ/qy5DfLVVR5mDaaSPHbadYoRYy/IDZwWriByxHC43HOIY8MBVDMdeAfWrcWG7rYArjSpfCASFnn4U6LN8Uz3CYLEQK/iCwKnix3O8aYGah//vOfUwgQ30aogeLE9XqN5CaN3jQNFtkYU1U30wzmD+7k+XwWSvoYKGOEEK20tXbVNOM4pkjn2V6vvZT65eWNEDaO8/XaN02DKihaeY6nYb/fv729yaJwMWhV9n1fVcW62SZGuRSFLFwea5qyuOx2tw/RA3GejSvKZrvbf/+7H//811+sdZwLWZR3j5/meUYtWhUldu3+/r5ZrSKh42y2+wNQtUho064nY12IX78+GeukLmKMr+/Hey76vt/v97N1+8OBUDZbI7WinEWXQopt276+vm82G5+lbiD5gD4S55wNHtcHegSvr68YAl/V9U8//9w0zTiOnz9/VlozxhIh1+tVF0VRltvt9qeffpJKJUooZ5EkSsi1774+P61Wq7quAB/hOoMghDuOHyIbBpjZNI0uCl0UwzgqrTfb7a+//iqlpIzt93swuGDclzQCRCkUfqGODo3C7z5/5kIUZckYY5w77wmlhdY//PADqBBQSdnv933fl9VNjaaqqlXTYBHKqgopWudmYz5//jyMY13XTdNgYhhchXMOvDuexYpAAwF/DGEx5AtBtcBPFrop7iO6atAY+N/zDf+3e7iRtbNXQDQXQmD8NhIOPoDl5ucUbgwlntthsMdZqO9Gy8UftBQUsxkINcawG6VdYj+ABr6+vnbd5XK5UEY5hzBco5QSgszOQkEMN2TKM3+AUCO/LsuSUJpiSiRl4OI2wA5ntyxL4I9FoRdsar1e48m11mVZgw+zoMb4nMvlwhiDYAO4ZUhrvPcpEWPmGMCbEsZYJaUQzHtgU9U43GZALsEvMDRoKqDlB4APkqHs6lNMEVRCOAAc/cWrLXAZyeIQ4C2AiQTEDLH8OI4oQixsdJG7w4BOLP1N8Ez4aui8Izkcg7+E3cT74WBwTnB8Ux5sAEsN8W2YY/6BFkwIgZyOMQYt1nBdqE6jNLqw8XD6ZVau3e122Cm8GVVuBOxSymmakHDsdrvz+Yw1WYBXOHVgpNZaJcThcMAF2+/3mRBCY55IjM5EbFaMMQS7aOBgaMQ0TTKGSG6zaZuq/umnnwRjxphp7J1zv/32248//ggUCExl2ILz+cjyixCiiyIxJoQA70UI8cc//vFy6bz3nBJsAWwWRmLc3d0N47TZbP7yl7+SJOq6KooiRV7oSqu03++besUon9S03W4pYZyDvycgOTfPdhimsiw595RSyERyLr2Px+NZa/2///f/9/3333MuV6t1CKks66ZppVJdN5jZITq2xjs3NXWrpH64/3Q+n5UsgAbPk/1//u3/JTFqXW7W/HA4CCXrenU8ninll0v3xz/+q3OurlerVTuOc1U1jImmqaCR/m//9m847SklKCjzDwMfcTuqsuJMTuPkbJhGMwwTY8xYJxVH8/P5fIY9JTcBggAbOo4jYhQEKH03OhtWTels4EwSwVJKlPCyqKfRSKGttdNo5nkmiW0266Zph2Ho+5Fz3nUDIUypol1vlNLn84VzUdcNY1zrIsXknIuE2OBTomVRhxAeHx+5lP3YIaff7XaXywVKB6fTiVImpcKola7rHh8fgbwgZaeU1nUNtX/OuTEzHBIaThlj2+0WrellWQoqAVqqDDj/E9wDZo4vln1xO8HeNKWBiiz+g6a/yXilLF7NOKdZAQpmN6aYMDkiJkop5By0UkIIh8FwVQVM7aeffvry5dfX19fgfaE1ylOUsWkyyab1ul2tW9RUQwjLOECUaEAfrpumHwbK6AjVthCGaV4iVmSsgN0vlwvwSqg+QNwi5VFFMDcob87zrLUMWSmFf2gaXxJeIaTgEgaUkihlQQgRQvRdP8+WsFtHAtyYzVqHCMA5v5Xc8XM8hrVumoaUgvfWe+ucMWZSSmhdxBiXvgFsCstT2K7X6+9+9zucDIQ/eM6PnQ0IQlNVoczFsngRwuSyrmnul0aFAAkNkCKblUrxBsDoKXfUF3lSGMIosD+xnvgn+/0e9xYOewk4Fh8DnwEW5kKcBcoEr7bb7VCIxiPBP6E8gFwHsaS/qSveJsViuWgeiwac89Pj4/fff/+nP/3peDxut1tEAEIIm6trcKh3d3fDMDDGvQ/o6kAsiV8Bwb5h6FNMjw+PRVF0XccoU0oXumrqliQ2jWboJ2fDer3uhyuK8x8fu+86qpXU6vG7T+M87XaHpl3143Q+n0QeNLLf7z99+gRNulW7bda7tl13w5RSatu274f1dtdNs1IKlKqClC54wqgui3a9/vbtW9XUq3Xb9T3nfLPbIs6QQnJGdVlMZv76/PT5d9/3fe+Cf31/ww/HeWKX8+PjY0jx0l1na6qmJoRsyq1zThW6KEsECp8/f0aK9vLyUtd1f71SSjebze9+97thGt/f3w+HgxBiKIftfhdjNM6mRMCuiSRd+24yswt+s9u+vr62m7XUigleVfU4jrc5aFqtt5t5nsuqwrkax/F0Og3DSEgSQkjHhRCo4acsGwWp4G7oGWNFVY7z1LQr5xzlbCnngsZNP4hH3d3dIZJD9+I0TVKrYRopZ0VVHo/HREmIsWrq77//fr1ev7+/Y+KeMYZzwQUvy5IJ5sytWWe1Wu13u5jS7CYm+Ga3TZQUVSm1GscxkhsG4/Ng1IWVBDCTUnq9Xk+nE1Ic5/wS6yyZNCi/KSXyIWsHwvxPcA/AgimlnHHn3VIBTj4pJZ3zKQsnEEJiDFDqApx0M5oxhhQJvwFNIeu8p5gYZzQmRmlIBBXOqqrKpkZOQAgJKZ4uly9PT5OzUumiXUeSCKVCK01ZpEQqiS4enseogmJIsxIqJKje39+lUsaYQmtKqCoFUBeAGKfTyRhTVQVlVEttjen7HkluCOHz58/zPM7zSGmilHhvjZkIidPcK1keDgfoLiAJqKoGDKV5MmXJGY2wKSkGziXnQQhhfRBCUH5TGgAGupA1Eb9rXSLoVkoZM3FOrZ2HYYjJc85hjzjnGqN9rUO6ClgfZHaEVxBwXk45EFuQr9Cji3lhTdMopd5fX7XWGKiJiiI+1hqzWa8JIR7tIFKioohfRAgBdYTlaXpd1+GY4kxDzDLmqUeoQLjcjofvCHQLjAtA/yC5GmMAdwAtvFwuoAagMQ2lIODI1i2MVWGMxTpM04SOaCFE1/WL2DVqTlVVcS4wC+H9/f3+/mG1Wm13u6evX4uivL+/77oOBxinEbljURRlWVLKNpstJkTit4Azs1qtLtdrqeVmvem7HsyZeZofHx+NnXwMUitjDJjvlKIzeVVVVdddORfGzFoX82SGcZou581m//j4+PT0dHd3t1mvzWzfXl4Jo0LKqqpSTA+Pj23bBkwpl1ALLgghq9WaUlHV67u9Sym52TnvCSFVUWupldBa6912xziTQmI26jyb3W6LXByJ9fV6NcY8PDx0XZ81wYL3gVFKKVNKd9decpF8nMMkpSxUUa1KyUVZlJhNXZYVKBjO+ZRIocvNelutGq2LeTb73f74dqyqqq4aJbVzTgrFGfvuu8+ciXEcY7wdicvlAkqCMUZJFUJ8f38/7A9a66EfLuez9wH0FsC/lNK2Xc3zTXF5GeODSkMIYZomxvnxdCrLsiyLuqrW683pdCSRILjEP0wklWVZN83ldPn++9/98ssv+90+pqSUlmKWrWrqVVXWlFLBpRQqhj5FwigXXGhdVlUdY9S6WK3WiG+EEkVReOXruqaJlGVlZltUZYqEJrKqmxRiqYq6rF6/vcyzGfpxvV4j1APmST7UG3DfhZKJEutdHIeyKJfQB6U7EC7qum6zNM5KSZYFEf5R93DY7nC7SA78ccNZ1gIjf88XWsq2wG3iB01glgfDLS9GKFCRBYh8fHyUUmzaFYDCcbbv50s32ySUWDWJM11XRNJSViH13ImYojFWZzl7RP2IfGF552kiKdGYeCK1KpRUMDSLWC5qEpxza02MXkrOOFVaME6KUjVNM4wdoXGc+qZpilLO8+yDFZIxRstKnc7v/dANw0AIFUJRwq+XgVIqhCaEpUS9B1W5VrIQXEWSQrAhRu9Cu14RQuq6/v7770FZQzcvLJExpqoqraVULJHYDxdK6TSYEELwhDNVVzUhxMw3trHW2tq/8SyllNBycM7913/9F6yzlBLtdSIPg8OypzynwX3oe0ekzBjjlNGYpJTReZZSpYtonbd2tV7DpSGZXWrU1lpUTZVSmAzBOV+tVmg4en19BcVo4Udh+1APADiGh0HdHpj+v/zLvxyPR2MMRrKAKgajXNf1MPbGTLRI681GcDmOs1LFPFspNaQ9gR+mRLQuU6LW+qqq7u4eAEhSypumVao4HA510zLxZpznUruQhCp0oW1/naZpu92ez9fD4Z5zHYJBAxmONxIp/FcrraS2s6OEMcop5bu7e1mUsipMCO1m+3I8qqJIMWmt+344XU5cqBDpw+Pn9/d3IYR19O7w+cvzl6qoT+9nEimnLIZ4Ph0pTYRLWVSb/WGaJp/I7HxiIpJUCTEZK4RUuuj6ybgwGadkpSWfpmm73729vTHGWOJ1USuhCx02m42ZrWBy7KexH6OPZVk6dGYEMnQj55wmtl6tx2GOgW7aHcC0SjcksOE68MRYolVZKaVSSN3pejgc/uM//oNz/vT0NaXYdZiX2dR11Ydrs6ofHx+EED0lbV3f7/eEkFuyaUxb10DPg7WKC8KEN3672iaXHvYPwQQ/+4lMUvpf/vrzfr2VJZdS+qa11l4vJ06TMeZ8fLuNeyNEcgK6RN/3yAkAVMYYU6SraiWE6C/DZrWNPs6DIYQEZ8qy7Ifr5XL5/P33+/3u0l13h01VaDtPd4f909NTU5WMMIZI7QAAIABJREFUpBjj3cN9yEwHdKSjR8da8/T0mzGTc45zWpYqpcAYGfvezjOl9OHubhgG793sfT8NNCRJBQ3EORdZEBsebSiEog3lnJVlwTlbr9e4VpQyShkhRJelVGq93XEp5nlelRUjFNknymbQD8W9ZnmA1YJV/BPcAwzNgpkgZwF9bSlI4n9R+o9Zo03kkQ9I4QkhIY/rIhm+p4mAUEgIQepQFIVSUusCFup6vR6PxxDC4XBQZSGkQqFGCNV3IxgFS/WGZkHKu7s7rAV4XYSQRTIIUBIo+Yh5XRYD8N7HFDnjWhfehxhuFUvvwzCMKSaA3eM45Vo3ZK7J6XQKPu12e+8DmMjQ7PRZyYBSynOHrVQyUnK9XqvmNnAYgJiU8uHh4c9//jN+KdKgGOM8j0Wpu67ruktKqdA1Cn3DMKBrCVwLkTWpwIOkeTgXHDPaj4ECgx4nPijT4dqAbgTpp8fHR9yo3//+94SQb0/PlFLUjYH7IUlH6Y8QUlUVxs+llFDBWwraWPkF2EH9A55soQwtA7mQTCzZT8x9W/LDpA0cPJyZ3W6HRx1GIqVUSgsh5mlijBWFPp/PKIkvpWy4n6WNDguFbUV4AQ2lkBs4lkafpmmmceacv7+fABeQm/DZiDejqRunnTH28PBAb8orIwgklNLXt9e2be72h+PxuGnb8/HEc/tVzENSARmjVCOlbJrVMIx1XceY+r43xhrjYqBKqUKXKRLBJSWMJMIZt87O06SUSihCjlNVVnVZMRKV1pQxTBYjhKw3awO6OWMICLD4OFrAD3GuttstEsTLucP+outQKYXc2szzbMyPP/5YN83ry0vf99fr9d//x/+4Xq9SStDzEEMopWLWrVouZt00jNJxnpe4WGQVemAy3759u7u7+/Lly8PDAyjLCFWXyVQ4yd77eTan08lYM01TUZazmbXShNKqLJFK1nWNKAG1NO8iY+xyuYB3i0ioqirO6Xq9vvRdVVUIXH76r5/LT98Nw4AjfblcGOeEEBSQXG5SQREe6S+jFEIsiH1DHieOtLWuaxS9QOHDMSN5rL0gBOjFokMD27jb7cDLuJWdpcB5JpTgBGouTsfjer0BQy/EoJRKhGB++42b27a4feKforn0/v4OS02y9u/yypWi9Dc6E6rWjAFIXVzCwjQlmSB7qwxbB7sGiOP266WUUnsX5sl0XX+9doRQKRUlLMZ0ONzNs+mu3TzPhDCQSlG+hx8Cw/p6vQJ3Q26LEBWFexwyWGSUVWHEGYPmBKcpxhidC911KIvaGMMor6pGqzIGIrjUqojJh+BiDNbauq4p4Z8+PT49fYMBBVpyuVzAw0spCX6LzcmHXj9jjBYavgRdMOBTQqUDNAwYINSxASaiDwDFg4W8C0h9QR5ReQZlAu1aC2yFcj2+OO7VQtaa0HafUtd1EOHRWm+32/PxlEiajVlv1immEILzriiKSAg4IZ8+fYIPAKxfVRUMYsoSjfDc+FtcTgykA1EKAhj4CksFAjolC2367e0NtThr7f39PSHEZinWhWKL83a+XLwLIcS+79frNQIXfM5HBxNjfHt722w2S+ZUluV6vf7tt99QAMCbSZ6hvSTBMcZ5HjjndV0PQ4dLAYgMVHchRFkWQHjP5wsu83a7OV/f1+s1VgYxHSpPPtxomsvAOBxaQpjWxTAMzoXL5fry8iqlYowJIZumQU+GEAK94iGE6G8tONdrJ4RAOyFJxAUvlBRKWu8U0UVRbLZbmB6wAGC8cPag+o7vvl6vMSUb5wrqSTjV1+u1ruuyqnRR3M+zKoqiLNfbrc2aHOB3qKygBfKCdU4XRSQEkkdcSiElSkoLNYhmLVJ41mViCrg34AtM0/Tw8ABfhfsihJBKTdY456vVqmpqcbk8fPfdX//6F6mUKERVViklSpngYrffVVX99csTjDUuLERBhBB3h916vWFClGWVF1k27SqkVK9WUuuyrsdxFEoxzhlnTVkuAlM4XW3bvr++Cs5l1uuEV8aOF0Xx8PCAEArpfozRx0gZs84ppSZj+Dj249gNAyjO2AtsQcyTYOqmUVoN40gIadtWa80JNcbEFAmjRVUyzilnIYbzeUAnBMmzNxaRlX/UPSx9renD0PYFVqJ5KshCwrlxCwiB6aF5YknMFPglmfgIQyFOAVAA3STr/DTN1jprXVEVQggmREjkt1+/TPMsuPI+CsFSIkv/GqwJYwyyJ8v/IkoCUQd1YNSvdBb7A93T2hk/6fuBEBJC3Gy26/Xm69cvWmsQZ2JE8bM0ZnYUtVNblZXW5d3dPfRTod8Ag5UyI1NIgSouy1NzCSHzPN9/vl+v14BWgHGBbCOlBFrKPgzMQJYDy4h1xt8ChoKhBLaGMDBmCUWQcVPOgvFP8JCofEBhCU2/eAY4pMfHx3mehZJSqZAipzzS5FP0GWZEpH9/fw/6B6IbXG9EXvgicFFYdrhnlENsVhsUQqB3CbAvDDpeeHKgQKi7YImu1ysmQhtjfHDOubKsMsHs1oOKUHc5wPgJ+EXw1mVZoj8o5ja0t7c3kgeBIKJUSg7DEEMCIFbX9TQdpZSUEpgPmgUNseBtuwrBh0CmaXTOUZrW6xXGW83zHH2Av4HZnaYppLAU2xeaQP5MamaDytDQD1rrqqy1LoFbUoqZNlxK7ZyniUglETlO01w3K601ScnMnlIaYlza8gmly73AFcCxkVluK8aI9QFMf1MkzIkXqjKEEFhDYy0yDEJI3dSrVbswBUieJwYbZ51dte35ckaQAWFNQohz7ng8IkCEi0UkscTXgN1B6aF5RPwwDOfzmTGG5rtbiVQI78L7+7luWikVJTxG4n1s203TNM/PzylRtA7AKyC/wX2hN7XjdL121vsYEuZuVVV9d3iY+oFzDhB4NjM6LodxQj8NUq6Frn18f19CXmwomLU4QrhuhBBwiqZpSiQxymKKUICPKUKWDaKcC4cFoFk2aDetMIR9q9XKGXOj9oWglUqEhBiFlMF50HAA88zzDOv3T3APAKbxJRFlkA/DQdMHklL4INm/ZA94LedjAZfgxILzCJceHx93u10ewShjTNNkXl5e+36glEmhKeEhJMLoPNsYyOxMWVbe+xAistTFXcGI+Dx3DDAcIWSRwEUfE5YbXfu421JqSrkxDvqaMRLOmXP+dLqghglRsJQSYxz5D8obhBJEYSqrX4G+ieI24IuqvHEuy6r67fkJBQakMhBxQ7iEn+NjcezKUmPk6pJxY9YFzCtCziUzQyIvsyotpRTPE7P4F4qogHcppXAMhBCoNMMgIlRfasJa60QpYYwyFlKKKXVZXAhT3bH1yBpx3JduR5SpcROgfYTqIti0wAeWPGYJ2BFPwKwAhcAfPqqPoUwNv4iutJSSUtL7SFL6aLKx6TaPgsADI+W/Xq/DMGAgJVwIqjVQNYCXCiGM49hdOyGkEOJwuFlYKCqS3E7l8/RNrXVR6HnGRBB6vV6bptb6pvl6Pp8f7u6HYeAZYfPeX/sryo/ee3hHfGXB5dvrOyFJcDkOcwixKpvg0zQZKRUI09a6XNTj3jnGhLWOUNp1PWHs2nWruo4pUUK01rQoCCHX6xW8F2QAyCHQGolbCcVsSinwHJi57joAJyRZulxKGWIUSq7aFZfCWhtSFFJSRiFat1CBlVJd153P57KqhBTH41EI4WNQhS7r6vn5GQ2P2KPwYcAcfMB33323eCY8DyA4ED3RVqmUGsbROltVFZcC9KcQ03a/k1wgZi/Lquu6siyHYTyfLwiVzucz1pxnIriUMoSIFF8qHYKXUj0/Pw99v1qtnl++FUWhtGacE0rLskyYs933iI/RDOTzwFSkFOjVBycKUAdanWQeTZYoSSkprpXWUikQIpYoHKEVynswupxzxribZ63Ufrevm8Y5Z42dc6MSoTR+YJf0fR/STcIA+Sv7e32j/6Z7QGCbMts95hf7+4FTNMcjKV9OvD7+7eIbllIEI7fJz23bQnnq1jl103QLSmkhZFlWQohIqfNgjvIQovfeWVfXBY4OnnZRsECZlGURK2wYJBPwVyZ72uU7juMEZ8P5TcmnLEtr3TwbMJcZ4/mrMUIo2oOFUCSR9Xq9YM3b7Xa9XsN4LVYGxI+qqjBeCf+L6PX9/f16vcKlgf1ZVRWEfZDve+/7vi9L7b0HOA5CJ+Dd1Wq18JFQkEcotHBt4RsWQVOR56zBBgEugIlvmwZgC5zKR21wxhlBVKK1saZZrV5evrFlZnUuJuG1YEr4KGMMFD0JIQsaBkhhIVcsdSz4JzgMdPN9+vQJ7gSACdIFeFZjzN3dHSJZhAcxRqX1OI6UMkQ2WB/c/6U2BooLrgoeA9vx/PwMWZ6XlxfGWNu2wzBM0yCVlELBfCC2QCgDFBHE2ZQHdXDOxtHO86S1ZoykFE+nY9u2P/zwA9afZRWTZTbcAqkBOgg3YrQax2m73TImKHVal1KqsqyUKp1zw3AbECKlFEIqpdGcQRkXQjRNXVd1jPHadZSmEEMIIabofZBQAZomrCFuAaINUMIQnaDeAwDnxlDgHMRoay263hJJxjkB4QNGi7KglA1D//r6infiYkIGPIRw93jvQzidTpSxuqr6YUiEwLXjKwMtxIIAf0cWst1uIbsEc4F4CO4HDLqmaZz3IUQu5DidDvv9/f19Pwyrdp08oAWRUjoc7pAqPT8/73ZbxGRVVT09PaFzpaoqRjmXYnvYCyESJafT6bA/oKsU19lYG1OKKfXDEGNkjCMDwFbi2AOXgwIHpXQcR/DrgKZer1eE2ghT6roepjGRtFqtnHec85hiiFFkQFjmiQbk1lbMhRCMUa31OE2R3ISOjLWE0phS3TTe+3Ea+77fbnfBe9P3QklKqVLK5lE6/wT3cLlcyIdEgfz/WqDxBYBIxKza/bHAgPd8dBh/g3F9mOcZjpR8EHVwzl8u3TTNMSVKmPchJRpT4lxQyowx1jpCSIpUCJkiRQoMRQTEpPhdIHcC5QcQj2AQb8BpE3lm1vKvUkqALzCwDDcQe4mSTsi92bBujANmcQhwENHbPGcipTTPs69rlDqgzQuOHef8/f29bVuIZwCjRPaHg3s6nby3lBFjTNs2sKQIOhYwHSAA2vEIIQh/0k0AKi04Es8qIPBD2D7ok4N0sd/vQZXDWqGWE2P0IYQUx2ligntrx2nkQjjvpnmmiYgsAAAgKOTxAzHrCQNKAjJLstYWIQTNd0u7HL4UpRSpBoApbBbNYiRw9ovOUkoJlxYrCRM/jhOeBOwGhPMIDFG6QJSzYGvn8xl7DWcDWJIQ0rYteOvGGIQFsOmA5rCtH9nlMDpAh4yxs5mNNVqrRJJ1NoZI6U1FyuXgA1uw2Wy45DgJqMdgDFHXdZze2LTYOKzPfr9/ezvCwePo4uqlmKSQRVH4EElKqIWQRFxwdV167yNJszHjOLZte+07QRnoD/ANwzAgmYabxPcSQpxOp4eHB5SCYO7Bj6yq6pdfflmtW2PNbOZLd8V7Qghfvn4tpOaM3d3dCSFWqxU2sSgKrYvZzMZZY8zleuFQu2GME4pSPHK4kCfEYC+Ox+P333+PEw5UEBVmdDsSQlar1ePj47fXVy5VXTeHQ6KU+hBSSqtV21+uSukQMCCWg8ILt4ciE24KCHKAm5pmxbkoipIwOs9GSLHf79+Pb5Sz9WqDvaaUjtMYfKCUwVDgiiHyw2BnnBbkCqgFYmExwxy5I+dcKqWih2hKojc1By54WVV+tjnm4MiwEdkQQvt+oIwqpXwMt77d4GdrpJQhRcIopKhc8GPX0US2dYUFhHuY/ymKrYu4B8n47OIDFvewgDkxz5Re8LKbOyWJEEoI4ZDE4cJ7LwQP3ldV2TS1UpIyAjUhQtPb+f3l9QUVeS64D976YKzxIaLGIqUghAouy1JTRufuNuxeZnFpJCLYDJaJXNg5UPtjrijCICIemed56aJgjEGY/vHxoSzLcZwQQ4XgKZUxxqpuGVfTNKfEy7pOlKlC66Kom+bt/a3v+9kaIUQiJBFCBaeEzWYezYzS3zhNSkrYOKXU8/MzyZJ8KaWuu14uZ+csYzR4TxLxPs6zW2rUMJoIzJe6DozCR4fHs/gEPhlGGbsGQAxw5DiOP/zwg5CSGMM4Z5yXVZUI+fr0pIui0Nr7UFdVVZbWOkrJNE4sm0UgSAuZGB4L5gzX2+SRcJRSYK9LArEgjUsLEh4S4AaYynVd//bbb6C+cM4Fp2+v387n4zD2X5+/PH73cO7OSiqSCGeCEiaF0uvy06dPiFgRHGBz2d9mYt90RNCkba0BtQK5PwL85bFTIinSeTaEUKAxfT9Ya4yZllQVmVbMAh7zNE3jyOGVYwzBj9PEhdSqsNaVZaWLkitFjFG6YIwyxqRQdVWHEAUXlNLVat2P05KUw4JjfIXWahwH5yCBZYtCxxh8pCGE8sZQCCSFaezbdh2djU4G70lMUggzzz1j3vtEKGMM/aRLdRqHSmY5FhwqmSfJowaAuggCR6VUd7kKIYLznLFSF9M4BucTl7oolJQhRsZZIolxVjV1u1r1fU8TUUISQoqiEFys1611zhpjrEXC570XnK/bFnS419dX6OOiOoICgMuzUVHFKctScE5pssa0TTNNs50Mp1wL2efxf4RIhCneeylFUZSE0Lu7eynldrutqhqxzvV8FerW6MqFaKq6G3ouOMSx23ZdFAUh1FlLEgFBkQueUpRKeu/w37KpCCVKa+/dOE3QwvN952MY56mqqhCD1AoDqGMMnHFKqLeOM84VS4QIzmmicN5I0OELsQIuBh+9oEIKkTwdxoFSlhINLhZacia765UmvmnblNLb9MLZTcbReS+kVFLa2fwT3MPHHGSphMQ8QBGGhuUeiA/X7zZG9PYTcXszI4QkIthtmJRU/P7+sN1udSG0FkUJrWl7PL1du9Pd3QNlzPkqkBRjcj7ghi/85bIsKEvWGtAwlk5anHKgIoi5lidEeM4Yw9laXB3QjMXbAVmCjaB50gulNKU4z45z571XpBhGI4RkQozTPBvbdd2lu1ZN3fW98y7EWBeFECKmxDiPKUVGvfckEillnf0rmCciz61Dnfx0Ok7TqNRN7MgYmyKRQqaUlpIGfBjY0AA98S1yi/XfytdYmWUkDgIx9IggdsYD+BB0UYQYJSZxE9IPAxdCSeXsoLQ+Ho8QERmGgTNOOQVrE1NTpjxCHJkZyZpOyJ+wpHOedYwNwttMHmEPZ4MQBEAtvB3+LSHEOqOVfH9/C9FrrYTkXLBEYyKEU0EIt9Yzxna73d3dHWwfGGu4YwsBKcYkBLtez8ZMq9Wq77uiKKq60Fq3bYvanXMO3VjGGGOscw5hklJKCE5IWqhiuB3gONjbCGgWYzLmJp3LGBNcmNmi7+TTp0+MCx9iokyVJU/EObfd7LQqx3GkhPddb41llFFBczzbYKEIIYzRlKL3jjHKGLXWgB0XYlRKXS4XY2ZCiHeGkpB8cLMRWjlrZ2PMbARkJSPG4vYonwK3rKrq+fkZtVO4cHwjhBdSSqQOuCA4YIzQ/fYmaiK5mBIRjKcYtVLw9NDiTIyu1i2JiSbCCKU5d0EWNYyjD/70fowpkpiUUkPXPT4+4vIiC3l4eEDJGremrusiy7MjCqGECkYl51IIyyjAvRC8c9YYhuWKMRgzIxoAvIljhrAALTtN26zaVVWVnLPgnZRCch5i1Eo751bNKqWkpHLWCSF1UZjrxTnrvOOc4Wx0Xae0jpQQTkliqtRFXb6d3l9fX7XWBSt9DEwKKWUiJKY4mplF4pwz0wyOpXNOcmGMQQEc6XXf95Aso5xREouysMYaYzkXJFKlNaXUlXWpy+jjZr11zgHavV67u7uDVIpxHkKoypImMo3jP8E9AFVPmTyz/PkjFhZzFeVj1WEpVzDGYoajWCIpJcE5Z1xKad2Mziao7yHn6LoOPeuEEEBPo5lxKME+BB5irW2aJsYIyA8IXchKVTwPyPwIi7ks/oGIG50ykPohhGDsAeTsYbkgoA37siBFiLlI5hEhzsX4mgVCxQe6PKgdKBYyWfwZySnArqWKDroOot2F74FsGqxQLBGwy67rcDeAz8KILM3bMUa4kAXrW9IUlHNRg4E5RtK6kDvBqkQEMM/zdrtF+vXw8AAoBv5ymQcHy46VZ7lxGq7OGIOBaDDTsPUoq6AesDhIMK8IIbA73vvD4YB8/3K5cUNDCErqtqm+fXvhXKxWlS4LQmhdN4qrFOiSCwJ8W6rrC7yJBVdKGXMTsYArnaZRCEHZTbzk+fkZ5+R4PBJCmqaBV0spYfgXy7qw0G5D6onVHoYhBI9KCRw2HHyMsSy1y00GJEuESSmJD4wx+E6e5w9a5za7nfMOgABsBKDLRUIG122pt2GFY9bSx5v/9V//9fX1lTA2z7MPARHSfr93s9lsNjD9RVHAKiFyR+PYUkr89u0b7OZSaFFKvb6+ptxzAJslhEAPOSHEGuO9Px6Px9OJS6GUQjxus6YWzTwccmO6J5zPpmnGfjifz1VROOdgPoqiQPvkarUCv67Kg2OxoTnEsXXd4KzWdQ2yHFwdHg97IfKMd7jbZTuwBSDdpawGhEttrd3tdufrZfY+huC8d9ZSStvVKuOODAd4QX2XZBc1zpSpd4hBAXSDs4tq1jhOJPcdo34OYhvMC6UUEwTgpAmjnPBc7kI0pjFUI8aAwLdtV+M4DkNvramqcr/fL1Q0mKb4910K/033EPOA06WzaYmYFsu7YE3Li/59pTqQm5GCwhJNxHijlHLeINqF4UMUfz6fQR6nuVHIBq+1Ns7i6+GdODcYMY/K/tLKtDznrfV6HIGqAwClH9p/ZJ6ZHkIoCg37KASfJu8cCSFIqdAZcLlcMAIFsAznfLfb4esfj0dUnJBRIVJOuVDpch/yQuh8fHyEXAHLXB1E/SSL+qHVGR8F8j5MACyUcw78nyX2X3RJwY4A1odMC+ByCAHdhSL3HOEbLR0hcAn4KK313d0dQCcUbxZf0rbtfr/H+JrNZoMmIJILVPCgIjcW4BRiy+CT8PxoL7gNJMjSe3hm0AcWO44MCYE89ghCFJvN1h1jCNE5//z8zRrnSBBULOcQC4tqPOo6yJkulwtcO6zw8k6A6f3QIQ9zzkGTERcVxn273S5YGflQ/ABTBZE4bATc2/LCUw3DcDye0VWAz196Yrxzt9Y5xvquG8YRq5RIQhEej4dFwODPJRNaYMMlZQ9ZOv5m5hgdx1FqTfIYbZYn+4o844xz/rHvBPsCai8hZBiGhU+FSv7DwwPmaaPckrJEwsvLCxaQUzrPM5SXKKV1XUutqqo6vx2xbpRSsPuEEJSyoii7t1dEzbgjHA2SH4SGhBAIFFCWw8PjlOLrN02DQaGoROYPp4fDAWAybiKiSbS8IcpEVQknWWRZGv9hrgmSlZe3V/zGcKMUllJK5105l1htoLWLa4l51C6KHKixI3FHcozvsoTdqEP0fQ+kd+GFI4fG3t1ixxgyUOQp5ZRGKQV4EDFqQtJq1VBKvHeUEilvs0wW8BnXaql+/UPuAd0ii/1aMCWasRFsT/p75QyeW15vISHJ7VGJkJQE4845yHVhCVCHTCkZY56enowx0P4llFZVFeltZiTKOyglhaz7hrvE82CJBdHCVQFAxDlf1ggBLHSsUAS+QSveFYXGefPeee9ijMZMjAlK6TzPS+MeIinEhinXrkOekgYridO53EPYdNgsxBGwj0oKlLIZY6fTCTccDKvlM/G7biTI6xUFkiUO/fr163JRSS77481LUAOjhqQBG4EzuiCHnN+kgB8fH5Hl4LVU4DebDSoBKH2jwXjJexBDgEBlrZ2mCdIXePEP4tsIncBiDFkwA/cZGwFjPc/z6XQC8RTOSWsNxDmFME5TiqRZ1bM1xtgQ4jj0q2q1wIb4Fji0ELgNIaA6jSBLCIV2B3Ebl3ZDS0C69d7/6U9/whyYmBtfKaVoFUxZchg7BS+IwxZzyxKlFEFrylXraZqsheoywwfiDCilQgwFZy4G6x2TwnsfYlDiNpgeNCHYETztcu+w/jhgOPnomQfFCKb/6esTIcR5V5blNM8IQVar1ZfTrxgSgCdHiICYFyECviMY+oiyYZtA01r0EFFSRihWFMX7+/t2uy2UOp/PjHOVG8HO18v9/T0uO/w9TARiduPMPM/gX9rZVFV1fzik3P+BC4VUAxO0uq5r23bxEJiCN89GCIGYDzA4HhViR4gCU0oQB9xsNugqxSXFri1sVKREgKYhaYdcIYbojH1/f9/v95xxkmWJSebuL1kjsljcYvRIH49HnD383v1+v9B5YYgQCEKf5iM4D3e4XDR8C8S1wQfvE2OsqpoQAucCGWpRNN57jBYuCq3UbZINft3HcsA/6h6whSxPg4GXAwYCx8DzyMn0Qcd7+QkcOCO3sj5NRArBCIUlSiSg9ggi8zRN3759e39/Z1mrY9W2ZVkO84TfhZQQBo5ktDeEcDgcPn/+/NNPP6FNBjg44gU8AyKmxa/gAt8gS0oxD4SQKCWYoPM8T8BhUkqMCYBODw8Px+MRSLrInFGYe/qBXYODhSQGxkV8YJ2inRt30lpblTUinQWkBjAVY5znGTgAOhUA6GN5EfVgv40xyEMRZOGTsUGw4zCUeDBwQFmez4xuo6VqjTOHsRNLMYZSCllGRI6wdDAlyLudc6fTaZmXR3In1PLJcx65ihUweRgqNhGXBx4CYSz2+nK5wKECnEGQu91u52FKkYaQ1uutfX/zLuiyIJHi/KTc3ItkHDwo/GqEh5zzruvatlhOVIyRc7oMz1oeDyMcYuZZ1nWNb4Q3LJEBImvQdpFsIa+OeR4DrIxzjhCLVnDEiThg6BG5dp0PgTIGWYXr9brZboUQqPnjCCHlDZl7hq3xeY4h/fv+x6qqcDsux4tSas6Z9wKQwgeApMtyIRoOeL1eL8nKApEhvcB+GWNg3VhBewvXAAAgAElEQVTW0kcXBcJbSqlUasnPEP6HEK7XKwkRuCguoMhKMCCwLaNiEHnEGIVSiG8eHh4YY5iOiWoB1sHmNvuFJh6ybPNCgY0fpi7GGA+Hw+FwwNHFQ3Zdh9IOVnW73aIMjucHz1AIIYV0yYUQnLVgWKaYpnHCqUh5FvRiFRd6CEz/ktWRLDGCMBGIxaQ1rgl2BA+PTyOELDcI66OUIowWRWGNDQEQd3Wz1il67xKJUvFmVVNGvPMpCfwimAjsLL7yP+oeIGAA3BPHi+S+rZSnxbEPHRY8T6tYYsaUUmK3ri5GKKNUcsEYOxwOiQTEpABJxnH89u0bTg8u2BKMU3ITAMBtwf1EKE0IadsWefqS6MCmIB1ZHDssO+4P3oY+g8Ph0Pf9PCvGqVRCax2Cc95bO3MuYrzJ/uCC1XV9uVzQlrWkoim/lvOBe6WUWkb9EEIWsikWCtCkzR35MZPNl2xs+UBEUjhzC8oksroW3owndHlQzOIUY5bcQHCECwZReBBwl8AQxY8F6pFZ6QGfj75CfPfFsizQGc0NiYvGQ8iDq3C7UmaCIsLCRiCF95loi2fANcZj4Hmg/IM9pSVt6ma2drfbd8PAOONcSJU4oTRLPL2/vyPwxx7hYGStNItECk1hm83mcrmU5Y2yjGAFqZj3frPZYEc2mw0GAhJC7u7uEEQ75wBVI59Dadfn4anIY+Ag4WyUKqCesghq4VVU5fV6DTEQSoWSKaZm1RBGfR7rDaorjj168herB/sopVwKVCB2I0j33jPOlkuhskrx+XwGWLTUHlyWsAQjA1E5jpbOmvYIaP74xz8+PT0hBcFfwY+iXgWsCWdGCQErjGLA+/u7pDeDjoQ+n1LqnNvtdnY21+sVQRIWv6prQIKEkOPxOGdpJtQ/0F0EnAcGB64anZJIvgEi4eiCbEoIQRIMdDelBOITYwyoHZYXXoFkut35fHbGEpJIiIwQRmgh1WyMdy6muNBAlhBhu90+PT1hQZ6fn3FZZJ59wDO5HH8GGen+/j6lhOAYMWJVVRDQRMqOrDrGWBTaOCuEmKYZ02vIre5LCEmU0b7vsLzImBnl8NzLbS3yVK5/1D3w3AiNcBteYTFzNBP7lvew3LC6gJsppUhvKTkjREkF6d2yLBkndV0uZQZMY6dZm9r7GzRkraXkJlCxpEg894wg2EEaCEu6RO6YiImvgMuDHPPt7Q1uBvHmZrNJKabkx2kAxMQFn+ZJKskYI0nQXF1MefbD4vwg1EUIEbkfguR5NTiCcA+LPMZizVNKRVFY4xAb4ouzTDGE48EzA3zEguPcI+9BSLKY4Pv7eyAwYBzimi3df0hXFzQcF0BjDqVSuGBAwFNKAOKwCwhR4cA2mw3aa2MevQdm13JaaO5LT3kM3/KE0HFCEOryOKPVavXp06f/w9p7NVmSpFdirt1DXJU3RVVXT3djBoQgSFvYGl/4xv31fCN3CRgMa0PMdKNFVYqbeVUI13w44d455BMxkw9t1SniRni4f+J85zsfmrTZMuiCqzJGol4WH+S9P5/PrW4RqViImimZc1ZK+tmhgRyfC/ejy8QYBKHLaJp5xpiU9XqtlNhsNpwvlF+EZrQoA97d3YERAL9VuWHAKqF8hYwN704s2naL2AZeE460UurTp08gemBzohZdnQShdL/fw85SutJa55gAgSLJi6XDHMhP3/eHwwG74nq93t7eOudQ/4A+D4KS4Tqs12uSife+7bqcsw9+HMcvnz/jvBhjdrvd4XBA1oUAE/4eLxdbd73appTatoWW+3q9Bq0LxSGYMKiUr1ar4XKJITDO3TCEEF5fX//qd799eXlRWtVwBHa8aRpjGiYEKi4fPnw4vR27ruuapus6/06r7ccff9ztdgiA4PWx8sgVACthiSoAjvgG2ZsoBCe0I3jvMdEB+wSPiayrNpr4MuJpu93mlA0E+LwPPpCcj8cjMJ9IElw1ds4SxFCaUoK/gfY+KcAvLo7stnYOktJCcH9/j5IeinAVJ8eDVBaZj8s4RSVxt0lrFqMHngZ/XMsMUihWZl+6Mjc+/GXGAb08U8ZSTBwHgHHnnJDCh5DeEfDrv3PhwucyLiaE4GIoPdM0eH+3v8W7bLsOUWHO+Xg8/vGPfzydTtoYYbpAGJOCCUMo4VzOpcOWEFLxuwp2IXDDCa+VAKw7LbKUiGLwwi6XC4IsADVN0/z00+hdjC6n4LuWWRdnG5TRzoWuVcMwSKVcDJOd5TxRwRMl6KVC+QQXR6zXNi2jHDT8GGNO5HoZYA5CCMP1sllvKaFGN8EHdIljKyOfFUXfFHYBVgxyRnAzIIPLote92WxqtYAQUoNZQgiSdDhImMVaxqzeFCenjlFD9a9yXoUQ4KIAd0KBBNBNRYRh9BHggGTig51mzwU1jb5er0qLpjXjZUg+ccIFF5OfcshCic1qs16vF7qLc6wMOuWcz7MFWrnf7733mFNmrRfUee9XXae5YIRyyt04b7fb8bIMIoW8DOccFRpdRkvWcmIIIedISPbeem+11tfrdZqsUpLSJe9EvoLKdtM0Xdeh8BDKWI4Qwvl80lpxzuZ5pJRpLVMK8zyiPRMLQggBDw2Lw5iQS7PLsN3u5tkpJZWUJGU323W/CiFEHxBj+ZyVMg8PD87hrDFKedv2gjIppdLqcrwwxSihl8uFUDZM0zS7tltppbTW0cfr6ZoZzYwqrUIIdp7btuWUoe0AuErF3HHeh+u03W6Dx+DJqWmaabTYDwBeYhE3I4QMw/X2du+DbVotlWCMxBS7vrkOl6ZvDocDISSnkKI/v70JRic7J5KbvpvGkXKmGq0bE1NMMUkmdaeH4dq2HaWsX29yzpSRSPPN3f50PT989YEQ8uHTV8Pl0pnu9uZWKbNabShjISVG8u16/fmXz4zREPzlct5ut32/OZ9PIF+AA6mUqiwVtehnRCAtqKYULqyfppExqrWKMTJGH5++3NzcrDar0/FIOCWM/PTTz0AIGGFGaz87mkjOKSdCYnLWd+1qHGyKlGTemC7G5GxghOYco4vzMDFG130XnHXTKDhNIffblZLKBe9GO04T5WS3uUHtEOHpUucjVAsZmN+tN0Jw74NRnNMkhAhSRxoIpZJySfloLV+UqvUwDCFEzqXWhjFOyF+CuRRSYlBbYjyllEhMKdnZ1mFntJTpc6lOV7+ERJ5SytXSheSsM8ZAQgNYGLI5a+3j4yNC+77rORdN26WUXPCMMec9ro/kERExfTdZzBgDG11rSsAZKlkI5QFWRCAAZFdVdJBVUkwhxL7vhZCZUG2apmsnNimtfQiAXH3JfEVJnEGujaUlihKa0jJ9DEfOlqnU3nvQEhBQkNKUF8tQ3Ioj4engOOEkIARWKaS+tByjCwG5FFbvPceJFIIvLCNiClbIx4hwQSDBj2hRTjydThh9HGPEYGFWqLGhyCyiP7FWFJCNMcYYpyxRUMul5EoB92dKqRiWjNPZpSVl1a/AWUSpEO4QL1Hrpf6x3+9fX1/hQlKh4QJLEZQl5xUXrWnOJT1KhSB7PB5Bo8ROq2ACyuaMLfMBsXRCiO12N5fiLSlKaqTQK2Dil2dkDISl0+l4uZwRZiLKY0UvQJfubl86jW9uboGbYQ8rhbnfaZ5mYEGbzQYi5HA/4zgTQlBFQNQFljAnFMEQY4wSCnd+u7/95fELxI5Q7wUeZURLSjm9YkQpRhRLa00bJRal1BAnnGiEw2yZsCRRysLJwvfhVK7DFfBsSnG1XrGliW+RyG3bFggStI1dSCEGIQTlDDP+ML1OckFy5mKhOYGDxARL3i5FI0pDjEKI1aoPzm+3hFJq7UwZYRzFQdTVJaq7IMpbO0spf/75Z0rpfr+H+ilg/bu7u6Zp/u3f/q2+JuCufd9fLmchFm7kPM/Wzu9QHWObZkepaZpcOooOh0Nswnq1Sikxxp1zjLIUY9s2+/3OGNO2CyruvW9UMwyXnNL5dLq7uwOr3THmnQvB55wJXfj3lNHJzrdFtLRmEqkob6YYOWOCCylk2xjnXIiR5CwWBTlJF308IAfOe7tIX1NCKdFa/gXcgzGGg39JmRAix+WoILuBoalFyAqPkMJugtMLOXEhlFLUNLvtdr/f4y2y8nW9Xt/e3sJChE8pL4I/sIPvLULl7KsycQwIfu3/RLaLfQ/eISx4zhlF2rZtgZIppdB6g8Tm/uNHBFPI0MH9UKXNMqX08vKC7YVwEmEgKeFhLmNhZBk5WVE4gPWIUuGZYFWx52Cy8Q/QP7quu1wu6HcDBpUKDQywMtCkm5sbWCs4GFpkQgDgLj4vJaQjtWMcFgqVyVorxos+nU6ohRJCwIsFqazidfWyKBrHMmK6VkegUjWOmIK7goghIUwIFdKkGtP0HaWUSSEFJ5xlRmlpEEEVClbp5uZmmiwgLJAC4WIRGUijd9ttu+rR7dX3ay5FLPoix+MRnhjrHIr4BASi4RGRoeORUUZCUInvwEBD3gousFpD7EOA9c557z1WDwv7/Pyciu66L+IQSDFxIiqLAcgD1nCahgpG4ZcZY8BPMHy4aRpg0Iyx4/HIKONS5JyFVm3bPj4+bvqOcY7yktZalfkWUkrVGFKAVqwJMA1sYKBDCFNKceuMf6P7B09RqRZYN4Q4nPOu7ysICWoNvvCm8LmPj49A+ZxzlBDBOM0EjdNYWO894zTG2PdqniVctTGGBBKip5R6667nC5hLnC4RntY6kSylkkrGGGXR0wY4Bh/z8vKC7pwK+wD+NcaAWu29/+qrr6DgC6MBSADGh5aZDZvNhhACsQOcFBiHEMKXL19q+xQmC8Apvp1OQghGaWOMkjKl1DYNIYRmdrmcED0ATgA86L1HxxVj7Hy5hBBCWvYz7hlv/3K5QE3EFc0eeDVaasDqT0XAKKVSyK7vQTpQUmolvfehMDX+XPewWq0YWu0Z55wzQhGF+bgYRJwcLH0NUqqtXKqjlCxYUM6bzebu7g52s+s6rRUs7+FwgByjkLLvVygM1DOGl4dtWkWYUYjjnP/TP/1TtbZwKsaYm5ub+lcw1uA4uqLfkHN+fn5erVb/+q//Gks7BerDo10aqnPOmWRYgbjM+7YgGODV4j3BfAshpFKQ36oRGXzkarXabrcwvrxMJqgsghqgHY/H/X6PPDeEgIpWLOxMUiTjQWSqMBq2Fx4f8BqcLi3iZSil1kgQh/x8Ptc6ZC2w4yJwusgV0FCK+3Fldg0r2jj48/qOCCEQ/AEwVZ2TtbZpOtww3DZqD9baaRraMl9onmfkSVh4RAaEkOPxiBky4JhRSmNKaHlhnCNVgmECqgb+FS9cZywanBz+F5g+KsYIF0hp4cSmpZSCMGaWkRLIyRhjLASn1Gaex5QCSgLA5WpTK9YHGx7TqgkhXdflHJUSaDJXSqUUAIU9Pj7i41BkwjrP8yylbsusVjg5JA3n8/k6DFgrhD51wyBIRy0oxdS2bYiJC45AYbPZgHqHkgx2BV4xLsI5l4q3nbFuoiwLyY6ny2xHIRf1b5wCxAfzPBNKRBEgqdYTzvLl5eV0OgGgq/XYTJahWHA81X8zRjmHQrVXSgrBh+G63W4JI/EdPR/2br1ecUKbtpmtVUpkkq/DVXrZtU2d3oHkGCYCz4Ujg9gRrynGiGzs5eXFvesnrbkX5xznDsuFq8kyoGK/3+PVbLdbFI0QtAHaHYfrerMxRmutnHPep3EcwAuAOUJIl3JIhbuFlcH/cs59DPBtqZAYEdJhMyDNreE14AfEpk0Zy4ZFZowp77x3UvC+76ydS7XjV0Xt/19ffyL0WvuKEaGjIQvEL5xhmMvK4sLawRyjSGCthZxhjNHaJRUAfwNrcTgcPn/+HAv5Wiu9Xq9ZKccjlEDEh4/DmedFv89a+/r6is4XYIjIoGFxoBKDp0BIiMQTVub+/h6AOx4WDQdIymCzcP+wyLnom+Ixa1SLvAQXF6VkDbtv3ilQYv9BtxzGF9/BI9SuMaBSmGcAeyGLZNgyDKPo61XROrB4Ea/BFSEj8UvjzMJtwEmG+Ysx1uY7MBlijG9vb2CvXi4XvDK4YVQp0K+QijY4YlteRjfX61PKBJeCS86Esz6GFEOihCWS+/WKS2G9I4zOzlrvhmnEJ+K1QqkwhLC0OOQcCrUfhZOu6yhnymjTNroxhFEfA+WMlgckhPjSxgwnhyMHcmEostvQ2sN0ivdvqnp6WVosl9NoJ+ct49QHN83jdbi8HJ4J/RNJylTG8+EiMUaIy9YGqM+fPyOYwEqCOoXFxF5FWxnIBdjGcMwwAbifrusY59qYTIjS2jSNkLJfrU7ns9a6bTs4exQnYoyZ5Go1+r5/enra7XbOOaArqNNg9+bCFUQEg22G4AAzwbDbQevAgZJC4CyjrusLCIxXgBQcVhhvB6mDlsoonUK8ni/TML48PdfhUcDWGGOAbQXjgvEc07pfccru9reM0M16s9/vQV7HlmvbRmvd9/3t7S0i2li43SjdpSLwVQNWmIi7uzs4YKwP4pX38V8sIwwqW9KVkVZwNuiFopQiMW3bdr/fa62VVClETlkKUQn5dnjNMUUfwI5DJTyVngHU80Wh+eK0wg+xopUEsxALP77CfbjU+XyG+m+t/oKHjQhpuFxJykbpdb9ys7XTzAil/0Hv8P9RbAVhihOac6Z52f1M8JpH56KrUdcdyWwNo8BcopRKvmR/8CUx0nmOT09P6AZAKEQZ5UJA+hMMhwrzIYYK71qQSKkT/ppypoQgN5f5Rc/Pz5hDgPcHhWd0CMfSWlmdyvl8vrndo6n19fg2DEOjDYwUrobwBCkeji4p05qklOv1xllXWWXwTDgbaPwB2lBrFfhNV+ShcLaxCfBrhBDUhGsUjJIGLXR1iALh3mDoEazFMuOwLiP2GU4+Ph2hKFYmvuMi14ujwIujWEESIQSWa7fbwccg+vYLR9tSyuDZu2719vYWQuJcKrF84e1g455OJzdP9eXCiANjYUwgiSaFso2q1XUcQFrFroNtHcbBGIP+j+PxCICClQm4qUgWGmPatj0ej7GMY8KZr3hjfDfRpLYfxuilkG3XIiAAEQBHcS5jsZGoAR7UWn/+/BntLBXTv16vNzcqBGftxDk9nd6klIwRa6fT6QSq6+fPn/EKSsz7a9M7K+wvLA56FPAel+LE4pyy975rW2vtNE6bzQa6kJvN5vn5+fn5uWbAYNzxovq+2+2AdXRdI8RCrDRGG6Patj2fr3PRVIdnRe4ITAm7Avs8pXS5XF5fXxGkY2thDFwIwdkA5AdwFixJ27Zd31yuJ8ZJ1zemUc65pjUph2EYcFcwo9XlR2+ttYQmQlkIrm0bLnhKEWYRbPJK9MCdo5AGSwq3MY4jqnpwBvCUuUiu9n0PMgKCLYS8MNC5dOPiaKB6hzgdTMWmaUyjOOeEJspyTP46nCnLq1WXcwJS9NVXX1lrKcvwUuPlSghpVz2SvMvlEkKc5znMIRYZDNBG6qHGFyKnet5xh/EdVyjGGJMPMXSmSzlM8xBjlIqb5tcxOf9x91DtBU2ZEELfyXS/P0vvsZT6X1Yn/zBGCGnbdrtaIw/ghZg/DJenpydsMlBour7v+26a7DzPyKqwWat3jYXWDUADjh0f54qeAbYFbqmaV/xtCMvspBrKgbsGtB0rjh8BmXVF7AjfB/cJJhKWBdv3r/7qrzjnLy8HZAaVeBdKWxMKmPbXGWQK0RnuHBOM8VWXF0t0PB5hmk0Z5GCMwbZGRIyr9X2/Wq3QygCjWdeKUorCNaA5jLFNhfgLRe76mPCXGJSEmjZiw5qTQb74cDggdMIKQDcJ2xG6fDlnQBBLTq0kYZQJTgIdppFLIYRggld7Bzielo5/IVSlNqxWK0TcpmmO1/NluLZ9l0iene37Xig5XpanwwZ4DxUimHVF/k8UXvJ+v0dGSN8NDkJ9yHvfNA0eHM+OCarWTs7NKYW3t8Pb2xshLMVc6/n4INxw1deCO7fWSik/ffo0z/Pr66sqoqepiCvgTNUIHbnmPDsgMzlnSJ5AXvs6DpRSwuj2Zpcp4VK8nY6ID3a7mxQjLoLrA3xD8EEIAbUXMSzCBXwfPd4474iga9AthGCMwCU/PDxg54gyqCO4pf6HrkZWJqzUmAMRG0glQxpyzg3qEBjkXkhxWAHsVeccoJsco5aSlu48TqmQUggxRk8pQcCE6IELHvwiBIBji2MuhCCEgq4NZwOwIZWq0vPz8+l0en5+3u/3ODK/+c3XSOjfx6PjOJ7PZ+xwlBIZY4+PjwiSkPBBigYYabAuc27HKYdIKeWEkpi8tT4uet1AftabHoEarowKDc4al+J4PDayAdhVCQvYG+8dFepSwM2Au/gyGxEWmBPKhWy1YZkIyjinLBP2F8ke2raVKGLERCmt2QOKY6kMNSOE4B+pUJh4aZ8mhGRGc863t7e79Waz2aw3a1pmeD09PQEWdIvUJeWM50xQSUPrPKw/L+2muDHU2VAIIgUWhEXb7XbIoGHuQXoRZTLBXGabVOI5FtROM5JcrfVlHFCY4pxzukj541KiDMeuqGi1JjDcdrYAhZFf4ziFEKogD0wwvFoF2em7BkP4RUII6uT193EskZfAiqHSgDANaEmtk2NnIwGvPWW+kH0BmoV3MuwojsGUkCLahcVHtAhwmRZGh/cePclAUXMhXLGitQXEiZaJdcrohafgPS19oSmlHCP2MVzdZrPB2UspUspQTlhgcSkpo6nI7NQXRwiRWnLCajCF6A8fwYo0N6hZMUZUPhCOVSART40clJaODRg+HGOkHahDhqIm4mwkhNRoFF4BpwDJR21Ods6dLydKqA9eKqGUIZRkkk1jYoymWbgSjDFsnnqmaqgBAH2724EggFXCgyD9ijGuV+t5mlA0o+gZCn673cbSsguhMFiTOoQK+2G9Xq9WqxhdCP5yOYFyllIYhsvr62vbhKZpuBDX4RpjWhUlUc451grHDeELpRQCEq+vr99++y32vDHm8HSo9QZ4cZj40Y5aa+885xyKh6vVivOlawoJASFkt9vlnGc7U0YpY8h+mqbRRjvvM1mYk6+vryEEiPd1XZdSTkWKHHn2PM/H4xEZMNrdsdQxRsD9WuvT6YTZxvDN7wfiAjlQSmHedcV8IIojlmZVLxVx1uZlAgIZhlFrQ9jSuk+IRAIH8qSP9r31zzmbtrHWkpBTitpoY0yKS2NyDdTyu0Y8WkS3AHPRopocQgg5A9EKZdATFvY/5h7+pPZghFRcGKlw3kzbmLaRWqkyS4eWhiDsV1H6rUThEVFKcYXOmL5vjVGZBsIiV2SahpeXl+t1pEwIqZXpV5t9v92DKslLkzAtug4wr6TQz2vtBWWMnDN04gBSvb29pTqmXMrb29uHhweE/KR0zSDeRAQNyxIKsR3Ib23dwJOuViv0eO/3+xSzViaGRDLdbnarfi2Fyim/N5SIPbGHwILAeagWHzE7rAAmhKCcMM8zAihsR1lkyWFoQKWA45FSotcXBgWk+5r+0yKNBacIJ4HKZ81AZRmQC/gFich+v//w4QNcC37BFMUxGEfGaIhWSBZToIzkTDDUlzJOBPMpZEYyoyEn0zVCqxiiHScSkuKSZ5p9XHcrzWWMWevGWu+co4yE6FMOmSTGaIw+5+i9ZYw4NxOSvvzyMyNUS9VoIxjvTKu4YJm2upGST9OA6F4pwRgxRlWnKKW8u7uD2zPGKGW8jykRcAKMaff7u5zpOI4ILeH/sLtSIinT1+PpfLoej+fhPMzD6Gdrp5kK6pMf7Tj7OZIYSUw02WATTYQTwskwD8fLkXBiOvP2ejweTuN5tIO7nq7Bxla3wYZI0uxs5pQKLo3mShBOTdswRmL0+K+10+fPP0/T4PwMjQDw1nCTovAGTWOkVm3fMSHm4M/TMJZR24fDQWsN4kMsIldAwEjJGEIIx/PFx3wd55hpCFlKY0e3Xm99SoTz6zgSxu8eHjJjidJMF74v0ixkuk9PT+v1CjMxQ3CXyynnOM+jUoIwFjN5O52VaVab7fZm3683TdcLrjmTKRHOlOB6f3MvhbGzE1rYYKmgplMhu27dbG76kG2kNKQkjb5OI2G07brr9Xo6nqCj7pyfptn7wBh/enpGmIIqNBwwDmYIaZ5d03RS6vV6O47zOM6cS62bYZjatrPWcy4JYdNkvQ9Ki5TDOF1TDrMdM4nOzwBynp6eY0xd12ttpFSUskCoTyRQfrwMr6eL6vos5BwiZYkLYhqpNL9/2EvJu64hJHHFhRaEJCEYo1lJLgVbdy0XyTTSGMFYsm7MJErFGSemUdpIyjLjZHezMY2iLGcSh/Hi/ExoyiQ2reaCtp3hUruQfMzDZLnUiTDr4zD9JeY9oDoK8KF+E6EusukaytV0xpeJvogiq610DuQfqjXYV/NPP/0EZWBjDKWcS0EoZ1ycz2dkcAjlQhEAqH4SXhTQcNu2h8MB3hLZPQKEaZrgJwghq9Vqt9uhngbZUeSzLy8vvnxVhzRNE2EUQdb1el11Pef806dPz8/Pssxt7/t+GudcZFkh51kdDx4Z3gtwB7qakX8gosQSAbAmZQwqfDDkslGzCqXTAqgFVh62HqkDdEFy4QshpsbVIO2LBB/WZC66LniJlYaP+0SpDYFeKP0iokhwI5rGyyKEQP56HMfrdZjGaZ5d368459bZaJeOOSEFoYRCUtfHGOIUJjin1Wo1T1MdBufKvCYhBOdsmgZKlspKzR0BMzoXUFhzswU/T3AuBM95UbcG3ojfT2lRKFmtVqAz1OrL09MT9i0vPYOI+vGFW8JCzfPcxdbO7ng8OedIiDG6nFPC/BJCeJkshGgDcLlS6n1e1TSNHS2wTaR3xjQxxNP5NAxjiNF7L9ZitnPTttvtNsRQ5TEqcIpG36ZVNZ9AHoldAWeGVwmMcb1eJxYqTwQNH6hkgOKFNXl7ewPPJ6WUEqGMa2VyIl3XHw6vkgvn7Cf3jhUAACAASURBVGa79t7/3//2+7ZtN2ntg5NKxBRDaTQhRT4IG6lOHqz6DSEEAPpQEi0DZBkhOcalGK61mqaxaZqcE3KLVNh9Xdc1jZ7n+XQ6WZ8opUY1CJCfnp5++OGHHDPJZLVavb1hKglfrfrD4cU5RUhuGpNzNkYTkh8fv3z48OHf//37pmnW6zXg5cPhcLlc/vqvf5tzopTEGLUWYBIRQqSUbadhdiil6GaVUqaYpRQ5p81m23Wt9z5GDO0RMSepxGynRFLTGsqI1FJrhckTSrXIaGG13nfY4JgH562dtVFt00opUsoxBkoX3gHslSpKhdhj1UTTwkaBIQp+4XNWK10/6891Dzg5qkx3wa1X1AgfY4tsYd0HsH21CgpsvWlaY4yQAqcXwt3DMHz69JsQI6VkGAbBFS/aztgZCJkrFbXmEDW9QkIHeo9zDrxVRH8VyKtDdFGJReSOrQZYE/Sqit6ch2stheWc9/v9x48fIXcDL4LuBJwEtEHAFiDwh1say9RSmC0caRRCeGGG4ITn0iMSY4RJqqVmLDsrdOFUVHHA6GVFzxmvA3Sjtm1xP3WJsCwoISLhCKWbGnkJug3sMjrNVVgJiTZcVC4EBJiYvu3ApEwpuWKYtNZSSBJ8jinHxIWIPgjGlZAuJEIIyrB19SDkgPVhZUYTthluA2uOUZekVIyRwaCHAKdFCO69g4nEs6A4FELkfNkegJhwJq31eE0oe6D+kfNCwMNJQ/nKez+OY3j0MUYfnFLq9fTEGRjGiZE8zzOny2C+MfjgbNO08zgpIXNMECfOMc3jRCNljOOeSWnF8t4rIZSQWKUUYnCepAzlBrS/AH1G5Gu0WXU9tooQYrNan48nJSSmsIEaxBnfrjdL2UAbgEjAjkBsrS4NB6EScsZxRD0ccBnJ2TnX9CseA27MzbZvO5pJCvHDw4Mx+vHL57zwTddIajnnUvKu6zA9HnEYYBwpF/kceDuw9VxRF0YiglAJoWHKDAQ/bF1r7TAOOaUck5ASXCDO+TSM0Yeu7bErat4MQBtbC0c1LaPiJJg/bdtCvBkGShdRy1pywH7ouk5KEZMnRTatoiYpRc45Gu5gr/FZSFiNNsPl2jUtI5SkzCntuq6Ola3Pha9YFEFwMDnnXdc6byHrkDNsiHqPQaUy6R1hB4KJWjJBNat4/YQ4Gz+C6fgLuAd8hihTbXkZWYybgDviZXRo/QctJWu4/c1mc3t72/d93/dd10upnLMvLy9PT09Y8dlazmlKiSsOm45LwV6wQq5npX0UCElFyXFjtAglVo4HWnsYY6g7hSJWAYYuKv51EUEJBSv5dDrVMThaKoCzyF1gnhDXwBL1fQ++LykDG1A6RlUg5wwWBywULUr3sMu5lHxlUa2AM2CM/fa3v/2Xf/kXUZoYYO7h6mDxUU6M75TDa4Ufd45bRdaP9QHTpr4moJDIP3AFGEocMBQnEZIAiMD5wQISmsHiP5+vp9N5s95ifZSSmeRY5M0JITjq43UAhziWYoMoQqpiEUpaCrk5R0KI0UsUAv9BCIG7JWQhWdUuB0QsWGcQ1bFceMXGNPAElYEtpTyfj0KIGP35fGzb9nw+Xq9nzjmwqZwj+p4YI/jfafYIvmL01s1aypSCHS1XS3kfaRbQHuc8IRQuuabOMUY/u/V6zSJpOxNCEJKlHOZ5FEJikVFPQmBhjKGZPD09bbdbbC3kc1IgX+ILC1xrtLBdhwvjJJPIBd1s1sAtQwjTMD89PaEpBK0SvLDjqpNommZJ4yATpPUSBhWCWU4pBGeMOp2SUqLv27e3Q9+3Nzc3w/XEGJUS4Q41ZmmLg3sGp8AX8XkpVb9qm6ZpWh1CWK07QsgwXrighKYYQ9u2ITgfspBLCQdH7OXl5e7uLufsrNNGJ+I4pyG62Y4ppetwdX5esW4YL+N0bdvWuqnrm7fjwRjDBe36VSZxtnae59e3l/v7eyFZ33dd1wrB9/sbxljf9zDEh8Mhpdg0fUrJe7Q3aiH4MAyYZgETDAMtRLpcT21nKKVNu4ggSMWVWXHOrbVCMECdm83KGAN6VS3/4OQiQXxfYYV701pzsYzCJoTmvOhnV0srpURwGctwFFJ4bnAbuYhH4fq+NDmiLekv4B6g+4p4mRQF1mqwXGlcqlkFeTeMGu707u5uu93udjutJbYpY+xyuTw+PoJUfrlcVuv1NNmu69qmZ2WEJyleEdYfGwVmEbsNLVRosEIujws+PT0dj8ePHz+CeKqUwoTI6kheX1/f3t622+3NzQ0M4na7dbNFMfynX35GRDlMI+JcpMlIyYEsXS4XuC44sxgjSmFw7LVpzhdNYBw/oED7/R7PjvnSOJ8IW2r8IoTAuDQAO+gVx3sFKxFvlxWKKinKw0h+ISqFmjn4cFBjxfbF6nVdh1Iqggu8O12E/mvRAt4lpfT6+vrx48dhGBBveu+tnWBx5nnm7No0Led8HEatdaNN27RSSiWkEjL6peFAF4kq1JNqQNd1HWPLbg7BEUI26xtUR5BKIy7JOQOUwOvAazocDpQSTIoGQav2NJzP12laii7YSI+Pj6ivYClgNNMy7jDPdkopUkZqmOmcO57elNGEkHG8ai0pxVB0GkPw4Rq9t1P21nJqxqtt29aO82q1Zpn42Ubk2Rray3GaRudcCH4cR63V6XS8Xi8kk+PrG+c8+tD3/eCDmy2nDFkgmHLAi+7u7uZ57rpIhZBoBpSSEpJiNEoTki+nM82EpBx9aLQJPAyXAUHD/f39VOaov76+Pjw8YPWmadpsNgiJEpQpl9fBBOdt24LRPjvXGLNZr/uuG4cheN+1LWcUvurDhw+c86r9RQvxBOEOZv2KMowBdTJo0tWGpFj4/jjpCAHRRioXmYBxve5hhaIPNJN5ntFFYbkIbhEJxhOFELBXVZkQE0vHQM65KrPGMtEW6TiyHyDb6/W6DgpDFLVarYBbmndzc1GJhNoCQHgkTKu2dc45awXn3jm12WDLIdirqDvMeiqaCAusFEKJeikwjxhTjInzJcHi7wbIizJMTBaxZzgbWRrdkVK3Ze4L3Hblzv257oGX3oLqBmCIa7RekYcKN9V/qCLTBuaJUtoYIwSf5wl8Mizlfr9/fXtrms4X5dfKTcQv4K0TQtAFTUphGZFyLt2wkIPHILZcSKj7/R6ib6fTCVGY9x5dtdvtFrgKrEnOWZWxEzHFirbDDw/DAKl61A+QOOd3Unq4CCG/tkphEfAWsc9gicCCAMCNBvKKLCFAxqMBh0Eg5orUYC7Nrog+bOmFbso8pQoW1e0CWBlCm9WvV6Y/WPM1JKkOPpXurVBayfDRtZ3ner22rfE+5Jy1UtjHKHVgNcy7uXWkDGjDEimloEyABA6PqZSQ6tfJHDUxUkph4hu4Qzkv1RewFbz3X758MUbHGICtqaKFN01TjItwJJSdeFHlk5LnnJxz0zTg4stxTb/KWMbCN79er02Gpp723gVvLcmZJOetVo0NkSnKKSMpj9chhah1M4+TURqQAsnZzTbGqBtTSxohhNPptEiu0uVogIMAT/z3f//30FvkRRwMCWjFS7Hn/a8jNBIh6e3tkFJqGt22JqUQo49FBccYA2QJzYygM+x2u9Pp9OHDB6zSzz//LPRSzxdCIE6+ns6c8+12Pc9z37eU5hj9drs2RoG6uNmscMSmCf0onBBaj0AsPQcxRufmlJLWcp7neR7b1jg3w2xO09R1aESd53l0rmkand51U82z/eMf/yikIIQcDs+ATxmjoC00jV6tOiGWnmHvfdPoEBxj9HI5oQPJOTvPk1LCufnLl1+0bnrazzallLbbLUq+0zRN88A488FO8wATRGgaxkuVuXwPk2w266Yx33///eFw6Pt2t9tRSinNlGZC0mrVDcPw8vJyd7dPKcyzA/2vNvTBVMbSfUXf1XSVUpSyEnkvdhh7OxWtdVaknPyfDhHABatd0mVMLN6++TPUvP/f7iGXLoyav6SiUAbrWeH1XIT56vatmTVsGYLfGNMwDAjwu6Z7jwvZ4K21zVLhWcZ98FLDwkmo53az2UAaAaEoCryEEFDTaFGWhrdHVyEro5VDmesCjhqmQ6/7VR2ikr2DxQTHVJSBOVClfnh4GMdRq8W8IqV47yCtta605rFSVsUgBARZWBxEx6CHwg+FMp1NSokiNiwUbiyU6TRa6+122zTNzz//jK0GYivyg1pUALESi0CXjgQuS09/dd6hSCjiPMMxEEKQGlYQD/UbHD/4jOv1SinJOaecV6uVEJJSKoVotbleryQmTqjighMaQkg5Q3r9crl8+PABHWqkAM2MMUozZUsxDQ4SPdt1r+P702RpGWAQS7/YPGelJGZElxQkXK/XGBfsDopeSARxitATjkY8XJBSynjOOcfo69uEz5ZaOOdoTikH533XtZzzvuusdUjflRKcM2MU5xQRAtwA+J141yG4pRZLUkrh5ub+9fV1mmbBqaCMZcIJzSEeD6+Xy2X/v+5+//vfV7YVQKdF/yMmsMy1kOM4Ki5yiEZK5+1wuSilaM5aymEYJOfoW4Wu1Pl8hhItXit0SioDGytMCjrBCvdBKXU8HrWWD7d3l8ulUZoQ0nZGMq6bhWxde0pyzm3beh94kW/B46NOjo+uKSwtk19TERNzRcUSUAGog+M4rtc953wYLtxxY/Q0DEoIEBy0lI3Wu81ms95oY/B+YXBwZUQngN2v1+tutytBZ0ZTOrIrzEMEboEDi22P84LqVyiaY7B4yISUknjw6/UKnUFjDMuEphxCSD7YcUo+cCHnaT4RAl8I4wA3XLbHwtqnpZczRsI4wt/IOSbfCIAQOJW0EFVI0ZFMZQYwmOtAcWgZ2pYKefp9jfrPcg+qaPfjM2ANZRkKVt0DFiiXIaOyUOxBM0UdVWslpZzm4e3t7fn5WSnVr3ouxOvrKxPcetd1veDL7IRUyPgVpq+VIlwW/okQMo7j119/jYjs/v5eSgkCMjKsCmcBlULAVWc1K6XQFeG9p4wZY7AvCV/8tiwi1cCF4CrwoXZ21loATSDUot4QiyK/KzPNaRmbE4rGPeinOA9QLKjutvLugdsiecLf5pwJAR8sEUpiDOO0CBHWAL9W4ZTS+EO081RiK25pLr3cUkqE4fgRovJQJsMgEKvnhFbZyBybpnl8PEqpYoyUUGOMEBI7HsUMnLRcutKEVph3XS04XvHt7d57N00TZQQ9pfg+XJoskv2Xy+Xl5YWWmhaWpfKy5nkMwSPzq7l/XCT7lxkPuCxQFO8doZnQPNtpmse6/oQkRmkMkTIafFC6MMeDY5TN0xSjF5znFGPId3d3v/zyiA4beLXVarPZrJ0LUhiEeGArAKvBVwWFS/qfnfd9z3POKJWBiY+oAihiLh1/eL/H4xE0OXRxY8Psbm5++unfYwzTFL58+dw0RgqB1WOMMra0rdW0G38I0VYc5BDCar0exiH45WQxSruui87vdruHj/fOOaG41IIQopU6nY9bvhWSd32Xcvrl888xxr7rU4rH4/F8vhyPRxT/kaYzxrx3q9VKCB5jQOYhpfDeouXFe5tS5Jz2fau1hKfp+/7777/nnO/3+3G8WmspIzEGSon3TmvddW2MO+fs+Xy8M/fYXJRmIRjn1Fq7WnVogNBaHw7aWnu9npumjZGAKoKkHNQ1SunxeBRCfPvtt2CgYamr1gPiPLCwOOerVT+OVynFZrOC+Y4x5hyl1oQmd5mFZB+/eojJW0dSDkDSQKqs+C15h+JiHwJnjjHJxeK7lHKMS30Rv4n9TwutY6HD6YYbkRPxPjrnGWWUMKWWBnVYQvIO4Plz3cPmZgeDmHzQWiux2H0fFyCoyusjLQJIggwopbTZbLquSykQGihrZzvO8/zTTz+nlJXSwjSc0DGmVmvddFxIxrhuWjVOIUD2wDZNQymbZ6vUInuJla2zHjnn0zRzzr0P6/Xmeh3aFtpE1FoLWH+e55peUJoJjVJS68au6wiNXBDOedObGOPhdMgss0yG4RJCWPedUPz1eGGEppSSdyRnGsP1dFytNkbJ4XIOzgpGBaM5BmXMGAK8Grg3uAF4CEopklPQhHwRJEBAgRpDLaeHEGC40RcmhJjmSXEavReMuGlYd7c0BRJz12wySc65EGLbSuRXjIkQFralEKLv+x9//BFlQ7iNChuCWyJLFwiyMZR2SwmLMSYYE4QQtCYwRqZpsKPtbrroAiHkej7d3Nw476QUk7Ncyes0XsYhURJJnr3rlAR4wt4RDcZxMIphu1PC7Dx1Rgc7t9oE50hK3rlGa875PI4JfZchMCHcHODgp+HCSBJChBwmN3HFffIkExcd4olwCSlFa13KIR39PM+ULW0NcCcV7R2GgcbgvV8XmfTgFGfMjmPsV7MbWU7z9dIIdr/diJz++ttvkk83N/vT8fzw4eMvnz8Lpbe73fHttW2b19dX04i206YRSnPnA8ncaD3P83A9M8ZWfZtTUFJsNreMibZtt9vtly9fcs5d16WctzcbSqmPLiSfaY45MEHnef7hpx+4WjrzoRZDBV1tNuHfs3PeGP32erjZrH/z9Sdj9G63/vz4yzfffH25XBPxq80NE0Qo5gIZ58H62XqVaeKShcm3nQGT6nQ83tzctF2Hj5CMHi/j/uYmZhYyo4RkJr88v/qUpTCU8Biyks1xOErhhQiMMDtOgrLhfOGEskyMVFrIHGl0yWVPIs0hX46DEJwTqbjKIbvJKaXGaVz36+1qywklJAtGc/BGKpZJ8smO1k5WCi24apv++elgdBN8JJkLoRgThDApdUokxhxjXq+3ISxZl9a6bS0h7PX1+Pb2+f7+w+vrKynKoYg/CCHPz89a68vlYosaPBRqp9E93H+MYcG7zufj/f2dblTMfaY8ZoK0H63CV+dgpljh0SBI6lYb1AIRdyKJd85B2DiEZO2iAdz3qG+lSMg8OympMS2jYhxHO3ttFpZjLQQSQkLMhIqQSds2k7XWeqlUSjn6kTHqg+PCcEFjDIzTSjf9s9zD/vZWIpaPqWkayRcNE0gos6JvlcucTuQNuOmUEiQ0SKHizvP8888/460IIQilm+02xrz8mHHTNE3TnBgzZYigLDOSpmlq2walbCBIiE/BZsPX7e3tMAwo+Q7DKMSCscAwgaERY+z7zloMKvAhYOpyTyh13jvvTRnqCz+H3Dn6gGKJKt+B9QFCghp1rcnXir0o09ZomT9xc3ODwXAxRgQgvLBLK98uxvj09FS7c/27KeQxJuRzSJaRvF+vFyaWfG4cxxCilGqa5pxJLAMwUAmUZUor5/x8Pq9WK2QzSOprwQ2JIJa967oQEm51HEf0LQPvjmVKMHRm4OqcX0Q+cK7AH8PeqAoogAVo0WKhlECaHwV89C7kzBDhoiYfi8oWSkTI93HPlFIhuLOuQpGA13LOMYRIgDItwBpyI/xh1eEBwSw4x3Ky4zQQijVhKWfO53k+nc6Ks//x7//2X//ln7Vg97e3Soqvvvrqjz9+0Up//fXX//A//c/TbFfrze3dfQyhbfU0jUrJtm3o0mwogfnW+lBcBhis//bv/s5Zj9oDktTvvvvOey+FwKQTYwyqu3hNYMqhxJqL+Aowh4I8zH3f39zctG0zpYQYBTV5rTXgVdgmVDXwKhchEClhKxHPol20aRofgvchhCglSTnDxnVdC7YY4FnO+eVylVKFsKCUCE1wEJxz1+vIGDscDqlMUsP+R7UJ93k8HpGtolyMjBMCJzAI1+tVCHm9Dufz+fPnzxUQ40XYwxhTlVFwY8Ca8KG5TGhHWq+UQlEqlTnP+/1+v9+jlQo8b8YY5+L19XW73aB45pzz3lq7jjExzoSQIUSjuXMuBWQ2C/9QFCmgaZoul0vMizgCLVwJQC/OLWSW9G5ms/fOmA7dmm3LAHOdz2fnHOMLbacWKnBXiWRCSEzhOlxTSoSRnDMlGTXCWiT2RYj+z3UP8H45ZyOVUsrbZdhIzR5yIaHWl4p1B0peKgeEMSKEQDkO8SnMCmFsmhZ5kxgSdgxhTBnDx1EIQTknjCVCmBCktOlBuVcWlnTTLHwvWB/sBkqX26s0YQi/cMVjjD7OIRHNpW66GKMybdN0aKPlXFLKhVApESEUDHHGzEhCALnM80yoBxsBbxp41/F4VGWUAu5wLkp2+HR8hxapbZwK1EWwbkCHxnGsE3FB+JmmaRqn1bpnUhFC7DjOPviUYyazD3wcKaHglWMToAdbyqXTrfK+sWKwjzc3N3hNFfqs47uraggMkFnGVDClMXPbxxBDSkwIyrnUOlMac445CykBuJkyvaDrOgAm/p3rRcoC91CrtefzuWJc3qeSYYwI4mxRAkazAikDXBljl8vZ+pmRbKfRTmPXdVqKeZ6jD75MkE0pdV3nnRvIwvYTdB2dFZRwkq2zwVnFmTGqbU3TGFA5+75XWnKlKSO/+93v/u2//6sQ4u7ugTOyWm8459vtdrvd3dzcNG3z3V99t93e/PjTD8a06/UWXk0ptdlsrLUPDw/ffffdTz/9hCMgpb69vVdKbzabebLr9foPf/gDkNhPnz4JzudpRm2ga9uf/v3HrutyTIILkHTv7+8h34T1QfQglZrGcbPefPjwYbvdCsHv93eM0OiDEnK9WueYUkrTMDoXlFKX07lr2h/++L0xRgkppbyeD6y0YVZZhJSSdy6nRHK+Xi7GGM6Y0DqnZK19enpCroytYq0VXEAhA049Ffo7amlIFmVRJD2dToz9KgSJ0wQuMtwDiAyYYoLtAQcAN48iIjoYnPeXywXuE/ycOr1DlCGMtDDLeZmglYo+K4T80MQAh1RLPkKIuSjfOe+u10tKcRzHw+FVKmO05ozZ2YYQ3g4HSokxDaIZIFTW2tvb2+12ezxfUcOrX7DXNdqrlVpaplOEAOkHtdvdfHl9fHl5AV0qly+sMCEkxhSnOYSgpfLWCSFyRI1hqQdXE12rAH+ue8Dpdc55bvF5SJQw7yEXAn4sE2lgUMCjSkViqG2R1MTX11f0HqMnBeNkoVsuhJqnRY8MRpMXmYdaORCMVjoNvlA9O59P3nsh+MPD/fff/5GQnHNCy1Utj3ddt1qtXl5ebLCxtFDgDcEEV11ctFijEr7ZbKyfQwiMUM551zX7/R716rZrQTJbr9dSyqUVrkjzV24JbDHOQ0Uw8VcwUtDRQ2iP7xSCP0F1hBZ9w6rgHWPE+M/qh6y1nPEYk5RUSom8ARwGBE3H4xG1aMCaeKfTNGHqMjAKvG4EeqlQXZ1zOS+lYEJSocRhLrGDwB+Sd7YMXnZ93ykl2tY8PjrOKbBgX5o2WRHCWq/XlLKcl3kGyALJr+PNLWMspWjtHIIvs1OWxKVOpIkxNE1j7XQdh8LhyaxwFsfLtZIOUbF4e3try8BkVlqxsG/brjVCxBD6rru7u0sxbjeb9XqtlbpO036//5u//u3/vlqt16vtfu/mkTKRc97tdowLzrkUEpor+5s9pQTld8xywDn/u7/7u0+fPmFvY8vtdrvVar3bbidtN5sN1vC3v/0tWvTf3t5AvgD+gIzZGLNe94eDjdFjGF9KquuacRyMMdSYcbgqrdCkdrmcGVsCDiklWqkx5o/SPI7Xt7fDdrv23islTqc3zjkhlJCktWSMcE5/+eWnaZo+ffr0+fPPKQW9jN3WjJGmMUJwtNpRSuECURr56sMHRNkAALHHnHNt21g7EbLYkOfn56+++irnqHWLF1d3Ail6ujWEQlkOvUGAfbDDgTCL0kYeQhiG4Xq9fvPNN6m0FJCCtmMPoMIMijamf5/P55eXl+++++5wOGC31NKgKUrJt7e3xpjnl2eloNFLYoyn40k33gevtTm8PAOSIiSvVmtb5kThKEECsg2LKQe6WwuHhCTGiFKC0iVRpnSpIw7DhMAoLhNiAue8trgCuUlLP0QWQgD4JSQBCEgpgtUmpYJAC6U05/iXGSb69PQkOI8xSsYppSRlLJkLi5ZGLQPUdiQk0ZUliXBAm+Z8Pj8+PnrvHx4ehBC16MrY0h6lNUHgcB0HIQSXgnNOOWOCt6pLMaaiKBeKgjc8BOfU2qlpmn/+5//rfD5SmtvWpLTEMjAQQBvGcVSNIYxTSimPTEghxDhbmnLwMfiolWGUa62VjIxyShjnIucspAQFqDZdE8phNOEDcPFxmpq2Qz4LyAJaF8BbEOnE0shW2ceqyIFAjBppEJaUlo5rFLWkVg5zVKSaZqtNQxlPKQbvqVy0TtvWOLd0/2r96yRhnDSMhq6Z7IJ0v9OpBWjzPhUt7okRkoZhWK/Xp9OblKpfryhn/XrFOJda+RgyJS4462ZjDLSYlJbTPArJx3Hhd93c3PzhD3+4vb0FQCR5B9gK5SsgCbDdxphhuIzjuNvtjscJ1fWmaeBv5nk0xhwOJ8YY4+x6PacQ5vEaY8zR57a11tp5nMarXK+H6zkU2pudiRRMCjYOl1qy67s2R7/p++PxrW2N1nK97pUSOUdj1HU4//3f/g9/8zd/81/+y//26dOnl+fny+USYlqtVl999ZFQ3vf9h48fnHUvLy9N0z5+efz666+ttW3Te++9ix8/fBJcPT2+MCqcDevVVmv98vLy6avtbrcX4lJZD9988w3UL7SUrTGtMcG5/+U//+dpmlIIj58/f/rm6+twmeZxd7NFNsY4fXk5CCEE1Cy46PtuyfuH2VubQrDTpJSy0+TmWQlhlDydTrc3uw/3dz/88INglKSYcp6sBw0dpHA0rk7TdH47bvrV8fBKKb1Qtt1uc4iBuGUGVFGfhXj+3X4vypRclP2enp6w56E0TAh5fHw8nU4PDw+EEEiepEKehgUHdAnZxC9fvszzbIwRQkBLDd+nlP7444+Msb7vvfPjPAG5hXNCvA8s5eeff64SVXd3d8655+fDapkAurTgMcbqCBMopP3yyy8fPnzY7XabzeY//aebGMPnz5+l5DHGNETxhQAAIABJREFU3/3ut+fz6fX11bQ2xRRj6Ferl5eX1mhKl6onLBVuD8dNl0FMMB206B2QInBS+8MLAnx6fHx2znXd6u3tNRaRlVCIfDX6jDEuA+mUSimlGNEZQ3JmjDIma1IC8IYXbdM/yz1cLpc674EQEn1A+Gm9q9ZflK5dhGyVHgPCJZ6TMYbYnBDSluHPQghSGgWrs4EIJR6g4iGEEMoWDdpa3sTcD0CcyF6///77Gl+klFFssEVqEXkiV8pjwKQLlDDBZU6EEmZMK+UkpSOENU2Xc54m+/JyUI3o+z6FGMuwI5wfbVr0ALMis0oIwTuoVDOEP0C9+aICtEwmSUWcHPdWcXkIdsKd5HcjVsLSpRUY40KozWantYkx9v36cj6RnJTSORPOhVLKOegKLDgbkGvQ+IDPwsPRIpSL3DmU0Rq1oQT/wNBjpRQhaRivYAGllAEPqjLUKL9r+6SlPQcfPQwD8OgKmBYZj/ntDVOpZW07R8MKTsvlckGraipjPKrdwf+iaVlrGZyf59nNNoQA1Yec83gdkg/Jh2CXOV+CshyiZAuOzwhJMXFCtZBqt7vZbFIIN9td17SjHnabLTDDl5fnb3/zm65p/vEf//F8Ph9Pp5jIf/1v/+18Pp/PlxATctO2bX0Im832cr6uVusYj9frCzzi3d09VFS8D23beR9Wq7XWxpbxVgDWvvnmG4xh3+/3+92NnWbBeLdabzdbwTh+B/xdKSVsK3KC0+lECW3b9nSE1Gs8nY7ee0bp5XQOzm9W6xCCm+1wuTZNkxlNId7f3m1Wa2hYnd6OUknCOOeUc9o0OqXQNPr19ZWQFLyVgp3sRAgRnB5zVEo1XYeqHjgUFUEahhGtMLDI2DzOuRDd+XJEMckHe/9wKxVHs14tGGAZbRkM7r3f7XZCCFQckWEgLZby1y4HxthsZ9ST0O8JwBbscGPM8XjMRQUOmw3J9NvbW875hx9+gKkBEpvKgJBvv/0WWg/b7TYG4r3b3ezmeWyadonblIoxKiGV0m3bMEqDs33feb/EIkDG8DWOo6RL8Qw3X+fBdV0DS7JagZefESHhnlE+RJxXel/a9KeiSVprdP+klCjlOUeYZ8ZE2zaELE1dOWcpOYiKfwH34JwjOTPGQLUm6VcJDYSW9cyrMoWRlA7AqqlnjHJ+fnx8hGWBWgOsYYqRc4wbY0imLtcrZTSRTBl1wWtidGPGcdRSuQIKQX0eWWfXNU2rn56elBbH0+vC/Ses79bOBeccNEeRveKGMyWUs9nZz49fvvvuOy6Fj6H2DcDbOedCCE1j8FwhLVpm4DXjI1KZu4kfHQ6HYRzvHxY15lxaLoEOQRYYvLRQlE/w+o/HY4VWamjAiwwG4AUppXXW+kgZbXLa7XYhBOudbowympSeOK0bQpY+TPjaDx8+AI1B/IKOuUoMh+NBrIQACj8lpZeKlW7qpmmG4YKpfFLKRHKIZJhG927arQuecea9zbkFBOF9UEooJd5ej5zLL1++4M9RP7DWpkAp/VW3EWWJlJLzs/MzYPrNdpVzjikyLp2f53m2bgohjBNxfnbOWcdTySOhcj9NQ86Z0GQaJSTzgexuNofDAW9QGxni0piiTWOMUVqs+v7rr75KOX74+MA5H6fh7v4W3eaPT19u93ul1Ndff/34+Jgz2d/dEUq3213Xdf/H//lfvQ8hpa5bpZS11n3f/8M//MPnz5+BbiulIBicUvry5Uvf9+fzGRvj5eVl//KCHl3vPeYVI66CXdBlMutut4sxzm+HFKLWWmtpjDLGzPN4uVycnQlhHH3OjFwul2NwhJCYGYgDoXRF4ZB+/PihafX5fJ7tONsRo0MJSbd3Dx8+fPj9739f+DP98Xj85ZdfgHkCpO37Hi0sTde9vL7BQ+Qif3J/f88p44wBtr2/v8eBYmUs2jzPt7e3gJcxcx7WH9fZ7XYYsPr6+np/f4+TUoWh8FcwHejVWK1W6By6jiN26cPDA3/XjV8xZNRv4e9TaVEWQmDIc+1FaNt2t9uhQr7ZbIQQSK8p5Yyxrz5+PLy+0ELtWwspjTTGjNOQUiKUrNbrtmlsEfjK7waKkNKXsMS7lMYylqpWenAPpeoQ4CaBpOENNs2SZqUysgImZbvdSs6T9zRlrZXqlt4AxljTGFK6UuK7/uq/gHsgOaNomXzgnIPfaa3FvIdYZA9gJX0ZHkBKASSEgAQCMQ4wuN12/3o839zcXK5jJhioQtbrtbN+GagQE2OZC5YIaYyZ5rnv+q5tD09PqZBEYeiFYG27Cgl7yHjnUspSqpyyUirGbK2F+T4ej2DLUMbsdSY5KyH3+/2q6w/PL7J0LbIilV7RFURGxjTo0wEJ4eHhAfyNmpwCGwUfCcGLKkPEOOdvb2/oX6vpYa3BEkKAO0HODxfJOV8uF4xxdUWzbBiGxFgKy8j1eZpwkRQzSVkIkTNGaUrETWAEwBbD0Mci3VW3V21pLuu5TNdgdVZHzuCxAM4+no5FTFdnEhhjc2nSSTFhVrDz4TqMSpkYc4h5HGfvo1Iq52WOgijyRJQyF5xREgQzOAnvnbW2pQ2eHabt5eUlpXR7e5uCp4zlGNw8CcYsYk/VESlTjJwx5xyjNKfEGOvalpWudXSoleArr/r/h7U367HlSq/E9hw7xjPldCfmJVlFWQLa3VbJQMsw7H9u2BDghiC01bLKUFNVHO+Q05lj3qMf1okoSo8S84nk5c3ME7H3N6xvrfUtyrJs28YYs1yuqqLKslQJiZUpwflFUUrGtVSb5Wr36rooMkAE1vs0K6zx6/XVMPbGmu12++WXX2V5MQwj4/z9+/eH/QHjJcRHBGUIUCDDRkrGABYoK7ZSz6SmtmviZC82m2shlLRD65y7v7/He8TMjFBKIkE7FYL79PB57AfO2GAc8FU37ZSmlN7e3uKEw4sMr/h0OhJCF+urvMgiCYlWRZm3XZPladvUy+XCOZvn2fX11Xa7XSwq7JJTSu52W6016tP1ep3n2fb5xVkHoRkAVYx5CQ1aJzEG7y+sJ0pJmqY60TiKWuvZ6wzm6s651WoFe1068USh4wE2ABfY7W43mvHm5gZWj3RS6aKSwy+gJjkxmIFlucDu4SzL3r179/T0NI6jVCrRGs88TLt9kFqCJ31/YdbBc5pzMYyDUMkFVE+ytjmnSQK1ylyQee8BqVlrhVLeO86FUlIpSUiUUoTgueCRBB9cJFA4OC5Y15mm6Y7H43K5YgwIDUJywGQU93TOvoSQSDxlREhGCBlHh6IqBB9CJIQQGkO8FI7x3zp7+BdOfmPXD23XNy1Ik93Qn5t6MCNccPtp9Xbf98DNkeJQ8kAhBceh8/m8fdkNxul84QiVOvdUbK7vSOTeESWTtrk8zbZtOWXEBzsYTlh0IVqvpUqk8iHKRPtIRKIIJzZY440nPskKKjXlykUWPRNUeRuddfAVEIKNY98057atGSPeuqooizRnhMIOKPogucCsD+ZZQjBKo1KCUuJdvLt9I7gihAFL1VpRGveH7WjaGH2SAA30lLKyWsyYPs4oahZGSN+2dhw5pXYcGSHOGBpjkWVpkvzud7+7v7+nnCWpJoLZ6OkkA/S/WF8shLD9qJiI1g9NZwcTjIvWZzolhAmhtMZ42Xdd45yJMRrjrPWnUx0jJYQppQlh4N2ifUbuwdlFCYPEMONjqDUm4QgLnlgTnI3ORkGEG1yZVVVW8cg54YnQlAgudT86KpRIUsIE4ZLJhHGJLGitXS6XOM1aJ5GywfvBGiZFlmtKAomOEh+965o6eic5G/uua2oz9HYcorVhHG3fK8ZY8KuySDiTlAbrcp0WabYoyqvV+os3b7NESy6UEIuyLPM8TZIiyxZluV4udZK9e3f/29/+2f/wzV+8vnubJ0X05HZ9s1ldhRCtdW3bWeuGYXTOcyXLRbG5XqtEyEQtqlVWVlIXxlMuE8Ll+69/c/v6zWp9nWX59dW1FHK9XqPcvrq6wig4yzLn7Ol0PJ2Ofd+t16u6PmNdetOeH58+ExqEZD7Y0fQhOs5ZICErMi55IEFp1XSNcUZK2Td9sIFFxiKLLkYXvfGmH8a+ExwrPZLHp5eH55dj09Zt++bdW8pZ23eH05EJLpR0wQ/eMZVkeb7b74sia8+nNBFff3X/26++ZpHuX3ZjN3DCBOWLoqrP9de/+TMuEh9I2/dpnlnvmKCM0TSRaSI5jcGZ6E2RJZLTcex3+5e8SO/fv0u0PJ72XNBzfTwcDsZYIeTLy/b+/n2aZl3XW2sHa3ozjs7qPJM64Uq2Q2+8gxAB9RNyIewDSIwkRqVUludUcCK4p9E6v1iswDwUQhVFlWWF1hnIVxBLos9QSt3f3ysuHj99JlgeQykXgnCWpJpLMVizPxzKqsLfIj4IxoaxU4kQkp/PZ2Ns349pmnMmKIuH455zFmMIgXgXgydt1+9PB0/C6ExkRCTy3NaEUy4YFzQS57zxwQrJsNdTSUUiDT4GHzkTlLD97hAjXS7X6/WVEBLDSJWISDwXLPzCoMlPNuMyUUma66ygXI7WD8ZFyo0LXTdiuwmJF6Ipota/LT38a+aSmDw+UdyRf2nMR6aFyTFGgPKAzjHxBzIzjuOnT5+MMctph4kPl4WgftrHm+f5crlEBY3GsKoqpVTf903TLBYLLNddLpdFUfjgJsWvgdzcOxdjPB1Pil/IAOM4Nm0zFaQWCsyiKAbjtP7THtCmaYDCo96PE61+XuGJ+QGMkkhZSClhxlC3jbW2LBZ937ftAFQKjwVvDtW31vrz58/BuYt+Ytq+wjmf/T/evXs3DAM2UqAtE4SB7Y7Hi1FNnudaZ3gvmOWSaU0eKHQYe2C456ddg5gN4OMnkxknOvrlcvn58+d5BI0iCz8RSMKs7EORQi52zRd9BhCwNE2hPJ8qXOpjwCWkk8E7TnCiFBamQ54CoTWlVClJYtBaR2/32x36MPxK+OVnule4bFT+E3qJehngtZ4Wc+L4wcfQe7der5He5jMsVXp7e3tzc7Pf7pxzqdbYedD0XVFVRVUNxoCzy6VcLJZKpc65vuu7rr+9vf2rv/qf/+Ef/rGsynFouq59+/Yto9QYa6b1tABJnp+fsUYQyEmWYZsFt9ZeXV1RSne73adPn8rFAq8SMOA4jsCXwDxmjIHwCsuveTIJ3NwY88MPPzDGnp6esCB9t9u9efMGXPjlcrnfH8wvbFFwZ1F41nUtGf3xxx+rMvfGCiE2m83Nzc1uv2eMvX79GngLsCk+uRqnqUY3ieePMhEz5AmBbAkhODOYkcyFPAw8YozQ0g7DAA8i07b4WcCO4E6B3rEsS5CRQgjAtaBqVkoJzq1ziOCMsXnPMfCMPM8hhPbeMsZmdym0YsYYWl5aZwQx/GJk8pdDLtFJ0vf9brcbhmF3PLx7927WNgMed8Fprfe73W67u1pvGGPH06nIc+MNozz4uDvuYyBSSjNaRnlZVF3sY4jBx+3LbhzMarWKgRDCjBmdc97DZ4GHQMqyTJJ0uVzCEW5OCZhc8sl5F9jjPPvBCafTZlY+7UcQQuAJodb8dUbT8yzhl6AVmqkZvcJvVpYl5AhwPZvnKn3fPz8/j+MoLiYhpus67wNjDHR4fBjAfHMUmCvNGbybNS9ZlrXdZZOolFLrpB+ddS5JEsqoSi5jK2Psfr+/u7sDkq6UAsKeFVWYlgOzyc8OvmwIqZzz4/EIwjX+LkKVlNJaG7UG3HSqz3gTdV17H+Gs2w/DbNuCgcQ8voaZRwihrmt00DN1YU7ASqlAYtd1qUzwt/D/oGhaLpePj88QUc+vA3weXN0kSSBcmMlRSAz39/eEkJkqJgQDDIhUN0sx2LScAycPkZ1OlAHgPPOfzodBTLshkQspZzT4WaAAOKvve+ovFR8eO7jqi8Wi6c/WWk4JY4xROv/oeZiPJzOjBMmEg1FKkeTcxQCZcc5Befzrv/7rsiz/7u/+jjHKOb+7u6vrGkMjxEeVZOv1+vb2drNa//3f//1ms0Ef7KLbbDa4b3iJeZ4nKkkSTSmXMtHa51lxE2mR54RRZ/vj8fTVV3dlVZ2O5/kwbzYbvC8oSzCMBT/nD3/4A57Ver0GvlQuFihNmqbx3oUQP3z4ALoaIWSxWPzt3/4tptDQGFJKz+fz58+fX15e8FoJIYBPu66DJXVVVWDuo74xxjRNfX9/D6Ny5AlQAPAAg/eLZVWW5UxuOR6PIH2CXd00DXCbGMmrV6/gOoUTAtQRNcdqtQohNE07Hx5jzDiOp9NJaw3nZjYJidUvDHuAlpxOJyg5AD+YYeQc1bqhk6rplyeZWTtO27mzPJdS4rq9evUKZWiSJKdTR6altkhLVVV9+vSJUYbT5SdJICIACi+MAXB0kbO3+93d3Z0x5sOHD3/+538O3rx3fhiG87lO09QHL4V0wT4/P+tMhxicc8bYcRxjvAg7Qgh9f/GzeXl5ATnFOZemBefC+4BMUFUVoi+wx/njl2UJSyXUu8BpECqNMc55rOvAxcfQHmUrvkmWaR8ALP8pev+70sNcNcyhKkwc5Dh9IWXBWGn+mkfHh8MB0KEUoqoWeCVFnhdFodQBkig1uWzSSUzIJ9HDjISwi8KWpmlKGTke9wiLx+OJMIHYIYQAjmmtdd4hdR2PR9CZMfxIdH7xGU2Su7s7ay2UmRix4gg+PDzg3DPGQJBlkxIYCQ/jIBB+6rpWSkM4TSYdL1QdhBA4BZlxjFNwxDQMv23btqvVChty8KP7cZi1crgDyLhgfM5VPOQ/+ETr9Rpi4BksQhKF11Dbtm/fvnWTwDJNU2tHBNx5yczMNVDT8lTcOpRsqM3R8RRFMQ/xMLVj04okZI5+HJi4uDcjHAPtpSFE49I0ozRaOzAWCfEhOODIkjMhBAkuz3M44qXTF5/cxHDD8blwMhErUUNg0w5i0Nu3bznnqNAZo+v1GtF/XrlaVtVltbJ1uHVpmjZt2w7N1fVVuahGa5CtlU5Ga3SSDb11LmRpwZhQidJaj6aPk9qOcy7Exao6xpjlad+3RZFTSj99+rhYVDc315xzmNujOZjrXKWkc/bqagMlCqWEUrLbbWdqVlmWID4AKJ/3giBqgweolCqK4vn5GYwXiNhRA2mdWGu9dzHGYeg5F+fz8Z+/+wPoUlLKLNNd3aRpGmJ82W63221VVYiMuMjAA/ColZLjOBoz9n1CKUPFM5ei+PhITohZ6EohJkDHD6EMQiT8o3RWoLjB7wxFwvX1dXQeCpVxHK+vr0HI9t63TZNlGWWsruvD6Qg1TJbmMcbn52eI79A6EEIgSALujVZYSqm1XhSLDx8+EM62261UCqcLEIW1Fhtf9vv9w8PD3fUN+tTvv/9+thlG388467ru6enpP/3H/0gj8d4zQvuuU0oba6z1MZIkSa211vqy1EmSVtUCZehisaqqZYyREEYpG0djrQ2BMMadC0KoNM0QVEFwAHPpUkSGoKfNXXND4Jxj7AIZ4UoiPpPJwst7Hy4qul8pPcwZAoPoGVCikzkaGlWERWx7p5NwF73ny8sLY4wzGQNx3qEPtSGK3Q6VBd7ZzJCz0zZzJCTYDABgwQ8CxYhN1pd938okc9a9vLx456wy2HLVdg0CE04wlHpd12V5lUxb1OHtDkwTggNwt5Gr8EqKqsSDxuvJ0hQFBePMGPP09BRCECLBSzLWjsYgkSDz43XGEMZhwFjSTPvsCCFaa5BVEGUopYRR7/2AhTD0IsHfbrdzHE/TFN9kLqJno9l59ggPDMRNoGf4u7vdbrlcSskx9IZMFO9uJp+g3ABQA0ZdnjNoOEC4fHx87LoONxzrQBDvoDhTUSmdeO+R3vDMlVK27613SfTO23N9Ntb0Q89mIoB3UkpGOPhR+BQw7OT84i8EnBPD7fV67ZwDNEEIIYQm0xLW9Xr9zTffoIIehh5eRh8+fGjbFkWuUur66grJdVFWs9hKSrnQyxCCTtOiLPIsZ5wRSna7XdsM//RP375588ViuSKUxBDrujZuZIzd33+ByGLMGELknCeJoiw2bY2e+O27N9jkKrhijH3zzTdwTMHrxuugNC4WZdu2dX0xkjmfj8Y45xzqgKqqHh8fUSUgW7NpUSCCKTgUGA6DGocIMgy9taNzjjGqtRrHAQN57/1+v6/yLISAKLNYLNar9TCGT58+IbKjNUSniHudZmlVLYwZm6Z2zuL6409xHs7ns/feWrNYLPHWOOfL5RLdJMBVXHb0WMAz+2nxH2MM7awxpqoq4sPpdDoej+v1uqoqzHvFvPzR2rZtf/rw8+bqqiiKvusfHh5CCK9fvwZChf5ptVogcKF4wqaW1WolmZxgBuMh2BQcTwOHeblcXm02r169IpPRwMPDw8ePH5GtL8hziPiJjDEP/ghWAyRKSIFDSCmVUhljORfjaLwPdV1nWR5CbJpGa50kGpM/XFghxOl0Ar6CY79arcDZgd3sPDUMk/EqagKl5JyncWFRg861YAguuMtehuni/PvSQ7jItS+8ST/tLKUTzx3/Cg4ZXipqUnQAh8PhdDrFGBkYb1ImKjGmOZ9rSil2tszfB/UmykMUVhDdxElOiVpyHEedJtaOeEAhxqZphnHw1q1XKzbp0du221ytcYVWq9UshD4cDlwINGUoRthEaEPOOBwOKMbTNF0ul8vl8nA4gJPK82x+sn+qi+H5Y0zXdcM4hhiLolitVpvNBgVgURRmGOhso6/Uer0+HA7zrfvxxx/3+z30pf04KKWcMPBOQJeNj08n9w6gLnPV0Pc9rJNQ0eBJjtMqCDqtafKT8yu+gPDgMsRp8ytKPNxeBFbIo7TWp9MJQyB8H2iPEbvVJO1GMgOlBKkR30opZZ3x3lNKGGO73ZZNyqPRmhgjpySEkGk1T27m0+8m/zL8qrD/hJAQjRRq2/V68+233zLGbm5ucHtvb2/3+x1mepzz9XqN7i3LsuVqBUiNhPj1118H72OMh92uWJVcMC64VOr+y/fe+8fHx+eX3W+//rOmaaxxQvDgo0713aubnz98GMcxy/LT6XR1db1cLs/n2nuvEhmCJyRyLpyzaarLsmSMeke89+DzABy/urqy1hIWheAxhr7vsiwFElXXZyEuwCYoc3OkAI6EWltMni5YKgc7EzJZ2T89PZVl3jT1YrG4ubmRUvRDnyTKOZdlQIrAdExhuv7dd3/seh9jhDkgTjjadywmqspquVw6Z7HdYRwNZmbI3JhvIRpKKR8eHgAqhGkNFBajZln24cMHQghK46IorI+oLcB/Wa1WwJcEZc45YGgIdoSQLE0bf1lhC8wkTTWllDEuphVPqKjwimGCDce96+treDdRSl24GGi/vLx8+dVX86pBFFu//e1vAW4fDocizYZxoJzP8wlCyITmE6WUFJJSWlWVNZZxnmVZnpfn81kIlaa595Fz7pw3xiaJ7roedqrWOs5FjNQ5HzyJgTIqhnHQSXZqTnlWcCZRDSDogdGLS4FEi+sG4HQcxzzP5tJ8psBh9oCajxBGGfHee/9vN23917OHMAm4UFDMIJ2bLKUQrTB3EhNDFPU+CqU0TRgnq/X6cDyO5uJCjoICHQaKX/cLUyp8KmRUdCrzlK/rOucvmCNqcMTEzWq9WCzscNmXoqY9B3NWQy61znV9D5gFjbn3HsljDmd8cmCHaQSESOi4sRGIEAIICFiT1qNzLSp0Ma04RWS/EEkJQUZhF5M7h6Pcdd3Ly8vTy4ufls/M7n7z55rxPXwrwHF4/WCF7fd7UI/AIsP3cZNUEO8LbRxKJ2stHJnu7u4AMRNCMNZGr4NaCS83TNoOOpGp0DogWIM4gMeI3KN0gryOKbSZPFfquhaC9T1EtrGuz5fBEhdJkthx4JwX2S0mjTNhAR0knyQa4KSjhUJ7VNd1VVV5nkPvWpYlhgcPDw95nqP2RYD7ZeWVap3npZTSW+e9x8yjPp2EFFKKNE2FlEVZwgH+5vr27esvnXPL5TpNxX7fRRbNVOgVRREjKcsyS4s0PWit81wfjrsQvBA8SdRiUSWJDCGkOpvD3MePHxErq6psunq1XgohhOQqkSpZKKW4YJvNpizL0+mEf7i5uYmT4SOdBo+Yt4cQUIu8evXq8+fPwFJQe3HOrTMqkeYw/PG7P5xOhzxPccvmE4V/1loPozmfz8baeXSE1uHm5gZF0qvXr4Zh0Drpus6YkRCKen++s7P42RiDtD1PtlC5w0sG3xb5LMuy/fGMtIF6yE1OzN66xWIBGu75fEYFDQSPMWaNsdZgwPOy3SKso5TBHTkcDgB4MfxArM/zvGma5XIZJt7q3d0dJuHoAq+vr733b968eXp6EpOZ/NPTk0jU7e1tVVWYR+JOueACCUqqNE2LouzaNp0wH/CYUdvNaEecli7gitFp/3F97oZhyPNM65RS5r0Hsg3kFtEPf4tN+7JwsHG2MSmB2z+dnI0wAuCTTM1aKwSbhmr010kPYVoiPaM9eP2RRZFwSpkPfnm1lEJywVWaSJ1EEgP1hIm2bR+fH611SaoTpTdX18YFIeQwGBqJw7ZurQHaPD8/4ySBKYGiGG1RmDYen89nyFWKstA6QYuglEoztdvu8H967wIhQrJqUcZI0KYliYKcWAhhnBecJ0oREpWUhERnLSVEcJ4WxdPzM3JvmI67tzZYRylx1tIYYoiJ1pQxpVTb9kJIxgznzDtCQnTGMM7NOHrvBOdCcMEZpRfxASEkzdI3b98wzouyQJ1orD2dz0opyljXdSREQZljDL+kszaGQBnru+7z58/jOGidhOCl1Cj9OGcX+5fTCamlKHPvLaEg+URr7Q8/fI/VEXMWORwOq9Xq+++/R+JnjC0Wi0guVo6MAe6P3gcpE2vdOJ6bpvXeSzmcz2csk0ALMgy9lHIYOmMG52wk3pp0HDtKqffW2kFrhbUQIbjz+Yz7j+li3/dpllNC2qYRQrRdv16tmJDMX2gYlLIQrPdyhYkyAAAgAElEQVQ+y8qqSt69eVsWJVK7UlKppGmaoR9vb+9QqL5+/Urr5Icffvj06eNf/uVfdl1unTscj5EQrXU/DMMwpFmWF0WikqZpSCTVotrvD4wxrmRRlqMZAyF5URRlmSRJ07Y5l6/fvu26Niv1uTZd3wymd3a4ulol6hU2DuVZjrRkrdWpdttwPjd5lhvjs6yglJNIfLBSSO8d54wQgvKcECqEWC83xtqhN0M/rlarNE21yt6/fw/sBQy3uSSsqurm5gYEClQV8GcFGQz8NCQSxhjnghKep6VW559++MCI0CqzzuY6UYyPY28BIklZlAuV2A8fnxnnuJuUEA7+cZIYY9q26bvup59+eP/+fhgGJSWljPGY6MQHvz/sI4mL5SLLMn7mlJDGmOvr6/3hQAjJ0rTr+v1ul2ptJrvQNE0TpaQQjJAs1VmaUkqUksMwCM4F55Rd1PiY1UHrQCeR6WDGtu+zPCeUnk4nwQWJBDuOMPnouhaN136/c84ul4vT6bhcLkLwnHMzmNV6farP6H2Hvt++vDx8fnj/7ossTxHI6/OZM6a1rqpK6gQ9McwQUedRS0Y7mnE8n0+M0vPpbL2T6rIHCT5m2NWKQgoalNnuAaA35zxE770bx9F5xznlnJVlwTlv6toYEyY2B+Mc9TTqP5RfGDBsNpuiyCklxqBw4ZQyRGt0byh5CSWUMELJv3ld3L9ID4jU+JqnGYxRRxyAAslEtSitsdYalSwYZ5xTlQgfx/1pezgdNuurolponfajsy4I68fRMEopuZiFIetA5h5CQBsVJqsMFLDjOKJ0RWNICLEmyfN8vbqijEZCGOOcCR+8sSNntFqskiTZ7g7ncxNjvLq6XiyWwzBSShPENUqi96fjoW/bPM9Xy8U4DFIIrZS1VnLeORecU0KYYTDjIBgLIeg0DyEYQzgTgXAuFeeCUkYjoSFYa7M0VVpnaeqtHYcuT5OrzSpJEmPGgpYyUZKoNMtGY9Is2+53UijUaBBGRh84ZYN1eZrqJBmU6rsuhqCkbIfBW5tlaZrq04kwdjFw5pxxzqZamyaJ2GxWfd8ZM9Z164N3zvz00498cpWxVhFCpJTv3r1DM8Sm7VGMXVhqQkjOpbWhrrsQgpSJEMIYG0LY7w9SJnXdzJ4cfT+UZcEFIzSMpicsGpMxGkPwghMz9g0JV1dXxgxwvKGT3h5NDCHEmhFN0rludJoZF2yIUrAsy7TOuo532N5MxVdf/dki1//jf/gPTXt6fHy01gTnx9GtlusQXVleFpWjQn95eXl8fsbwgwtRLZd2t0vzPM2yr7782nv/6dOnh4eH+/v7sNvZ4NdXm7fv3r28PEP2jLPnvd9sroSSSUy73u4P275vx7GPJFhjy3zVmjYr86IoZ/mnNY4zSQnf7Q7L5ZJEZo2v67qsckpjP3TA2bru4jNPSFxWMdN5mmR9O2o1Dp0piwotGno7lJzoZtAlvLy8gKWGMf7MX8A9BY+TUmqN65q+a/tE6bvbV8YYnaR5VhAaLAkT05r5SEKkIZJxGJIkMSEIDGOqSghRlaW1rizyREEnFDOd4VaWy5JyyhnXmS4X5eZ6wxhTh+P5eI4hxhBJiEPXe+/LorDWHPZ751xzPhd3d4IxKUSqdQguUVIKxlLdtm1wtqlP9fmY6hzybDCLoGqMMYYYfQyEsTTLhBSjGeu2ac51plOtk3HsGSNpmlxfX/V9v92+GGOUkt77uj7vdrs8z7uu/ennj7/5zW8en59ubm4WVZWmaXBeSam4WORFjJET+vj8Yozpmvbbb7/9X//3/w2tAHZ5TeprD1XSmCQtb47Hfd8Pb968cb9wjgGlEwC19/6nn356//59lmUzgiql9N7qVDlnm6YWgnHBIgnDaMaur+v63Dbr9dpYy4UIJLrgE56geqaUptMi4WlXZzTGeC84D13XA1CdxsaXDXox/ErLRGcsmP7CPSpiQQNlnDFKmR0temcpVJZmWish6X6/3b1srTGLxWKeto/Tfju0WmQaceO/A1GBkg5N08zYxfC2nxYd46xYa0EM3x32E2nSG86dNcYYQjlwf+R5DDbwPfHNQfwAMwQANxI+5xzzT3rZ3OAwssYrn5n12C7HsSZo2u4gk4QLzhgdhuH5+TkEfz7XQvTexzzLUW4rlVzKVXbBl9I0hdEIykNoQTEfRiWIJv39+/d/+O47Ntmk46VMj8ij7xaSa63r+jwRii5qDDOZsGLmoZRarVazObBzLssyH1y44KTOOUfpn9xThmGA2z5+k67rVqs1Srk01ZgZXtC5vu/ajlHmg3f2wrUYx9FZK4XgjDlY4VNKI/HWmXG0xlhrgw9Fnnd5AUE5gG801IvF4ne/+93L8+7V3Q0jrlqU1SK1rn9+enHetF17PO5DDHhH1trtdovayln7/LwVUmzW6/V6U9eNEGNRlHgRVVVhofdms/nw4cN6vVmv109PT23bxUjyvNA6TRK9WCy1Tvq+G81wPp9j9IxxKeSiWkqhmrbmgus04ZwKySkjbdMWRc4Y/fz5c12f//kP3yqVxBDeffGac/74+IwTiM3PWZaF4J1z8BrCqL/rOoRFcGlmIcXMUrXWGmNub2/btt3v98Bg51eJ4w1WK2c0y1IhhdaJTrVScjSjkpLxP/kiI/4CDQZPGhN+NHAzyQpKAhyhx8fHqqoIiYuqopTGSCghnDIB2nG4gFd4iTM705gRlwuXd2ZC+mkfNaJbnud1Xfd9n+oc0NbPP/+M8WHTNMBFCaVSydVymeaZMTZNUxoJjSTGeDgczufzYrF49eoVY+x0Ok8W+jLLihDip0+fy7La7/dSyvV6DUASj/ebb74hjI7WVGW1ubr66cPPfd+VZcUEf/36NWyPUdOEy1piWxR5VVVpmnEu9vujUsoYezrt4i+stgn505bc6+trUBBn64d5Sgw+JO4+ELau6/qh//TpE9CwYRh836U6nSMzAhdGRGpaRAgAHAHNe0/pxdJ1/h/+zbnhX6cHvEKEJDt5shJCYojWWSKlkiKGmMikyIvlYpmmmVKiH5rt80vbNDEEb10IdObwALG5RCsh+WQKD1AS5wPulb9Ew4dp0Rgw6JlrMROTCaU+eEZJ33UUGzbsZYI6j+DstEs5yzK4myH3gvUxjiPUvJTS8/mM4cHxeIwxyMlcD+NQQDR5elkETymVSl1cHIqiGzqsmmiallJ6OBwpZdViKYSklCmlhmHc7w9CiCzL27ZBvnSXXaEErCd0x8BzUVmvViv41Mdf2B3iBFhrQ+i01lJe3I+NMUWR0wsL/yKNwT9jOHF1dfWvvAoA3yFuJglchS/OkdZiuxFULJ2UEqscJ+IAGcbLvAS/7WG/V0qRGM04JlIZY16eno/7Q1VVnLLRD4wxbx2jVArBCHXWccqkEtEHRkiR5T2jSl322Vlrb2+vrq+v63MrhDRDu1qV53p/Oh10qrAQ5OnpMdH6N7/5DQjjs6ZP64wQqpP0P//n/+XHH3/0Lgguq3KBUT8shDHURRLSOl0uVzhv+/2hLKs8Lwghh+O+aZtIsj/84Q9FkSVJ8vPPH377229iHJ2zhARKSYjBe2etiTFIpbI8FRImwe1+v8NP2e12j4/PaZo2dffw8AgcCfPY9XoNkQRqZMDWMJiLMUJkiu4EFxM2O9DrgJiLDwWWHXLGZrNJszSSMI69Mebm5gqZI03Tl5cXyLn1tIYIgQPLCvm0qACZSQgxDCdC4n6/D8GfTqe2bV69euW81eqijnx2nhISnKeUOn8hxXZdSwhRSta1Z4zf3d1hOcRMugUm/vT0dH9/D6rIYrEA3W7msuNeJEny5s0bMO4eHh7KqhJKGigbzuc0TXOdtk2LiSNyJEKK1ulut6eUEkLTNCuK8nyuCaHX19eU0qurq6nc9uAKW+eE9z4GmShgjEKINMtQSjLGkLBh2J6mKeeiqhZJoo0xnIs0zcbRNE2LwQ+d1qDhTQEbvEiDvX9+fn737h0+I5BhBBk5Wb8458ZhrOv6H3//+/v7+xBCkmrOL2aFYZLCIQ4cj0c7bfZGfYDcL+VlVy6mevOV/xXSg5tcPMm0zjtcnJ5oCIFJJriUXBZ5kWf5slpSSiIJzbnebbdd2zlru7aVSco4h8HLMAzL5RLjKR8JchrOwdxkIMsBFp8J1OihMBfCmBqtAKJhkiRmNDpVi+VSSblcLg/HM6WXpgSRNIQA1A/ysV/S3Xa7XddNOzv7HvGUELLb7UAYB6UBt4VNij9rLScU1Udb15TSV29ed2Pb933fD4fDXko1DKO1Tqe59+F0Oi8Wi74f+n7AeWWMlWUJh2E1rS2kk0INkBrG4xd9PyGcc3AiMaRBpIB2h3O2WFRlWbZtlSRKSJFnBSbYZlrBFEKANQ3IJLPpcdM0WZ723YC6I8bonDcGRqfw21C48JRKzlnbNmVZeO+9d2gdLmPzvo9ac8oopU1dU0JpJCREQgij1Fnbdh2qp0Qp4PUdhD8oR5zjnGc6ZYwgTWKz0M3NzT9/+8fDYd81u/dfvqWUgsRQlnma5k/PL0LK+/t7GPdD/NV13TDaJNHjaBaL5fF4GoYxhCClOh6PAHBhVghb2ffv30Oa1PfD4XDY7w9pmimlOBfjuGOMff788b/8l//7zZs3q9XqdKo5F4fD4XDYM0bzPFuv16fT6fHxQUqZJGtCZvFRfHl5zrL0w4cP3333HSFss9kgCiAe5XmGK4AKGiVRnFZ60cn9DSJq9Jd4R5vN5uPHj7gaODzb7RZnmxACyVsI/nQ6pqleLpdpmhyPBynx+tq7uzuUfSg7Xl5eEJ6w8gRi3e12ixBj7UgZicQnGkG8TLQcTh1isRCibRoKGCAEaw2hgXE+DB1jXKlESB4m/4U4ua7izAAjXSwWaDIQTAFCQo+GaZmbfAegvwENYaqRjdbajgYgwWazweNK0/TTp0+3t6+m0VcUgqdptlgsGeOzPIJNFHnE5WEcF8tljLFp28Vyud/vuRBf/+Y3nHOMDSDXwJaX29vbLMuEkITQ87nmXIQQMbpjjMECkkw7iPCmhmEA7xYDVziQQ/mP1zGOI8RG+BTDOGITcFmWZVVxKYwxjFyEhDOBbRgGqCyRIMFoqqoK7w4RFY8UDQdCx783PfyyE6G/cNQILiRar1frPC844ySSROnkgiD1h/3hfD5bMyZSCS5SrUdjEMPitKiIc346nuCgi3uC+EXn9QJTFAY8gtkU8gdiPZn2e4zjmKZp13ec0yxN8zxrmmbo+/k3b5pmt9t1XXd9ff3Lqhw1NcS9xth8El7ODXVd1+M44LdC7kVJi7OeKKVVQglVSg2cO+ekErlIrTVaJyOxhFClEin0crnO8uxwOAohk0RjXuScp/RiygQfb3CrcBmw/hPNlrX2+voaT8BaC9tLHDJKqZSKkBFPLNGJuJh1k+AvK/zSaYsqMiVE1DHGoihmJzj8KaY+9MIc7SdM097ff7FYLF9eXsZxiDFaa8ZxADVrGPq+74QQ3jvOOaMsOD+CQDWMqLwE52mSMEIE5yQEGqOWSgjukyTGqFXCGWOURefHfhCUCp1GekEnUFQio+z3+0RGwZVzZyll1/VJklLq26YbBnN7e/vx40eEXaiiQgiEUMY4IdRaxxgPITrn0deDbQWc7cKlsfZ8rjGP+fHHH1+9er1aqdPp3Lbn169ff/fdH0PwRVG+vGxDIL///f+3273kea51Utdn791+v//555+qqoKBHVQXIJLc3t5gvY8x7nw+6+TSHGitnbPobpMk2W63bBKg9X2Plg6tAzIHADS8F1SCwCexmOF0OmG7FNriEMJuv39+fmKMpqlOEmXMKKWglAC+AAKJ+NI0zdPT0/XV7dPTE/679365XKK0yvI0xjiOuCmjkCJNdQg+Vcm2ros0U0KeT6f/9vKSZRkhFClzGAbGqQ+OMZqmF42FUgq9EcgazrkkTRFGcYZnHBvKJEwcoS1ACQ/dQ4xRTNL6ROugEkYZ+NZAycK0WMIY1zTNbrvbXG36bkgSTSlDDMFXnOTHwJZRfsHrm087Sr/77juEeM75ZrOBmlIIKaUSXCqplEyscVKqEC7bL4Zh+Pz580y/lJNTcpqmHz582O/319fXuJhYVIVFOKBLIQymaQoTEfSF1pi+78kE383pNk6qLORyQggKICxi6rpmdiuYrdd/nfTApl0ZfPL0R3AXTKzX69Vq5ZxnlCupF9UyUZrSeDrtnx4f26bhXJDILVZFTsjJrOHC8JlO8jq0qKhzkQZmSIQQIqVcLBaopxAs0BFzznWqrXfOORKiUioSgvZTyqSsljg9WmuQU9EMmWkhBshncxtrrYWvMiwcwi+W5cZpiQeZSGaUUsa5c867yx5eSqn3zgerteJcdO0wjr13Qessy3Kl1Diaw+GIEhWmp8PQol8Baif+tD3qskACsxOQwcG0A4xY1zXAAe+9lBc28DiOzjoIFKQUQorZkQafSE7qd+fcH//4xzgJZ/BkhBBUMPxiGL557601McavvvpqHMf9fscY9T7AbBJp+5dDKcZYkijsWoBLKOdccE4ISVSSpdlms4k+qOlLcDEMQ55mZOLIxQtHXhtvEMEJIaiG8jz//rs/aMmEkG17/vDhR5WkwRNGpRQyK/LNZnM+n0+nEywcyrIkVDZNlyQJFElaa+fC4+MTuiUcbKhGyrJs2/bj5wdj7OfPj0mSnE+1td6Mtm7Om00lpTgcdre3t6vV6tv//s9ap23TMB6E4HV93u93Hz78jEqw7ztCAiGRc5bnWdPUjFHO2cePH0+nk9bZ8Xhcr8TNzc2XX36Z57kxI9AGeAcAFjsejzpLUc9iCd3z8zO6IgAI2DU7i6jHcYRWnE7LOGFZRqJdLhdZpmMMXdcYM3JOF4sKvEnUqjj5KJ5urm8uMkZr+eRHMo6jMUOSaKWkEKm1pihyrZV3lhJS5oWSsiwKZ+ynDx/v7u7KRemcCcEzRpNEWWsJCXmexcAQgnHCgQQgHPvJzwZhNE3Tw+EwDAO6W3yBL4BqiTKmdEIZHY1ZrzcqUcQHayxUQcfjcbvdfvvtt0mSBM+CD/vdvuv6b9ZXdd2sVqtxMEop1AGHwwEFO+R766srIYW1LlLCucjynBCipPzxhx8w3YG1xhdffPHVV191Xc8oh9OQEJJzEQPp2g7xDZz4f/qnf2KM3d3dHQ6Hr7/+uqqqeYsJLvt2u02nnaP4/cGxzPP8bI6vXr36u//nv061DkE4CtMSnTmtQoeL9OCm9clISG1bY15FCMFsA1Xjr5Ae8BUnge4MWsVIpFRZliP6oBzgnB+P+4eHR6ANJF5wN0+ZnPIBIWQGIpD95iJISgm+Ns49HhkwkNkdN0kSCKO89yjHFouF0klTN0Weee+zPIvBa60J5QjiMBjAz8LUAX1AnFwuEHBBDiOEfPz4EYUVEAA/bdVGaomT8TWUq2PXQx3WdR0UkkJwpdQ4uBgjZwL0svP5HEmE0tJ7j79yc3PTNGfOL74080WFDpZOiw/xo/u+RxWJXwM5bxiG7XYLcAl47n6/Z5wOw4AB8Cw0BaMcmBKe3n6/nz1L8IkYYyQCgILHOKnrmnNGCLu+uXp52VZV+fLyHIL3wcGivG3buq6NuThwKCWVkF3TIPGgAyMXiT/7pU8GynZ0LXDmwMOBTjtJVDu0jNGuuyz4/fTpk5TS+VgP7T/8t3+Uip3Pzbt3q+32OA41NvdhxIp5Y5ZlV1dXhErOz5RS53xdN6gZm6blfPv4+AgB3TiOGMPWdf3x4yettRlN3/XWOu/86XTu+ub1m2shRdM0i8VKKRUj6fvBOnN1VVlruq49Hg/ff//98Xh8+/bt6XQsy0JKKQQXgnvvpRRSCndZP0BRXsCqCCNlvNN55IaKdT6iCPfIeWgxAT1BATdT4BEXCCHgLAEhJIQsFtVqtQIRHIQGQiLgSqXUzc2NmXbWns/nIq82mw16Zag767oGSToETynJstT7qixL3AWgT0DnMEG8vr4mnBBKOGeE8BB817XDMHDOrI+g5GLKKifPOCii2bQjC5kSmQNwaFVVKJBxQowxPniZJM47H0IcBkJI/AXUgcnc6XTKsmy1vE4S/fT0/Pz8o1IJeNuEkHEcf/rpJ9wRlDKr1SoSYsx4Op+RIwkhYz1Cvf/NN99AMNH3/atXr6CmFlzEeNnnSClLEg12Z5pm8CpHow/0Aoo8MFmvrq6Ox+PhcAA6h14E+c8Yg8IX9poQUV+A0L6TiSI6FYzhqLA/0Q4vpkdyMujEoZpX46CwwGOHNc6vkB7ctDpjnhvjVwFx9lIeXkQZxHn3/PL4/PxorSdEMK6k0jzRQiqhkmE0idYuBJkkiU77ccQHm2W6M+IPtisSKXIGrhb8IVD1z3px731VVvX5nKYZF7KoFkopH2nTdv1ovHfL5ZJxQRknjCU67Z+eAHeSyUzwwggk5PbVK87Yw9OT1hrs+I8//xyDl5IbY4J3hBIpWAxuHG30Xgs1hPbq6iqGaJxjQqRpboM1xo529IQwyXWeCC6kkvO0A6kOzfv6apOl6fF07IbRGcM5Jz4mMlEy8S4MxqokiITfvHr98LKt63OW5SCeOheenl6E4FmWQ0MmpRzHfrvdLRYViUzKhBFhxhGvr1osqkUFJyjO5WDOzod+GL2P49gnQgkpqnIphFRKYQmPMUYIlmWFMUNVLvpu+NkYQigh1BpHKfXRdUPXto11xo7DOAyS5ZxLTphxAw3BjoNgKQ1RcpnmeblcMMF1loYQCGWBRC5FnqdZpkMIQmR5kRd5XpRlCEHUTHDunSPBF3kSvctSSYjdXK2fnh+ZSItycTo3kUSVyFTrujkH77//7vv79++XyzVljHG5ubkZhvF8Ov9f/8f/6Ua7fX5++8W7dui6oY+M3n/55Xb7cjyf3rx7xxjrjwfct5nS8+OPP1JK3757m6Xl0Jvjsfnqq9+O4whHZUKCMY4QFgIZR2uM9z5mWbnb7dq2j7FnjB2P9enUtO3QNH1dt5RKpfg4jqv14vb2VmtBqBM8khiVpG1vlFaUUh9ikmo77Y1Biwa+GedcFwXh3HgfKOVKqTTdHY93d7dcckJJqpObmytKabDueDxlmVwtFqvlIlFJ1w1lVmZ51pw6QinxxCb29vo2hNCcm+DD6XjarK/h9Ya0jduXpldXVzfWmiTRWutxNFpn1lrK+Ohdkeq+aynnZhw3m816vT6cjlwoIVQkViUJbfuirIqyIjFkWU4piZF9/vwpRkIJI4RkWscQGCHe2peXF85533WcUpiDpVmaaDma/nQ6UEbzPKcsCM6GsUNbxhg9n89aaWNckviu65fLtTHOmPF0Ol9d3Uqhsix7//6eUsIYbZqakMgZyXT6/PyspIwhNOczpYxSGj1pT00/DsFHa219bqRUaZqvViVMFZMkgWet994Hl2Z5Uzen+sQoG81Q5nme5YEGFEBsknOiyTuf6uVivVqts7SIgXoX23bI0gJEu9evX6OzB7TY9/3xfB7GYbFYcM6Xi8X+cAjO0xixVAblFxBIFNx4ZUCnnXMwMK+KsqhK633btoS2PlAfqfO/hixuHv8iMaDKMMZwKuaSH/2OtXa7bc7nkzHG+8C5SvOSc1mUpc4y6/wwjtibwxhniiOHA16fhwFicjxGNY11IqfTSUr5/PzMJ5E26np0Bq9fv16tVm3TYFiUJGnf93XTQVE8DENVLXwIUqksy6rFAhUrZhgoN8BBjud6sVjsdrskSSA/qapKpYnrhzRNx3GMMVBCZ4w+uji4oWv73XbvnGWMN0379LwljBpjvXd5VnDO27ajjDpnGGNgH6ElRN9HKOVKJomWUp6Pp6qqfLAxxL4fnPNKJSpJrHM+BOMsl5K4oFQCd1Lv/ePj42azsZPNSZblfU+1TruujSFyzjhjOklUkugsFUKoJKmWi7puXQiUsTTLzqczpQwVjVIJpqbb7RYfM0mSLMswIBnHUUmFgZ5zjgvR9z32gB5Pe05IDCFRKs8KTljTBCm51lIIPo5jkug0TbGtW02m4lIpY00IoaxyQkjXdZTGq+sNfK3l84X1gdmAc2OeayFYoi9EBlgvWNtxzkFpBNysdToaK2WS5mWWZSF6SsjjwyfiXbUok0QYz421Qojbu9vtbhsJiSS2XQs+MTABQA1gRpBIrL3oCRaL5fl8RpQJIRRFRQgRQjEmqmqBtyalIoRpnWit27btul4ImWV5lhapzqXiQojNZrVaLZQSlClGCOcihiDExZ0lMs6VpOHSp9LJpzLPcyp4lmWJ1hicCikpY2PfJzrRqU4SlWUZZ4xTVp/PQ9LnWZqmyXK5TNNMyZYSCvvItu0JIZzxsiifn5+HfnDOSSFhlIT48u7dOyBdQojlqprbQXgEhRCkUoM10tosz2MWtU5urq5Xq1XTdZEQylhZlEVRFHkhlSqLgjGSZdr7iNBGKQk+GmOAGRQTXcJ737VtURRCCnRRF9CVIUdK52KIMYw+XjyDpRCOEvxpMgyjcw4olnOeMco4ubndvP/ynVJqf9iCyK6kYpS1TQMv+OBDCK5pmjev3zJCE5lw6r79798eDgfn/W63/6v/6T9BnzTRZIUQYjQm0th0NaMsTdNh6ASjVVk4FzjnTdOgk4MFL6W0KpeUst1uC8wQaWaxWETil8sl8CX6ixW/jDOMwQEV3t3egtEEoPhPMwlr3WSpgGVTajIP9d4nIsGStDSN3r9wLoy9YOb/3vQwj5FnXAXJME2yKaYoVPqU0vPpPAwDYzwSkmVpVZY+kCzP87I8nc50cgIghIRpdTD+rp2+5h8xERM1+gmcSxDypJQwbJBS5nl+e3s7Q7Tb7XbuKgghaKM2mw2QHFSFSEt4vsBe8SaqqgKyoZTCVgnOueBSFzJJVIynNM37vk9USgjhTAImzvMc0zMgAPvjwcd4f38PehmSOSEEghQ52d7leX59fV2WZd2ez+dzWzsB9sIAACAASURBVDcAB9frtR0uW8hBvUjTtOlaKeVgRjw3bGHik48jctXpVBNC2EUOcqFalUWRpqkPgVCapVmIgUz66nnYI6Ws69pnvigKOxmBqMl5aYIN+el0Qs4gk+O6FmK73YLbl6ap6XsphVIqS7PoI+eU0KD7xIw2xkApo4xKIXBqZ1Z+lmbBjO/evRNCPDw8wEU1y7L1er1YLKC0APsCGfr169fLqooxHg4HFAcIwZTSLM2MMWVVlmVZMd73Y1GUlLKyLGmIm/XKmPHNF+8Wq2Xk/OnpGfSe0+kEY+2PHz92Xff+/qvD4XB9fQ0BVNu2SqnD4fDx48eHhwet9eFwgN577uv9tP4P10EIcZE9aw1IExNmMGivrq4i8YSQqipWqxWl8XQ6KSFijM6GLMmG0eLdWWtZJGVZSikPhwNuCprLw26/WCwWZVWVpRlNkReC8yIv0jRdLBeSC0bp2PUY12W5AqAEVCFNU+zsw/lXKmGMCsGzTIMye/fqlnNm7UgZH8a+bpiQbBg7QiqMx2dLlXEcKWPW2oeHz4Lx1XJZFiWczGMMjLHgfZqm4NfhaSSJ1FqMo4fdFrBqxDgUy7M3D5wu+eQMDb00/pRz5pzRWvtIQojOB+s8F5wRxgiDfgikA1STsMhFpt9ut9vtFoBeD/2HFC4GwmjTdevN2jgHtvditYIhB4QX3vvFYnE4HABpIPTjpVPG+r4nkUwULzsMg48x6iiEeP/+/U8//YSQhY+glDQmonjFa/Xep1lyd3cXJj8SxCW0jwCiYS/tp/2AYNOgbgBcjIxuJpdfRO95zi+lTNIUsfHq6so4++uAS3Ry8CDT2h+l1Hq9LvNKa43whzDddd35fAIDUmstpaKUScWzNGX04v9MCLHTCgc6uUdNJUkC+jnSKRykvfewHpz5vHgcuJ9IA7BFBHIHUy08VnxnvOOZkFrX9fl8gqro9evXxpibm5svvvjidDoNxr558wa8wHEcy7KEOmnsO8q5VIpQGrqOcgaOOReqH+qirPp+MNb7EJeL5WiN9X6zuXLOEcI2m+txtLvd9ng8YjQ6E7Hi5AV9qs9IV1988UVVVY/1Z3xkQggYFGLa9XF1ddXXLfz9Z3R40ssYyBGUUlVVWpt0faeShBACe1opJaHEOBtjBLtxHAetlfe2rk/LqsiLlE3hG/UIjimu7t/8zd/AyhhnkU2+e1mWJVLN04u2batiAfj4L/78LyIJ//j//r7rhqqqxrEXlPVtxwmNkeRpprgglGR5jpEPFmxsNhuU7ZApYZCAADEMw83NjVYKdTSKfYxqiqIgIWy329evXjPGbu/uvv/+h6Ioura1xiglvvzyi3EY8qJYLJdltdxudzCfSJLk3bt3GGhvNpu7uzu0IDiQYTK3Gcfx97//vXPu6ekJaQmML6CUGPtTSpEhCCFYXoKyFzQbay2sk0J0Wus0hR/XZakDBomEs5QJY0xfN1VVtecaNw7vAnVYlqVFmWud5EUWQ1BKFEXGKAnB73Y7GglVxBhDGYN5Yj/UE1OZxIhNy2GuwDDd1DoxJt1ut03TfP78abVaZXlmjDkeD3V9hgEGhucIXvOV74dBKUUJISEaY6xUl6gRouIihCAZl4xHxhIhc51KTTknIXA2bXhlVMCCDHAKHEydc2hS506dEPLw8ABgue+7rmvSLNsfT1W1KMpqGI2gIvoomJiH88A88NmNMZjHHo/H5XJJCFmtVpTScRxh6W+MSbO0aRohxY8//pjneVFVUkrIaAA2wHsctEDOOXjAq/U6zbM0zU7HY6q1lJJTliQJV3KeId/e3p5OJ875/f19lqeM0xDdMPZcMJ2qaZR12cwBz4w5fDXNuWnqoijStLq9va7r+vHxUamLOhIDxXksCl5+mIwn0Eb0fe+N3+1266srKeVyuTTGnJua/iqG3ggW5Bemrf8/a2+2JFlyZImp7Xf1LSIycitUoWqAnh5Kc4Yf0J/OBwq/oIUjmBY0p4BibpEZ4bvfzVY+nGsXKXwkkAKBZEVGeLjbVdPl6NGj8+5cmhEhdMz5PC0yppRiTDGEelOXdS2EKsuKSwlgimfpUKBGVVUt8BySSvSd0ETFJucYI0RU0L3EjwOGQtjAiOxqtUIMgLoWoCeedb7++te/vnr1iojQl8Zlhomv1+uHh4fj8YgOIS4zRqlhEPiM1oeu6/pxmpxv16Y/HL++PE/T9Pj4+Pj2zfF//A+fElNSMjo+Pzvn9/v9jFmlJCWqFg7NRSiLnU4nPE4cnbX2/dt3P/7442l/WDq62EnJ5SzJp5SyWasVqf2cuVuLcjKlmdKQkgzeF6aw08SItNbE6Hq9Ts7e3d3d3W0vl4vSwnlr3SQkhxBQ3TSgc6BGgSQAgjTIvs45hFs0k/txgMH9z//pKcZAKYRwPB299Yyxuq6EFFVVte363bv3f/nLr5jbUkodDofNZoPPLjjBKW+3291uh/OHhaR53VOBYaiUUl3XKYSl7faHP/wB/W0lZVWU+/2+qqsQQgwBD2672e2fv26au59/+rHrb+fLtVmtm/UaFRJ8BwREz+fzjz/+aIwZx7FpGtjD+/fv9/t913U//vQD2n2HwwEFB2Iz7BO+JuTFTdM0KaVQ0SK0IInbbu6UUp8+f0ANLaUchi6EUJRljJGRIMHdaEEWMMZMclgcXAhhaUrfbdZCiL5jQ3crtDof9lopwViM0U22rqoYY1PVUoj9fq+0TmkWVBgHqPMG772UvCg054woEUUhWNvWbVu/enXPOVdKjmPvnJ2m0Lb1atXC2SEHx/u31saUkrOmaSlG75zOMqLxu5FpnDARSck4pxDI+5k5bbMoPdZOLLcSgxdSSmSRC5cExfc4jpfLparqAAUh4DxKXs9XU86rcC+XCzAJ0DfBysNilb7vX79+rZRSWn95eooptasVEZ3P58laI8Rms1mv10op7z1a0JGS9x6i+hhCgvGAGFL0JTQWd9vtMAx1WSmlQkrDMIBkbIy5u7tDoYnOB/TlUGnlQDi3jmFCMB4Afe/evYPoExLiYRjKshrHCR5goTBh5gMjljwLkCBBsda+vLxMzoGPgEwdqMzfGx4o70pDKwaMFOfcuq0R68ALCiHcbjfG2c8///zp0+dhnJqmKcrKej+Oo9R/WyiI+4zSD9aDGXqk0kjwYRxEtDAcQPnCcaBEwHXVWqMmBTcOnIeyLJ+envDmQQfieR0CY2y92RwPB0SXpSJDNP73f/938IVxpfOgBtvvjwDHOBecS2s9MX44nEIIv/zyh8fHN09PACu4UqYoqr4fGRN13YaQsAUab2Cz2eCt4im+vLysdyuEq3fv3r19+7ZtWwhSYiQQztGdjjEvNKcQ4bJRfg3DsF6v/+M//sM5t9lsLpcz8DHsheecOedCjKWUyM6YmPfKybzJJ826fmIcx2F6AdECxx6z2sE49iwro+EYy7Ksm9rHkGIEbYli1NpUVVWVVRc6DB85b5VSTbOqqlowTjE1Vd0Pw7pdFdpURSmE0MVqt9tdLpe7uzuMubo8mt51HSgreJQolW6Xy4K94qyapqnKstBGa73b7sbJcSFWq1VKsazK3//486otqlJJXhGlQHGz2pZliYGm+/t7HGwIAYwdvCYyMowOwHSFEHd3d1++fFmsFy4Sfgd3EhUDtDHQKtNao68I7HHxnhjLQFDBkQouSMxkMyICfQUVAJJE1CshBK1EU9d9103WcqK2aeBhd1oho4RTqMrycrkYI0LwYNZPoyvLMoRYFAVRXNRVQQLGuHK7WcfvNPwZkzHGz58/rVZryigCDAbIZKCkpJRcDH2fKGFZW9/1SirBuRKYhGcxRjt6CqS1BDqK2220iXnOBhytpmlQnfd9T5wjEwIBHVk87s5qtWJSdl0fY9RaK61CFYyetVSXkabtdguRvuv1im2swHOapgmZF8M5x/fgIRZZNBvOqq7rxGgYhuTd5XLBgtiF9X44HManqSxKTCyWZdlgQ7AUKc3rHICKr1arpmmEYM5NoPJM04AgvRTlKGhgFUBB2rbZbNYxxqoqtJZFYbBjKuatugDMU96XJfJWLp4XRmmteeIQcfHev3n31hhjbLGM3/9d4YF9N6ICE4ebACaIWRJUbd++fROCwasi9h5PR2NKYwjrE1Cc+rzzKOShZfzBjyyrezDfQFkgYcGd8PW3b9/iFvV9fzwe9/s9eALX6/Xh4QF0LiklJP6v1ysSeRAHm7o+n07IXkEFuVwuMUZrhz/96U9N04CI1nXd169fU0qRmPc+MQoxuBB8DD4GLiXjTBuz2W6ruob+MzGGjTR4Wdz8ZafbAsugGwHFDuLp1ncxTzlhNwb8xWIiOJ9ZVoE87LKqqvP57L1/enparzf7/ZGIMLEBW4k5kBPnwzAM0ziOY7NqiWiz2Ux2YBRvt5vW0rlpmobL5dQNjvJsPPoQzrnD4TAM83oykWdfqqpKMaWUBrBRyyLmvRGMz8tHy6rUQQkpldJ1XTPOIf2tpExaPz4+AqUx5cxuhPrN8XjEtCcg+6Udgsk1+Git9eVyAQgJmdKyLDljSqm3b9/6EKu6KcsaOjzGmLosjOZEzjjFRSGlLLWZpunLly9ArpaEF8jJ8XjEiMn9/T3E9OGYXr16hUUL6B5BBH6aJvBTUd0DlkRnaPZcSqVZN8WFLFSeWYmzwrNzjms513lE/NYhnOONgbEKrelV2xL5GD3nJDgZo6ZJDMNgvY9xyzk/n8/OWuw95pwvMm0x73TDfzpnuZhT77IyQohEkM8hkFY5n2XhiUTfj8jJUHX9zVNIebdqq6L0zg19P02Tt04pxRmLIUzjCEb/TL6klCiud2sMDNd1LaWsygZgOuwWbWpInXddN40jYwy0V/giCMXXdV2Uhc14RiLvnPPO974HsAzXgadwuVwwdPbp0ydsFkKe6kIoqjJhCxCl12/f3D3cf316Gq4D4Huk9kop4iylNHYe2yxQJqI0Wa3X1+7GiHHG0NLY7e68c0LPatvAuKCInlIKwdm8lBfDW0QEn4mDRbhCYurcxFgJT4vKLMbQtg0RJ6pdXqSKHGWZoqBMKEUz3Dn35tUbY8xoLSoJIkqU1P/f0Yf/L3MJb3ohC+GMFugKEe94PIYQmqYFVzeEaK2VShVlQURd18MzTpnMir/f39/PqVPOxVDaoxVjjEH4RY8aPR+UV4D/IEyNJYta64eHB9QHSL2R4KQ8XHo6nU6nU13XdVmgWMNYMkId55woQMRpqU9hiCmltm0wvljX1dL2oMSXSUsgbPACm/UaBfXS/YbvAACNHA0Pqeu6djNP+tR1Ddp+yFKACJkg9Y7jWFQlZY4ActsF99NaIUnBE0lEzjk7Tbdbd7teldZKa5wkPhc0FSQXMcbj8QhvjllixEXgeCFDNCEEnDzeAMATnzUpnbVGGze/iO+6znlnCoNmHVrNxphV297tdu/fv//tt99evXr19u3boii44EWpwY6DhYA3hXoRm4i+x+WccyyXOyhbL5fLarVar1bRByFESqS4MFoXRVXWtRH68PzNFIaSLbT+v7/99fH1j7fLlYgDsYSIJqQejTFQjvv27dvXr1/BXwKMK4R48+YNrgCE8CCZF0KAQBuGSzabzfPzM8QeYFQ+j+VrrfEoYVRorlo7Oecma0MIXIhEBE+KTAgXDTcf+ExRFD44SiHOmyM5hF7gJhafGPNtSjMdPiWo/yuFj+mcgzrv0gJERpxSwtIboP8IKimlpqmQ4qjvlgpzziVRWZZt0yopu64bbh3jbImIAItwqWd5Lh4ji8O8G05qrSlxnUdlfZ4Gl1JKKYqiEEph/gC7LrTWEAtgLEEPGwyOREwqaQrDiePi4DsxJHE+n6GqsF6v4QG01oxzxflutwN3CHYupWyb9t//rz/h2oYQDBZkaZVSSt4A6SKitm1BSKnqOoRgJ8uIpnF8fHwUQjCiqmnA0MWaijdv3ozzQjAuJWcsGWOEYHOYFyLGab7mGV3ous45X9WlkJxz7oPbH15Op5MQQisj8z7HpYNLmb8Dc7XWggEI88blXdhKPEvn/b3hAS8K2BeEB0ifSs4ThaLkxMbD8fzl6UPXXX/6+cdLP+qqJSLnvJJaKm2tdd6FQF03vHnzDkOYfT8CcoEQFT5h3/fIkZGnAEdCmYk5DoibogQReQUxEdVlURXmej6l4Der9nK5aCm00f1kbYiMc660946EZFIJbd6+/4ExFhKdrzddlJPzQunhfMFlQH9/WalYFCZ6JznnLG7Wbd/3zg6rpr5cekY0DkNhzMP9PYJ8WRTYsAbhLUzwIkQtEauu62EYkP8G61lI0ohpGGdNLqNt3s81juPT01NilFIyZVEUBY90OBzW6zXmvzC4JKS4DtfbeAsUBzeV1k7eRyYYl6P1UheMieBiENGRuxwudV0Ol3G72cQpRpeasnUufvv2IvVscIyxV69eXS4XMLjK0nz+/Bn/JPOQinOOIhu68Sa7mJIsSmfHyHh/uV6ut2a9+t//j//zejm9eXydKDjXE0/D1Jd1Yd345u3v7+/bX375T5fzmXF5u3QxpKqoBZP4Xwp0Pd/GftJy1pMpTamUul2ukeZnVFTVZJ3zabXeKa2YiUJJrmIIJDEyznlI3qcYuSFupmG0ydwGN9h91VZI7hhjz8/Pd3d32HX417/+dcybz9Eng3pzWVdFWW+2d6aotCmlMihS+1tXlY3RJWeSMymFNrqMgb4+PdvJ9/00jLYoy8RIVdWtv1ZV2Y9dSD4xXTW1mGRIxISxfgrEKHHOeV23j/exLEvnJqghISWqqgLgnsVWKxfAIGrbdduuu67bbbaZDsSIMxv8erc9HJ5D8HVdJzZin5WU8nK59r2N2Hiodcn5ZrtF8hfQf/Y+hWBySwPJHEIU5TXOyH6aouRESkrOmI9BahUZlauGExNCmLLgnPsUuZLXy4UxFgON4ySE1IXERSYpGY+J4jhN2mjGAzEfQhynnnGFEvP+/h4xhjGmlN4+3JGQpij3+z36ZES0albJx/56Wa/X714/Hg4H6M6i3j2cT+fzWSjeDbd1XB9P+8JULJFRehwGTowTSyE2VfX49jHGqLSIyXOWUvScpOZySlwIOQ72cj0VRcFYYizF4NZt7cyc4gjBRjtorWP0RLHrOqVE29YpBa1lSqlp2vP5XJazSooQyvtZ+SZmkQzOOdokx+O+rhqjZ1B9HCajixhjjMm5OYuVWXUU3UegMiyvwgY/ShXKGFMQFaV2dmzqsu97N43/gPCA9hrADWgbINEQjHMhhODj1D+/PB2Ph0RkimIKcZysEMKHgCMbx9FOrqobniWa4feJaBGiinmuz3sP3UHAQTBNvBPEA5UV8ZaAD0hhs9lgeH3JqcFnv1wv4LwKKfzx2PVdXRhT123bfvnypaqqP//5z7vdLuSGJ3JntCVgdpBkuN1uwcfClCnSQmkNeSWO9/7+/h5gzmq1crlNB5SGMtwP97o0lsHhIyKepW+OxyO8IaDn29CHEISSaMmcTqff//AjHBkaX/f391prkcUthBSo4n0IpjBSzeUL2hWgGHLOo/dj309FAXEeZP3W2rJuAZ0jLUXRBvtDHopNsd/TQkIIp9MlpWAKvVpvYgzEWMOYUMqezxCJulxOzk2BkpSyaau2rZumbpraGFVXpZDFOM0br/AHHRoEflBsAXNLKZumudwuQoiyrGJM42g551LqcRqJwrpYK6WCR3kxS9dxKU7nC8Jt066HaQwhvH//HiIrRPTu3TuIWECmBZ8URR6esspbqmKMWMUTQgBPOoW5i/B9K0IIgbRAauVS8CkSI2W0994UuqoqaOEBRvDepyhMUToXlJqXeYFxUNclmh9Ak4ss2kopSSnBvsUpLVgWyNCcc+RYRFSW1el0OB6Pu90O9RZu8Wo1N5mQb2Fq1zl3OhxiVtCBlSqlOBea/U1AZfmlIst9j+N4vV4RWZGZhhRjSJN3nHOJDTlKTv3EmEhpzl7xocBDGYY+JRZjwMAT0mHOGehDgBOQFLL51867xwE3xRgVk0YbwHFEhDQc9hNjFDexBP6nr09t2757LIP34zCM45hiXG+3ACrQCYNB4gqrEEJ03s96a3jBsiwx9NPUDdU0juPpdIpZDm6puaEgkrKuqsK0ExFl8ToistZieWrM8+F4RsaUzoW+H3ENnQsY3CNiLsuhIxKguA95yTEOCp4T7xlwC1B3NHF9Hrr8e8MDyrr7+3spJTonxhiWCAM+3aFDIux9OOz3ow8AZ8ELxqNdrVbaFOg/ww8C2EETYpkHAQgjvmP0A8QHGrNkrzNYsVqllPb7PVgKaE6IvPSGIEuCVUExpRCdc931VhTFRV7RUQd3Yml4UC7tMzjroBSitcayKsiTlWW52+2myR4O52EY9vv9n//85/P5/Msvv6SUALZ2XQfMqpg3gEfUuch00AtFbQhy/d3d3fV6/fz586dPn7bbLeX12rehV0o1xSxNgwbA6XTC+VRVBbtpmkZLFX1giQTj3roUovd+SLN8zW63e3p6gluv63rob13X7XY7xGO8SE5RKzDHU0qHwyGftlr6JUhYkJXg0FKKbdtoo52zi+QitsqUZeWdO59OVVUWVbOqG5QjCwtFCIGFKkoJoui9lZKvVg3n/P5+N45jVRWMMWsT5xSjxwULIaQUlVREpLVhjHHGpNJqXu+KWU7nnECnCnQpNCr3+z2MDakGsMHD4RCyZCkAQyAe8PV3d3dVUyulgL/jUuAbCm3gpNDS93kLJpqrMSUphLVWKAliblU2d7vYNmvOOWfS+Sn4dL2ecRpgIsB0lVJaSxx7ynRSKaWSUkk5t3/SLHuF54UGCdo2eDUp5TQNaL+hDbaEh9utv9167z1jQkrBuazr9na7SaH7qfcuppSUNN57RsK7IM0sYIc7siRtMUbQi2FpeDoss+F9XlaK2OPyTs3lUucwxlJiSkGw1mKFFxHBIy/zB4j3FKO37na5CiGM0hd7noYxpRSaeQ4AzrEsSxwRqM8I/1C5ePr2dRgG5Ado2iMXhJg5bitKc7DscDvw4jEGHGMIDlsgl+eF7rTPuiboDjLGsPCG8qwVpMwQgJEdwrPhrYa8RxlUHcYSDhCHtsCwuIBodOGQ1ayyDAbjLEBCGXFa6Npwwovs1d8bHtLMj57H4oAv4USIxDiOz8/PLy8vIaTClIzxaerh2nBtEBJNoXyYI+rlcgEyyDJdF+j88/MzfgRxAl0+gPUwOKC3YHGklICSE9E8kXe54FZghogxVhUFYL4YvbVjjFFKPk3D+Rx3u+31egXGt5gdRMpQMcBrv3nzRmutdbFaraz13nsptVJqvd5eLhdjRjCmP3z4gJQH7ZC3b9+i/YUSBJpU8IlYJ4I49OrVq0+fPuGfYKbTNCH7gxiZc06XBRpQQgihZkeAxwFXhYz1ershbiFO42djjBa6S5k0BdpGURSfPvx2OBx++OEHBAM80xBCW7fA03EgsOZxHNEjRxWCnMjkdUkwemMKLtgw+BACYzylNE22MHq1XtlhJMamaUpMPNzd40fQ9xNCcMYiMThreDST1bHQ+aBcv8LvCCFGO+f46826uw3G6JRYUSjnpzwpOvnRDsPgnGd5IReMBIcPIj94zKDK4O7dbjekV8iyMYWEX6qygqbKq5OQ/hdFgTI3fCc5A1eOoEWMXa/X3f0dSjH0k5YUZJlDXnp7qIkRmbSe9RnjPBKPeCB41mIDUZIoOReU1hITVVKWVQX/ggsI/B11pJRSa8VYVRTVMuK3/JYYI5Miccak4AC1LZNSJudQMS9E+6XWf35+Ri8QTwr/Cr+BVyYim3enT86hPya1Rp+5jLEg8iH4EISUwzhaa/thgMMlmsXGkQLzLP6P14SfQU/Re6+EspO9Xq8gg4CcikbOUuwCEkRb5eXlpW1b1GpTXrt0f3+PpHBJAsa8zV4pBZJeu5pXdcUYAIeEeRfb37ZTpDzVhNEleBXEcvhSBLClB4DZAJ0XGeC0VWZFcy6JuPeQH5Zo9+KXujxKvLwBmxepiTyQUGV7gN8+n89fv369v7//B4QHvHu4ewALM7ikOOPpfD4/PT3FGMuyqquWKAGlwUoy5NHW2tW6tG58fHxEtY4HvIAJ0zRBVw6qGLC2RawRvgODLaj0OefW2uv1CmykqioMUt3d3QEhxTExRjH4GDxjTAoeKMXgOWNlVS5gDlynzeu3vm8nIjUrimIc7VLKMMZQQ4xZgD7k2cWZUFEUeOVlMhkkDc757XYDfROvgN649/bh4QF9Y3T4IUccvG/b1gaP9Aq/Gr1rxFSt9bt37+q6fnp6KmLcbbZG6ejDOI7YusM5SzFO3nM+t2HR84DxAYJDsglfj5CMKAXU6/vsA8XT4s6WVhhwHiGEDw5ovlJ6mqbXrx6kEHYiOwzYhmZvt77vg595nDHG/X6vpVJFiSViIXjO2Xq9gq2v1ytrLeeM83kwkCgZMytZwn601pwLpTRRMKTwQL2LSdMwYCe2RjYHdwyeIopFk9fwwe8XRXE4HLbbLToumHxGsIcxO+f2+/379++HYXj9+jWbd7vPispEhN9VFAW8T1VV58v5vn14fn6GY1XKzHuniadE1nqljLU+ZvHnBWlEOBRi1tLIsKr0PkgpuBJSUggUmeCCMUYkpCr0MIyqMEIpyTRNE94hLByDNW3bcs6l5EIY7yPwSM4hKudjJGs941IqU5QSvomYEEIk4jwPJ8H7IDyAUI+Kakkcl6Y0z4qw+LtzLqRYKBlA//O+G3oX/Gina3cLMeqy8okuXU9Ed3d31lotOEofALaobsdxFJnrfHh5mYbBjiN8+rI7oSgKKAtIKT9+/MgYwyx0CAFVL3xIn1ePrNdr8C0/fvyIKL7dbtF6xPkTUfCpKIrHx0dtVJq3GYqFvwCyCcvEIcrDdNkXMbwOBKGR8Jm8/xlWhE+kspCiyBshkV8iyPHMTcLlYCBXdAAAIABJREFUxcANfALCGH5XzFRXVBKIMQtfYLPZIF34B4QHGMESvbG1A280RHc8Hs/n8zAMk/JKGehVoB5ECoaMFZQMgK2o4Jb+ATIL5ALIVUGnA6EQzTRMz+GxIZvAkaG2wubOkNWfNpvN09OTc+56vXIhwHy43W7OufP5vNlsYgje+/P5/PDwAL1iPA9o5fu8XwigU13XKVHXDUIoxrgQSinDuRzHOU9BlgQWJhwEvCranpSXCIL4hBQGPXakiuM4P0iwWmFPfd9jvsEUpixLZbQxZnK2LEvbj6DKINcAIgQiFoBpYCPgH+N74NqQI+PMl5THew9FeABo7Xq7kIIwuzeO4+FwMEYBpcFcW/hu6cX9/f3xeFRaxBjwCLTWQozeBwy2CCG0NkIISsxamyhhPuN0On39+vXH9z9s7++4IKl4TF4qIySLMUrFhWTEIhcUk0/EhOSSSSJCzae1ZkyEecMeSSlDTEhd65ql1E3TwtsJCG+IeWieLc5robQOw7DZbKTQcPTIYwBHEJFQEgaJj09EmBBEBgDbhiUj3CKb8WHuFtZ1/eHDhz/89AtWTk2TlVKez5cY42+//UaU3r59C9QRZ5vLC6K5TMcAEMbrKCVyjhgjIRjnhO0C9N3eX8pKTYwxqivOCUCHmIVAiIg4n4VtkHoDpr/dbtoY5NQ4GbgVhISlgsfrLHACZX1ppDLxO6ER3At8okUlAqeKUJ2L9Ylz7qxnxC1kRawvTBm9x5kvqxcQZjgxUJgopuB8U9X39/dKqjP2+YwjBBRA7X96ejqfz+fbFeXgmBXCYau4IFDLR/IEr4LmE94kLg4JJqUgSmCLIV3DxwRmLr5bebA4t6WFg+eCR7yAgUubR0qJ5GzBZlJKdV2Pgx0HO/QT57y7DXVdf7dYfk6yQdtB1xatGq01YuRSeuJzwaXf399DW/ofEx4W/AFOASebQsRgFzImSsKYgou/yawiLMOkpBTYBgbRklwxcVT3bdtSjnu4Y4tIOApYmTevIovEvDGiIi4bQJXVaoUw4OZVKpPWKjjHpJj6LsbIU5SMUqLn52f4cfS1Fvk/sHLx4ymlx8fH3W4nxK2ua3wRZo3POE0TOivee0zcLGFscSK4PygDwVJFpYJgAGOFYeHjI+ZJKSc7219VVVDTcxef8kpUZNDQqX95eWnqulA6xTT1g9Y6Ok8hBu+5+Zu2PswU6RIRITDAeaEiRP204CeYyEebB8O9iILYxwJW3+Vyy+uayxC9c64sIcdtvHeOMyXnz9W2bWKCMTqdTnCm0OS4dTdPsxwpUFEkBPAvcDQx96gXc2KMGVPgksCfGqNjEgsQivMZhhGmyPNOtKIonp+fZV6qAccHDhgSRs4kRC4Bpi3vc8Ho8v7UGTZs6wYF7u12W7K8BSjzIRAxRH0UcIfDAbn8u3fvTqfT4XCAof70009IeOFZ0szGnhN2lmXKOOchJR+jUiIlCiFKzkNKkRLnXBnN8pQSlzwlIs4KKrWSd3d3eMM4Pa3lOIa+G50NlKyUchzs5XwTXDXtmjibpimF6GIiYkwqxiWlWXV/8YNL6gPcDJeCMqyETGjh0gzDoI1JxE1VXrrb5F1RFD7FbhxWWimtGOfWu8SobpthGCZnnfcUwvdpFkCScRyjD0sOh+dSFAV2qCx5Z9/3GAT7+PEj/27fCX4QFR5sYOlmoymFBHSZPMXJF0UR5iqUGGMhzOwMnHbMSz0X4Au1CzL0xdJi5vTjZTHTwxgDuRGtDtQ3ODc4d5GVveGyhBDO+5QCEU3ThG/GB0y5H7koMi313IJkAu3H/foHhIfNZrN0aJfYyDkfxvFwODw/vzAufIicUllWPq8nwzVGMUHEnHXDMDw8PPj8R34nNYO2FWwLlxlwCs+SZ3CpiCJw30B14NpSSufzeb1eY5vjv/3bvyGY9X0fY9BaQ3dsmibs64A2PZrGeDb4CvIgOPflDqxW6xBos9mCDg+mPOruxfUDVAH3AIojqJlgvkA2AVhhJzDSVfRIEOdgBLB10JawoGbZ2aKUqsoKyylxXGVZPj8/I24ppTjN85aIoLBFF7wpCrCS7u7u0IVjjE2TlVJppTvq23Y1jpOUappmLYenpyfoHaGnLaX88GFgeVATXa/n529KaVynGMPpdFJaxpgY4zGCjOG8c1VVOucpUYxpckNw/tPHz9vNeppsVdX39/fbzfZ4PSciREqt9SLNNtuilIvLg5Xn5DpyPm/QresWe7CFEJCtM8aEECmvycMMOT4R5SYnn9eWJJ/J46jThRCbzebLly8YfNntdkQUQgDVQghxuVy89yHG7tYNdQPGAbBjoB8LqhO8Z5xhZA/jltjUBLipaRrEKq2VzIIlKhPzQgjez3gFcAm8Q+djpOT9HDZSpBQTI5rFMeYeJueceRdTTERUFOX9/cPLy4tznjHg4+RcAOcKqCOye6UU40xrA0IePrjAMkE3t6MBHiAq4ABh+bjRxhjGeZ9lwSjDBn3fO+9VVtrAFcMDvd1uiZjifMn/6rp+/fp1DPF2PiPbw1PDkFpTNy/Pz/AP8B5IpK7XDsjEMiOCcLvdbg+Hw5uHh2WKBdAWdmBgvj2EgIW+qKUW2AevMy++ZnC1ebooJedcUZTwV8ghUNCj2ILjHobher1iCHSJEPj/y+WCaScUCijgME+OnEMppZQ/ny/W2sPhAG8shAzjgJAEtZ44y49qrXWFie4QofMvBFGacb8ll5o5Tv+Q6gGbVYCcoHKBc+zG6/74IpUhKZUuulsnysqG4EJiMTGhy7p9fn6eXFCGOx/wFpFf4/5jONkY8/z8jPAAdAxCQ5SLZZy+lFIbc7ycpnEMKVrvtFTWWm9tVRTeT0WhV6tGa3k6HTjn0zRY54jzuq22dw9fvz7d3b/6+Pnr7v7V8/MzikqEBDgOa60P9nK5GFMwTkqpqmq6bvA+nE7HV68eiKKUXEo+jv35TPv9c07lnbVjXZdddz2fj3Vdnk4Ha633dhxB26Wi0Mfj3jmntSSKnFPf34qiaJqGc+bcdL2evXec8xAc52a326TEEs/StuPkJjuMQ4oJJQvODQl+WZYhRp+CT37yU2vaxNNgB2OM4AI546dPn2DuMcbn52cipkw1+XS5DVKXIXHrkwuUJsfYtN8ft9s7Y8q2Ze/e/cDYp5eXl5RoHK21L5MdheTe21s3Bp9i9M7ZaZIxRucca8Q4TDGQoyA4s1Mg4lKaFHnyLDH+8PDa2dE5lxLd3d1xJRmTzlqtRNusKTFU04wEMeG9JyZmaH6aWyMhRAQh54Zx6hlj1g1FYSgxax3nQSnFXFKaMx6lYqaQRmtj1Hbb+hC4SCFYmpgdbFmW3sUQJ+nCre/evH5zPH4lYiFEbJPX2sSYiNjQ9ZILABfSaKEVpVg3jXVjTP5yPZtCxaRDdE1bdV03TGNVVULKQuk//PxLIbXhEhP+4IACdQGqY+2EXtpPP/2kMlfVORcZxRiFNozxwbq2bQNifgxFIQTD/ptAISgpGCVrQa4nToklxhhPjJPQiVjdrhMJ66LzY4zJ+TQOA+NJabFaN6DwjeNYNysWg5b8PI1+moKdKCVvJ8lpzBo+PIcNzrl3zkhFIaYQI1oOnDHOp2kK1qWUdGEEEHwleRLGmHXTPj995VIZqaZ+qItyGEfng5vsZrNhxCQXoN45a8G2Op1OcLjAG5q2edk/hxQSS8SpqAoiWm/Xpiyen5+FZNvd2hhTlBp05H4aBzcRZ8poJK9VURZFUWg1TZOUbJqG2+2CZicRoRGItA8eD9l9UcgQgikwlhSl1F03hhB9Zl0u7FLAA0jLnp6e4D8Xp2+yoL3IQtxKqaLQ63UjBC9Lvd8fiKjrLFGZKCTy49RbN3DG++EmFY8pJmI9chGhEyMhRJxH4A3nvKraEIgzYSefUioN00L2RH3fr9brdrVShQn/EM0lBEk0EpZS3drxcDgA0lFKcy6EVC54ISR0fqSUnItxnGJMUGOOWUsAoBPyI/SlIfd2d3eH9Bm/JYfEeSFtkRelhRhlXgSNZKFtG2sJeTfLY43ZngR+9TCMl8vVe7/b7U6nEx5hyssehmHw3gkhvHdKafgjCHtcr9fz+QwMCuzA5aECrEAKhmwXVOvD4QAO0n6/X+A/GMHd3V3KawkwZxuCgzwvCueQZd1CSAv+hp4KyKyL9hRMEN9zOp2mabDWdl233W7BPlpqWLw+umeAQYRQS9sNVQ5YgACXkD3B1jnn2LBGM+8iIHIzxpSSgv9tGRzLjPgYo3OWkfIuRB8YY6bQRNE5R5RYIsBTt64TYtasx64ulL2QIVNKrYxB+2ccR8jxAnUkYgvIBlSX81tKMwWesi7F0tYmSkqrojRcMBYpBJ8oxhgwM+GcM7KYpul6u/2g1fF4RLqAama1WsGo4CmwA7KoK631tbtVVeXGYXkKxhg4/b7vb7d51n2Zr4aCJp7vAoITUVmW3s87vEBYQkSPMUqj5zgRI/AKJIDWO86Y4IxzJhh556SYK6EYKYTAIyNiMcaYeROCi5TS8/NzypNDMQQUi8CF8MZQS3nn9i8v4MYA9EC2C6IgHB/MkjOGW4yfSikxIbiYK5tpmriSLCXnnE8ReesywxxmTSeWYkTdhqquLEu0Mw+HA0sJnHLoTn79+tV7/+bNG1w3tGQBeQFgQB2jlIIW3tIBlnnFC1bGGmO22800DimlYeiFEMfjsSiqZSwX5RpErlhWYVDzOgAOo800uQHfQN81kykjgSmPZyEVACK9jCDwPIUOR2SMkJJZy7P18mkacUkZI8aY0n9T9hyt5YJPdhJCyEwYq6pCiHk/B9YUAjEuzTyBuHSn4EL/AeEBtwU0j+WeeB8OhyNOodCaccn5dDgciqKsqxkCOhwOMcbNZnN3dwcJTHAD0CEBx+twOKCeoLxaGVAg2rZzNDTmbyoOmS3KOV817Wq1ev76NeTpcCzng4lorZ0f4MswwdA0zXq9Xg4F3W88KiIyRQEGfYyxqqrg43a73Ww2eD/AoJ6fn6HUBHwZhrjYE0paoAqIebA2XB4EjNVqhecdQkAbJkaOVjN6xRh2u7u7G4YJvGa8vePxCMIDfCUq+vP5XOed6Rh4Qvk8D08RIdB+36w2xuz3+7KU+FxIWhck1DkHCAVKNSCeAWPN9DgLuCZGXxTFOMwXEu075FCIECH44L21o+KcM2OtZZSC88fT/sOHD2VplNKM0flyO58v3749F0WBYK+UaprG+yCELMtqHEdr3TTZuiYplTHF0E/WOljRNDrGGGfS6MQojYMlIjt5Y8z5dO1ug1LncRwfHh601kLQMHghRAjRlCqE5LJCNSLo4XD48uXL69evL5cLhD1UHs5gmQ1SFAVGN6x3RVHwFJG1sDz6lFLyIYBWH0J48+YNlGPatp0miweNiRN0y+BlzuczQBuZZ+W89yQ4brgxBjkNcqDEKIQgOROcC06ciHNmXbDWwxRlFoyD716oMl3Xffv2DRarlVoIbClLmaEE9HmxDPza0lUG7JOyKh/nPBLNiDZWtnhPnEmlqroO0l4ul8QZh6hMmPfDwz9KKVFI4S5Mzl+vVwSq8/kMtPZ6vRZa930P8gi+AdkeThsAjs/r6IFGoIOIpAeYTAjBKE0xeeu0VFbI3W5XFObFTt779Xrd96PWuqoaWLv4boeNtRZyQVBRxfHioNCZizHgvgDn8HkJMZrDjDHoduA8EUsQGpd2fcyD7l1niYhzDn1iyhNtgP6Aw2O8P8ZkrVNCTtPEEjGpBAM1eQ45eEH8FOKoEILYjDJ572OK/5jwADKMyFI/RGStPR4Pw9CP47hebXVZEBNIZonQt5m7AsYYtEqKvHJAZt0e8DSM0Uv6ybNWHbwYy9yJxRZhxDElGAFQe7iVb9+ewFfzeWYkxohVMED9nHPb7RaPJ2Z6j8jCnLfbTSm58AFev3794f/5iG7SwsHH6yPCLekDgjxjDMQ7dJ7xbSmLD4ICyDK9HcEGdXqMUQiFdGy/36eUwEDAB8dFBekIl2F+PHmhHnBbhCvOCRNe8PKvX79OKTHGQTfC08ErxDw5AceEogRcICEEhC3RnkW+jM8Ys6QSzjxGH2OMKVg3CSF8cEKIRNF5yxhz3gbPUvTOubowMYrjaa+FFly8fHs+HF52u12kZCdrrS2q8tZ3IcV2vbLe+RiY4O16RRlJB+NoKa2KuhJaXbpb13U+RSXVbehd8E1V4/misMMQb/A1oiBjLIQ5/x2GnpLgXMJpKq2fnp5O18uvv/56Pp9Wq/bLl8/WTtZO4zhIKZSSGHMzxvTjAP+INgYKHZ5Xqi2zmVgOjOY/MiFjDOrXH374AfAgaNOg2J3P52/fvv3xj3/EI8bYTelrBHVstMfEZQiBSzFNkxJcSqEETyE0Te1D6roBTmTpJeA6oJ2WUkKhBhTlx9/9DlUm7BB4F8vKjzBanmUbWB7sYFmtYfaSWfjIB3/rOillYmRSUkUZQrher9JOddMwxnjuZoP4BPOG2/LeX28dtn0sZjYPQisFRwE3vVqtttuttRN86OFwwJPd7Xar1QpayzxTqhaQp+u6aRpi9FC28N6uVlB109AdcC4svcOlZH95ecFqBBi/1joE5701Bl06V1XFatXcbhe4rAUbxyODwwQTEmgBThJTARBBwgXEY1rAA5llkWyengE5tciybCklKRmkXhnDqJBmLHnvLEv49FprbWSIjgIxPpOAeRbL0lpHmu/C3xsewEGCg8aYwuVy+fjxI+BvFFw+BJqZgpyIYF5QcVhYIoBT0BECxgIa2TiOWP6DWwTPu1QM6OvORZzgIUQhZx4Yal6KEVQoDA2gXED2pJTabDaIFq9evdJaL0taQL3F5DA6QsYYVHBt226326Gfn+jz8zPlqXeQU3HZwHBg3639W61Wy7A37BKeesFqvfdFUSy7xkQWJIh5RRTCCQ75dJrBUJFn8ces4YobhWV5l8vFGDMM/Wr1Bi+LvAPFdUoEuhRjDN1yzFt4H0CTEHnSDRDiApKg1URE1+sVpwrneLvdsMsTHwT3LcYoBEspeG9TCkpp51yKQUkegvOB/+7HP/73//7fSaboHD0+dLfb+/fvz6dr1TaJhVXbMhLn03W3vXc2CEGX8+3+LnImY5j+8utvj4+PV9ulyO7vXjk/mLJy/tauNlIZACCn08n7KJh8evqGlJyx4XS6pJQADSPo3m79ly9fiIgzbq1VijnnpNZwx8jjtJFCMsbTetOu1+uffvrp27dvZWUYCfjxsiyBbBhjvPd//fXXzWYDxwTf+uHDByHU7XZ79eoVBuDBsASmipix2+2gKMwY2+/3fd9/+PChqipQ8vCLnp6edg/3IECfTiesQYR3KKpyt90OfbfbbQWloe9Xq7ZdbVAQY8gGlShyCIQEyMpCLX9Jt8uyxHQY3AfqY3TXRZaGjplZy/MMF/yvlDLxSDGWZTk5C3zs5bD/4Xe/Qze7qqpb30mlqqo6nk4PDw+mNEjJcWHRgyzL0vqAvn2RtxGAWxi9v7+/h8GjGjgej9bWIXhgR69fv0ZcQY8dg1CogBHwUkop+Lvt5suXL239hxQ8S9EoiWG6wpQLMwp3cMm7X15eAAMsc7gIPMj/UEAssQRYNOhSC+mmaRrkAaBCpSzEybPwD3zFNE3YZ4WUApTTJRkSWUkIr88zsbCuKudclQuRlJLRGlmjyttEVqsVCpGZOCA4oFohhDKz5srfGx4A8lBW+xuG4du3b2gtArjYn47alFJKZUxdN2iG5NRYIDzg79jJtYwFcM6zcxHgg55OJwTYlImDIevF9n1fVlWiBHc/16TTtLB0cBnwYFCkY2842koy73jC3/GzSJmNMff39/v9fhgdz4pSUsrb7aa1BkT2pz/9CaYc8uZt9p2OCtJ8WCe+CMAhJ9ox5WVHoJYWWYMWgQR3D8kU5RoZr4AuDupWdFMQqHDBABl571OKCwbisgT0MAyQYMBU0cKP2mw25/N12U+wVLuIc2ja41eALH+73WL0cMSI+mjSUNYSgBw6y1QWIYSUfOhGZ723k1Giaeph6IVJyYdxGOxkQ0iRkjblODmhdbtZnz9+PJxP8L9SytvQAw/c7/fr9RrOwucZe4QupA4Ynh+GoVD6t99+g3wW3m2MEZqkQojffvvNe384HIQQ4zStVzX2yXApRzuVZbk/HRljDw8PMcZluy/ywa7r6qqFCRXVPJ2+2W3HacRySrBa8binaSorieoWNOiXlxcEbCnlfr9HRgz3sezURLb08ePH7XYrhPj69evXr19NVTrnUMldr9fD4fDu3bvj8Xjtbv/8z//88u3r+/fvtODeuZeX5//6X/83lomMOJOQBz6W2wFBESTpuEGwFvhWIkK2Aa8Hr7QU9Pj4MEsgeyml4APsihItcKUxRmldatM0TUgzlRNoqmTz2DD/2zKVVNc1LqDIikYhhPv7+5TS5XTCQlncF3ik2+1WFAYNM0x4YeZ5HKf7+3tkNufzGQNo4zgWWhujBeOSi7EftFQsUd/1356+/vT7n/EGGGOXy22poZE+pu+WKBAlxhJRCgFTWecYfVmaGAMRw08tsLn/bq4Wns1mafqY9S2WeREcKYoqoLsL1/Z2u4WQkGSnPHqptU4pck4pQXMpEUXGiDHinISQjM3/hG+Tcm6BTN4hR48xNqyNWeD57woPS1REM+p2u0ECGtMZ+Bgyi8AopbyLiG8LHA8+OxwZSi2krlAUAY8TgAlcHh7q9XrVWewIZ933fVVWXPDr+YJeVtd1WspPnz5h3IFyAYUNvf0wwtsu7P6maQ6HA8siVgBDkTLcblepBIBOZGGczxgO8hGUeAsnL2Y+Mub4EDgBj4LdCF/GvuP/IV2CrS8YZQgOdo/wg+TFGINN8TFLtZRluVQbS3xCrDLGeJ98FodYUIKu65pmpRRDtogzREQBWgWkS+e10jiQhR2AR4Z6CF8JWWw55YFM7y1jaRg67xXl1Y9aK2tdP3RSsMKYlIVS7tv22F1vtyvQkhCJMabLypji8fHN8Xh2LmBwl3N5OJy+fXu5u7szpqzrdr8/DsOUEiMmmFRM+ui8VNINI1faVPXhcCrU5FxwLjBm67qt69ZaS8QA1CytIBSUXddtt3d1XU/OYdBEStn3XQjeGEXEGUtScmvHGH2MAm3nqqp0Yfb7vSkKpVRIhHiANj6wvtVqJZVeeBygRRIRdAZheCEEhD2kcqAzSCm/ffuGxhL8xcePH1NKLy8vEPPZbDZ1XSNB7rruL3/5S0rxfrvxzo3jMI5jonmYY3HH9B2LF04Z1xkp0fPzM6LIdrtdvA++fyniYTlwbUDkl78QkXcuhaiU4s6hU3Ltbs65vutYTCmlsqpCjHAFi9whDBiy5yj0KzNvWsaqDBxd3/fTMMDScAswCNV1XVHMG9E55+Be4ylUVRVjxEgwPlpRFIUxUvGmrRIFLqQP9nw5xujQklmtVn0/EhEu7ELNUnlyC1oDMQYp+VJq+Lz9tO97ISQYyYjBY96Ig7oNSRvLzGnco+VCIWEFYReBE143Zqa7tV4IAdo0ji6l5J3jiSTjRCSzOiFaszyHc/iZnK5Jnv+4PKNnsjLg3xUe4FvxGZxzIIOD9VGYigjxShCluq4541LKV69eQQML1RnqAPBq8EgWhxWCRzxECEHOgkISdQMeMIIeTqcxNeccLPLffvvtP/38My4DwCgpZUxJai2VonFOt3Uuu7z3x+NxSZMBlC3pfCIVwlydNE1zudywiZ7noaqUV0cQEWPscDhAnItyA5+ylB6ODosroJMBawAjRQiBsJfygCURYYJ0oWRY62KetcERISSgI+Lyvm74a8wKWGuFkN4HY+jx8fFwPFXl3KZDywSfizE2DNNyu0Iep/zxxx/X6zUC7cLGUUoNQw/hTCEEUez6W9M03s9dUxS8qIVxYQA9CcHHcTRKhxg551135Y+PwTpEtc9fvrTb3et3v9tUzbqp1ZcvKSUfvHVuGEel1PF4PBwOyDNCCJOzVVV/fXnWWjebLS7b0oH33j9/e+7O834OFJRaSyHYZrM+Xw5Q4QU4Y4xZr9e3bmSCkSBF80IOzPknJQtjyrK8v79bMAQivpz/7XYTQrw8P//8y8+fXj4URXm93j5//gLiw9CPRVloXbx7926z2SDXgUtF+wp7KIG7wtPpvNwbGjhg6bx9+zbG6FkKzmPwoq7r3/3ud/Aju7vd2zdvn5++NFW9Wq2D92VZKq2k0lVdppiu12tKGAmahZDnnrn3oCrpvGEGpQCuNixqzCKjS5GN/DelkFIMIXFOmMSIMcbAiLgQXCZBgldVeb6U1/N5lL1iXGsdnSXGUkpNWblxSikSxRQ9YwaASdu2IQRT6IdXD0PfJyIphTZGCFFR1d9uqMC+r3iE4EKwsiy2223T1Jyz6/XadbeiMMZoYwqEVdwUpK0+2N1uh8CWUoKo5Xa39d69vLyM49T3A9K1/X7PGFtKB4Sfz58/p5QSxcnOG3yL0lg3YfjfmAJAN2r6MaudSykPh4PLA1ssM6DgLhA2EFRwiYgIcD08FRIOzOovA+dwhsMwpuSJza1WLohxxgWF4FJSRPO4CWWBltkOBZd5E7WzLsR/xL6HRdg5hABGB0pRF31kvG63/eglV13XN6ZJRChZUkro5sHiU6axEtHj4yMYdUKI87mfNYJyHx87lQAOLENM4PBIKZMPLFFbN0rMG0tufU+CC8kYS5GRMaZoWxLKJzJliVYhEuEpbxWNKSghrZsSRSnlw6v7YRiss1qbsqiU1HZyUsppGoSojFFSavSHAXbBG07TBD+IsITKAO9T5U28jDGUxmNeegUPdTwe0fDAN7ftmvMOaD7nHBK+iGGbzeZf//Vf67r+9OnTfr9HMrJgLMhT8FN+iqoqUmJtu5nsFJKo2/U//9N//utf/vL582cptdamqmohVFFUXPSmrE7XG1cyxuiiizxR14LkAAAgAElEQVTKQshC3oYbE8Q4MUpEsVAyOjn5qa5LraX3lCIx4tPoGSdOceiuLAVGNNnRO6clD97GqLiQUhZtu+U8Ka2lErvd6i+//sftdrPWDzbsT13ZbIqynm5HHpNkZCS7uMGOLrYrZUpEvr7vnw/70VplzLf93nnXbjcpJVBXGEt9fzNGvRyeV/WKKxlZtMHaMI1dl1KETp/3noiklIARpBTVqmSamONSMzOZh9cPdT9KKbvLFaBW3Wwwsi1ksd7syroIyetCNasaC0t4CNPQD6PVxlV1q03RtOt+mLQpiKgsq7KspJRd11VVTUTO+fv7+3/5l39B7obsEr77v/yX/+XPf/5z265iTJ8/f7m/v3/16pWPqZ+GyVqpdNd1SmpGfN2spn7YNLVi9J//8EcievXwGLIOSlEatDf2h2/Oj9tda621o+OMgauqpBScKymrsuRZUwROBJWrMWbs+mOIeJM+BKX14GwMXnHC0l/K6lJKKUtBFzrGkFhIwdvJGyWkIEaJSeG8l1KKlATndprapmHB8eSNFMGPUsr1uk1caK204NwoyUrnPGNJUEw+cIrGaMbIOauU2mzWwzCE4FMKw9gzRjG5stLX2ymRH8ahLmulNWe0Wa26rnOTf/3qDee8G3opebtamaI4HA6TteM0bbdbqZT3PqZg3aiM1IWx3p2uZynM0FsiYafAGPXdNA6OMWIuMJLBBzsFRvJ6uShZzIJj3i8gAXKUvh+0NkQc7P+i0NaO1o1KqhCCMWpeFxE9ZEM5J+cmxpJSgvNFLsWjCY8Q3vc9yi+ldUqiH8/eR601cem8T4n5QCGQUiKE6L2LEaNv3Aa/KHZIIVJKLgts/L3hYQnC6GsBNr27u7sNPcZQOedlWfT9UBTFMI6YKUWURvaHmhFxFZ6a8iwlBBThKK21l8vl559/5pxjOShAN7TUhBAm74DFHYDPlVJWVXE9m77v1+16s9mM1nPGhJLTOAXnvfd93y+LvZbcP+aZdaDnt9vN6GKz2cQYwYhA9gReEJrMMa8IhiksrFBUmt/DL6h1UPqgLZxSQms35F3EMUaQMSC0B4WfpU8D4C5liSRk5biTS2a0ZHmMsc1ms1qvu6GXUiqtgQZAoAIIMuiYfd/j3LTR2EU6jmNVF0opa6fz+Xy5XmDQTVUvBJWQxcWARKO8YInqsoo+QF4FR4REtVmtlE+ykK9fv97vv/Vdt16tUkqrdsU5izEcX55NUaYYGCVnbYzx97///bfnJ2SvbdtyLgFkA1SEcw/eobeGbAvYF9gv2/VGSf3t+durV/dYJnE+n4Zh8CEgR8NRAGBRSoXgxmnEGeKfnHsJITRN8+OPP4IZgS7Lq1evHh8fT6fD0p+H2OLT05O13lqLVdIsy9osC8iQpd5uN2w8xcdRSh0OB+Dd4DuidkS/+uXl5e7uDs/rbrcbvz4hoRmGQXJxf38PSdHudkETi+c52Bjj0A9GKylVZKHQhmISQpSmSA231iFJwoaSqqqgHTnOm9fm4RVAwYIYqsnVdsOwuYGSTSlGt9yImMUkUGJqrS+XSwZGEsAhwAMyb2XAvRBCEIspxhiD4KS1jozDt8BXFIUJWZAj5aXFy5xE7tbWRSnRHgP9ZqZvaC2FTJHw+IQI8BiFKJQSEHsGqmHMvM3JGHO63qZpkkpjaxuxJASXEmJcEK+klKIpNIShpmnS2kC6HCo7qPgXiDITTDQ29ozjaIyRkoNipKS+v7+HQiKgPNxf8R1fFqUDiG19jx6ths9BTlyWBfAVHC/8oXNOSsXyNog8MMFgxjBIeGZQAcM/altcyJrGoCpjJ0w/jTFG2D120FNGydEIRQWntd5ut4+Pj6gkDocDwJZZlzQ4QE942PlEep8lD0GoRyES8mQKuATr9ZqI/umf/ulweCGiOLe8xDjeiqIsy/J2uaK0T1kKf+6QX7FZ2goh9vs9ESHHx1v6+vVrWZaPj4/gKVdVVdftmP+kvMkShaT3frVawUCNMegvQfpfa40eANobMcbNZgOdO/pusYRSarvdooR8eHiQUl6vV5QjgL9AlFr2JiHjw5gVhiSklOAXznfVGAgg7na7hQdCuWM/DMPxeGSMtXXTdR2n5K3rrjel1MPdvFcLpYnPu8GFEJWpYMf47PPX09yRK4ri5eUF1R7aRU1dJ0OUeNu2Q3f59ddf+65vm9XrN2+sjULwYeyv12t3vVg7Hp6ezufzf/tv/+vj4eHh4eEvv/5WluX5fH16ekIr9Zc//mGz3UopfYpKKT/ZzrqmaT5/ebpdLl8+f67rum3btllJJTabNYDpsixSipBfhBfGJAdGQ7q+p0SMCaMUi94Uat20z8/PifO7u7uqqp6envAQN5sNVh9//vxExPf7IzD0X3/9q3OuaVqVV0njgYLRhKPY7/e32+3x8RGufL1eA0Rdeo8QnVZKvX//XimFxnXKm5fO/e18OS9OJ6V0Pp+323kWB3w/nqdBh2FIKSyU6+8uLylFyHtMXtzLszLVMhA3ZgXi6OYNDSEENaNSkYhut96YEtJ/MRJjnIgT8ZRY+/+y9iZNdmTnleCd7/XhzS8CgURSzBRLJRZLtKIk076tlvrVWqjM2rSSNSWRSskokkgEYn6Dj3fuxXF3gupeUYxFGoAE3uDu9xvOd75zVpvX1xNjIoQwjk5KLSXLmczbHhYIDyFECC4oEzyzGKAFC38bmuKXxQfSiXPO+wgMCskGIBghhFGRIhFGUcKDTyQzSjghNMboLJQEpzqg7/uXt5ef/ex/9X1PCJNSj6NLiVjrrR2AOIEE6H1MKW82Gwi04GMAi1uv14SmYQAreiCEQEIRdT2Z5xaQwFkmnfgAuM6osytWlUUF5WOEe9x9VJaoJheYGkMX4N5CCGASuNdgUi3F6JJQ8YPfonDH4FbOooq4I/gW4Y+yNU3mXvLl5WVhyI3O4fminG33O6GEVOp8uSA9YI6H+sLMnk243MDO+Lw/WVUFigswnfF98KbQAZSzp3GeVGgCksc333yDJuZ4PBKat9st0ENjTF0HKQU4ACTlhbiZc8bZyPMCJ6CGnDPoSbvdDrzVL+MpIjVIe0sXxWebb8RH7GHitu12O1SjmFWk2TgFTC0cRbQU+F6oKJesgxOyNKrgt0B0BHdXTBqovyMX4Ukiiby9vcWc4J6Ifwg/YRy5ZQGHMSakmmkFcRzHfmgxJkGHkWMyxuSYGGMuOsbYaEdkyi/5OXYcVqWAMG/XNNfrVTBWFYWWqtCGGUEyi94XpmiaU1EUP/urvyx+Uf3TP/+Ccy44lzSPXdM16dP9p1/9+3f/+3//X6ZQ2+12vdo+Pb1gtv/4+JhSOtze/Nl//+9CCCanEX3f94fDAVxPMCNtaXPOMAvb73dSyqIoY4x5voxfDodSSkLKcRwp5fv9frdbLXcQTx0aL9ys3W6HVo9Sejwem6bB4j1eEOrxUA4Hf/FLeTUEaxx+VDbIEJjZoOVCxuKcg/6wdLdCylVdj+MoGPfeD10/DIOYF3FRU+NG47FhlDJCcowuhOg9pxRdhRBSyoyB7UIjxDlCAEKyR8UdY6SE2NlLgE7KVAwnAjQN0BnyLDqNP8cVoLNRY4xpyTF8Nid3zmWSGaWcM04yoZPor5RSyckQEMdkCZ2cT4YHeFPsxGGCguuJRTZkUMy6rbuEFJXRPqSY0rVpoGOKo8q+2Ipgk2HGxN/TWo/OGmOGtmecDmNHaBrGzrqBEIK1c6UUykGcx8XOAWcZ0+yvv/666zrAPqvVahyhnJqFEEJygIoLJpFmLQmUWZvNZvGaRHOptRSCpRTIZKchF6tqtA503pyIMaZEhLC4xcDD+byejTiQZ34jYwyMiT9CeqCU9n3/+PiIBqooipQTk0IXRUpEGp0IiSRnRoVWyTlEcDRZOJkQ+ESjAFcGXJcFxEdLDgoNVgpQ7IBpt2xOoqbGff36669BQ4L0rinMMAw++JRSTtOE2Y0W1eI4q/iu1+tMJljpcrngZtzd3SES4ZkWQoDJjuZ0GOzizIXLjUcNyQbGcJAKx8VRsz1InuXeENQwHEaCNMbg+6aZ6OacQ9eFwzAMA/xmMZXCE3C5XCCWRwhZLsU4jqt6ZQcbgod4GybD1+t1/eP/gTuIBxdwlhBCcElSTiESmrVSQ09SiK8vL0IqpRSeeCZoWZac0BCCogoqrfi/c4/PGJ30ADCgqqt6tVrVq9Xt8ebl5Q17UErKdb16fX443tzsNvvffvzkYxqs2+82Y9+ezmdvbd/3h8Nu5Svn3Ga9Y+zf3t7O3k9xAbYWOWcmxaqqP338iKArOFeziUrbtDGkqq5CwHicUkqv1+s3P/whGOWI8pjyhVk1JKZAUo4hkJQF49v1pijLBdlD9wYrtK7r97uDD4Fz4Zw/HI7OBe9dCPEHP/jBwq4B9evl5QVFBobSwJ3mOf9wOBxeX1/Bjp2HTys8BofDAZPP0+mUchZCPj8/C8bx+m9vb19/9VVZljlNMqVsFngghGzWmxDsVIIQkjPhjEspQ6IxekopLKoQO1AC41DgSdaz50fyAcvG1XpVr1bL1PT+/qGu14QwlPCUUrQOnEshlDGl994Yba3XusgzmVV84UxO8W8Yo4xTwRmXSqmQCboHOjPr+OyfEULIebKXQIW6VNkxZkKY9zHnUSkdQmSMKmOathm9rVertmnrzUoaLbQCPTSE+Pj41LYtnOkeH59SCkIIysU4jiKmoqh2u11Kqbwrqqr6/vvvU8roDxBMFgolChSkKCB1UAGhX5DdpWQodmcPGOecW29WKCkASqNI5bNaMyrpZZL08PAAbjTya0oxpYxxrJtdfuPMs0+zBBGbjTJxzRG9i1nBHoEX7jh/MLjEfu83jDVNA5B3u93e3t6WZcno5B+31C+EEACICHkLw4dzjtiKCweyEC73slCNhIwGYqmjUVJ579/e3jAzULNXD7j/OEhlWVo7hhDGYRyGobk2XdcN47gcV1QWenYHXK/XsFBB6IEzzIcPH/72b//2L//yL+EBgr9fliVUYJHS02xKQWdmG2PscDigVcLSGXq6YfYhQYkHWHCYf4B34YvzeUPCe3+5XOCJZGf7o7ZtX19fHx8fX15elqsKcQUokQCMwnhmtVoJMTU6xpinp2dr7c9//nMcUQw2MMAgs+Cwc05JJYQoTOG9f7j//C///C+oZ5d4B4NrZE1UTLjFZVmWZaWVqopy7AdOmeSiLApO2dgPxhgtFaO0KqoUE6N0v9/fHG/v3r//i7/4i3d3t1qKuiqup9eHTx+dA42Ky9naE4svLy8vGPQhsuM8cM7roqQpR+cFZXVR7jfbuihLbcgsPQJgF6Uiml3cNeRmVFtGqdKYHCKJicREc1ZCvL+7+/DVh7ou8dgDYcPP9dIaU/bdoKTJiTIqSKa77QFNMKY1bta1RmUHxjM4qUtQRhMGRjU6PFibydmrFe3Iy8vL6XRKGV6teb1eH49H8N9w1NO8+o4nOecsBBeMwz0XGq4k5xwTTdPS7JcXoZzN6xFzUXuiVls4+ISQ6/UKeqiUEpMerTVn02EPIYAk452fyPvtJEpBCF0+J0ZxeXKGoAuXhs0/S3lLZyvQpaIHOo94t0SVnHOMSSmdM+m6HjYYUH4LMRBCqqrabLdlWaacvPd1tXbWS6G7dhBcUcIFl+NgHx8fJ7UJQuAnsVqtttttWRarVV1VJaUZDg6Q0QRdHqZeYRbhB/MbjDI2e7ctaBI4V4CGEKyfnp5gZE3mWj7NGq7zVGzahsNmtRCsKLRSIueUc1xG1m7WjV+eAUxBlnZhAcAxJkGoRHOMBAaU6Q/4+b3uAQNY7CUtuJWPYXXYW2cJyUoroaQpC8Ko0pPJDD4BkgpGi3Qm7GOhFO3VOPb8i2UuTHFRFy90IGzSobLGc4YyHJfPGDOOtmmaYRzqlZJKEsbyLASCygsFCIjnRVGUVYmLiz1MbIq+f/9ecAlIDnMwTHGRJDDtQG7HBAKwfppZqhhy4oE4nU5fDo3x7ofDoSgKWBXqWTkcW11ipszTeb0DF2e32/V9//d///cvLy/Y9Mk5397ehllsfLPZgFwrpPDeM0bxGRhjKSfG2NvbK+YNCD3AE4wx/XUYUhe9pzkLyoJzdhz1dvXu3TvrXY6Jcy65YLO7Ie6dm23vkGv7nPxsinA8HgGjxxgzyYKx0hRSKM75+/d3wY3N9W1V19a63W5njFlvVjHEl+enprlsBTSQE6XTjBGXCGl7nD3UUGpoxZFKAeYAAGnb9vHpqR/G1ap2zl2vl7quyrIghID0iYHkzc3NsoS/1GhkFpuilDHGKCVdN4AHHGPElceAB9U6WitstO33u9e3l/v7ezSaGMt777fbLSEELDWlVNu2zjkEFzpvrhFCVqsVkgH7nSibxgJtjHEYx81x/5Of/KS5XMuy5HTaQwbGiH+I16GTLp6MfrI1XII+Ywy+QnTWY1/AJcDfCMSIv9N8zgdwJdq2ZZwrpQpWKaXu3n1VmIozSQhV0qQ4Bp8opc6GtnklmdnRoxnyLvrgOJ+gVIBsUw7IE2CVSRYT3TA55zKjC4y+wGU555QmPXNAtdiUhpI8ThZgKDz2PgTrnAthsCN2InPOq826UBpjpMvlst/vgbl/9eHD+69ut9tNtd5wzvthXK+3iG+71YpOu0ctLMJQR6KVhLQiIeTLiQiEBhDB0DVKOSH+eG69n5BAdPDIDXKyRv+dwdyCXQOHwM31wYfgAUV47yilOU+LLEuOwcNJ5x9kDjKrwSNwActZeo48O+79l9JD07UueCb48fa2LMuu65zzlFJTmNHZqqp+8PXXp9MlxlivVoRSQjJYH4QQIJtN0/R9x7lA2tRaxRjf3l4vl4sQDNXcgmnCUo1SZkwhBI8xMTYRhCGQgBuDec4wDL/61a8u5wvJnFEhhBRchpA444KLnHKmmVDCBcdhiDFWdVWYIsaYUxZCEEqd81IpY4ySmjFOKev7HkKWSqmcyTKKQP+oZk0b9AqIX/v9HmeDzfLjfnYVR639ox/9CN0VuqKHh4dh9ovHsw48F1gqnnh8a0BY0I6Ps0UM53w2UvbGGGcdeAvejn3fm6JwztG+L00B0AA1IJClsiwzIf0wOOfolTBGTVFIpYzRzjtKKeO8LMvdZoshudY65ux9cM7l2SWtKitGMg0j58IYY4qCUQpmSFVWKaacwu3tV2VZDH17d/tV318450rIdb0qpeL7w7qulBSCkNKUl8u1aVophQ8+xVwUZr/fPj1Xx5vj/f09gpcdbSQ5OY3pfc45Rt/3Y71aERKd7c+nN614jFEwXmhTmlIyhY/ddd3d3R1ayWXEgkOeUkLNsdnsGOdSqufn5665Ho57RlJwFgYAT89PH776sN1uQwza6Jubm3EYpZwc2eTMGBFc9EO/Xq9NYVJOnHPKKE20ruuyKnPKy/4E51xKUdeV1oZz1rWdc7YsC86ZVspobfu+MkWpjdwxRpkQPMcoBEspckoKYwglMcacIuNUSknp5EsTU4opUkZJhMcZJsB0mQTk2TwD5RrW9BCXQSsXWlnvrLWE0t1uRwnxo+Vw9EseFUYIlJAYwuS6wRhTSiLjlmXZD1lwzjnvu/7aXAEhUEoTo5mQmGKmNHiPSRAhmVLW94PWk7NhjInzBHvJGCPnrKpMzqxprsCc264rjCGEFGUZQ/CTrW/iTBSm4Jx7FyjlEJSlgl+bZhxHHzylpCjN7rD54Tc/yClyzqQppBSlc5v1pigLbz2k6rSW2+3G2lFKScjvRinorfls940LO86mXigjYoyjdXlm86eUYkiKK5YZwXLJMCilNut1JoRRGmbX6HEYIHpWFIXgfBzH1boGpDaMQ855tFYIwSgHxsCmTs7FuOg4EPI7248pf0DTBdXJ3JSIvv9jaC6dm/Pnp89fvf+667rVZpsI00XyKejC2OcnKXU/DNZZwijjjDBq7Xg+v2E2m1Lg3Fyv5/P5vNlsYpz8yNr2WpblMHRQMPbeL1pGYPFX1UoIOY4jhJEZ4yEkkAgBzjw9PSEZokMcrCdU5ESd9V3TrVardVVvtxvGaNu2kFTxwWUick6FNkPX+xyDj7owMaWiqurNJowhhKi1GQb79dd/cjwevfefPn2acETGwu+bS0MzFTNtMmu1Nk3z+fPn3W7HGMN6OQCN29tbzvmnT5/EF1ZlcX4s2KwxAFRHfmHkNLFIZxGVul61bUspDyEaUw7DmFK21nVdt91tU0rWubZtTWHW63X0IaUEtozW2s+G8oTREbrzgjvnuBT748EHO/SDUDKF6JxbVTVmSIlkQpksDBOiqlfttXEuHHfFhbKiXA2jDzEXZXU47D59+t5HH4IXlBmtbm62x+NhGNrNZvXbT796fXkysvB2sEN/c3Pz9dcfttutG4d6Va/Xm+fnV6XU+fx2OBxu3x3KSieavvnmm1/9+tecczoXpyMf3NiPY09I2h+2D4/3zo+U0hCtYGnsrzc371LK0cfL25lmnuPvDIXQxWOTMc4mdMMwhBCRtmNKSqnmen55+LwulLWW5fj4+Z59+Pr+4b6oCibYZr3p+55LrqnmnL88v202m9VqI7jSygQf62pd1fVqs277zkXPiSCcrncbrXWwjnOWUqyqEieWEJJSjMFSmjgjp7eXzWbDORGCNk1DfPj63V2IzlrLGFTTM8mOUY3vQ3OiNOfopVKEJMJoToQyyoUglGZKmOCZMvGFlgyYMGzWT1vYCsgcQFLKqjrc3ECqixASfWCc3t7sKAnBB8YYJYJkr7XuB4c4yDnXRqGUjskzQgpttFSU0q6hNGWWiWQ8cuq842TCiKwLOWeaMyHS+wAtfyFUjMlaTykdhlEpeTzu15s6hPD29rpeb6py9XY+Z0ISIUYpIaVU6uXtLT7H9XrNCAs2BOe977fbbfTOmLIfB+/9n/35nxWlcc7d3h6rVUmmlXIiJTVBlqUkNKTkYpxYT4CSCCExxqenp8PhJsL2bh5QM8agv4L6HUcMgFL76fPcj/KhHYILXdcJLpumXa9rzliOMafEOSc5c8xmcr5eLue3N0KIG0eJweTgpSSUiqpcvby8kMzKkhNOYgyMMZDKGEP/RxibIET0+tg+5pzbMcTQxxjLyjDGuq7RWofwB64+/F56sNb2fX88HgG150zats2RjMPYNu1+d+Sc23HUWt8cbxghvVIYhnzJZNCz9jWQGSAqZN4dd/M0G2uliMXAx1Gwk9muT84qcpTS19dX0IG11j/+8Y8/f/4MMgmKQTzufd/hry39lHPODTaE4EMIIfSnEyoCrbUgv5MCNMYcj0dr7dvb22azAbCT5qUNMmvNA1XE9On19RWV/sIHxfkfxxEy49iL3u/3YKkiT0DXaJ6BDxi/i1kHFCUJqF+EEDCsl3Hfwlb0zoEMc//wYIoipeS8X63X+NYLXwUzcID43vv1el0UOoQQY9Baj3YYrS0Fx+tDs4SmzBhjQiilhCnQdINtrLXerCpCSAgOQx3nbmKMh93+r//6rzLJm806pcA5vbt7r5T59OmhLivr3Ply/vbbb7kQx5ub3/72t7zcSinv7+9jjL/5zX/89Kc//dMffZNzBhg4bSOGjAHP+Xw67rdCKFx2rfXnzw9C8BBCTH6zWX/4cBdjEkLGmKQiwzi8vrxqbax159O5qipTGNRclFLckRCmsR6NEQ064E1rrdKadL0QAo7cYC4uZNCiKCBWD3oFQBtQDwTnhSmcsykmQrKUKvgAjAJLkSjYwT5su4l0hxkbkGKkLillJtwYwxiJMVo7LEMOAPKU5TjpZQn8Yon4yxEG1MBmnuUyQYWjCcoUDMxSSj4E4Gno0QFnxxjLskopjeNZa0Mpk1JxLkieziZGEXy2nIL9xjIvJfOGRCbEWielDN4751LOwXshlVbhfD7jjuMgqEmsgXBuiqIwRqckIJqptEJUwaQHSC9uKEAeLENg6x7YclEUNzc3t7dHpWUIYbNZrVbF765Qlk07xkgJod4lZzs6b0QBjYwxArpYvgtyBnodZPoJtu0ntJxQ0rZtYQwGwiRl0GedtdZKHElMI5aFCdw1II1pFmpDEYw4iY6Bc06IK4oSoOKS3eMs3q5mv3q8S845RRpjGMdxtYZ8U3TOXa9nQr75r6aH8/lcluUCeaM9l0Z554zWfde9PD9LISnlJOduGLDbtTANAJtAyIjO0n6AO3CWQgjwyZsISMZ47/u+e3nJbdu2bTNR+1NyzlZVSWf1kuWZu7m5wXMAQAYjQQxUu65bqG+4yl3X+XFS7NBahxSVUv/0T//0N3/zNw8PD/jKoBujosejhqEcKAqYNdFZxwbvKKXE44h5e0pp2RXHwgTGEkt8QbAGoxkWNGJW3UrzKj+KfSQA7EaklO7u7oBdQgwSU03vHIAdzjlsILVWjLGm7zH3RrWIx9paq1WJqIE/t9Yj9Z4ul6quYVAKzkahNNAD55wqq2EY7DDu93vGmB1Hvl1XVRmCE1xUVdn3pZTyZz/72TffflMURkr+7//+74SQul4Hn37961/Xdf2v//qvPkafYjf0XAqh1OntNI7j8/NzjPHTp08/+clPQL6qYzbG3NzcXK9XRvvn55emaT5+/Fj+9Cc+xK/2h6KoUs7ny6WuKmN0XR1vbndlZShlJFPGeNNcx3H4/vths96e3s4xpBBSUVRlYS7XE55SFBMoHVJKwzAiQC/kGaV0jBGjL0i5IL8655Bu8WkRp8CYKCvDCN2sVtdrbpuGUrrfa8Z5mn0UjDEgKWBCuIQDzHVyzlD4YLNCqvce0hgI9JSQBGBEipRDzoz9vjEqYAS8YPoiPbCZ1oKJMQIcWtil8uBCLK+DGUAIQSkVbCIkMiYoZUIoAnIK6chsLr0cTBBpCSXjpJcuGeNKq5RzTJlkOg7TyiFjPCVPMkkpYUp/OBxA/MXkchhsWRbGGCEI5haCcwQWjNaxVwFO/M3NDbDrca4AACAASURBVKIkKEAQQGWMQYO5LEvOJycGpSTnJOUMFjEhRAjlbBhH+/DwQsiUtiml+/1+lmWbdmmRNQGGF0UBUZBhGFBH4iKkSWYx5QxttEBSpowMY88ExfRoUfig88aM/MJ+A68fY+RSL5zj5fYZU3yZqPCLEMJC+0QVjr88jiOj2JefmEtCTCo4f0Bu+M/pAexPnIf7+/txtCklEnwgcej7r7/+E8H5OLhMImdczyouM4GBHA6HRfobnxi/wN/BSAB0DjQohBDIWKYUUs51XW63a61FjC5EvyRYlM+4In3fv3v37u7uznv/9PSEe/P6+hpTRO0DgBjPk5SSZdp1nZvHd5TS3/zmN3/3d38niOCcIxC8vr7e3d0JIV5eXq7XKz4z5uFy9rzVs78uphp+tp/DRB2DSsR6dJ0AjlBfIEPgT3B0kduWKmC/37++vqLefP/+vZh/5hssQD4pisI5d7leKaXWOqRbDGP7vr+/vw/eL+kBM54QgtEUhEtCSFVVWsv1el3V5fPrG1AslKhKqc1mI7WWpb6/v1/UKSYge1q1rcZx6Lr2crl+//33q9Xqh9/8cLfbKCVTTlCwiTH3/fh//9s//K+f/hQfNYRAKBVSbnfbc9vCPw6cpc1mIwQnRLswMdyttVKQnDPW05q23+73TAipze5wXL+8aK3LVbEu9W631VoyJrwPKeWyLLQuPj887vf7x8en/f4wjrZp2s1mvRRZOJDLJorUCk8L3trEtL+5pUz0o8WaAtZCcTeddYhii8gu6gDOubOOEtp13TiM2mjBBZa0U0qYUQP0g/c4IQTZBRU9LnKYt/koy+M4Ujo5UIYQFDfeh5yZUjJngkMhpVgWZens9MIYs9YvW1NL2gghoINhsxSVnBe4rHMLBREPSdd1MMMJIRDOQyaJUqk046ysqnGcHKvw+uDGWGsJY5SzkCLlPBGSCIGaxTCMwzB479brDefwYkvG8IXLBwCAUto0DZwBKaU5E5LhIcYYZavVCiy1ZQy72Wxub28fHx9RouG04oHZbHY4ZUVRMI4cxmIkhGYYx6ZIL+frOHrv4+vrybm+qkqI/+NfgTMGis1+v0c6R29hjAFTzloLotf1et3v94RmIXlVlTlnpWQMQUoxjgOXbOG2TbF7TttLxQkkIOd8uVx0UQGyw0OCkfVut0uJoA7mX6hiky86G9zx5WHmnEkpTfE7HWjgY//V9MAYA9GoaRqti6qqQgiXy9kGC9WKFKLgnDHBKGV0Om8LkyfNPpqIxcseGdh1VVUBCQVlCKMImAqknMZxDDGMduSCL3pH4JnUdY0tedyhr7766u3t7fHx0c+qud77mOI4Djnn4/EINQusX9webtq2dRC9okRKyaXQWocxLIwpcCvFvNME2sN6vQac17YtpPk3mw06jCWx21ngjM+W3xgHPT8/Q/Dut7/9LXYXln+1XB/suAE6AwW+bVskV1w3QExLXiGEbDabGOM4DIzQy+UitYopxRjbrl2t16+vr+vVCg8WstRSFuG6zbVPjjEKKZSSOGyYwqHjccEHP3V+RVHYYUwp9XbMOS+jo6enp9H2gDvLojAG+sbucrnAfO3p6bnrun/55S/LqizrKuYspLDeFVU1uOFwOABYAG3a+7C8OKokOQkCktVqZb0tq6oo65iI0sV6u6OUKMEYCULIsqzbtvMuMMZTSsNo+364XlvnPGM8xuR9iDHZ0e32W6NL7733EUMjqVQIkc/y8lLKbrTr1aobLGA6FAdgakoppZCACoEEYpXdez/0fXRuvV6zTLy1jJDo/dj1hDPEZVA2oEO+7EbR2XJn4drhTUfbxxilnJZvldKcTOtdjDEfEqUkxoiLxmbvmjwbu4YQQphAp6U+XWTjlpCEdxyGQc0zKrTF6GacczmTmDLnMoQwjrYoCkaYNoVzHoviaIxQJxFChnF03ufZlodMvnKcc8mYj9GFkOq6cC4QkgELowZFJwdx1hDinAPIOAbGWEqei0mYAFo4QHUW6g543kilsHFNiWitsWWWMkSdedf5EEclFWMiRtL3w+Pj8zi4vu85T5Dh8bNSDrxssa2ptQbJEBgj4hg+A9zAQNw3RofgpRSMsViVjFIh+DAMTEykdjMrjY7jCLQWWQHha6Gr4rHns5o1m3boBGOkmM0e2PwDQlCaLRIAPFCKzocaY4xR1tq+Ry/yh2WH/0/3AGSj7/uiqOq6fnp6GoYhRF9X1dgPwzjc3rwzWish+65D67RMU5GsgBQjRvNZIHNZisYb3d3duVnLqO2uIXhrR++91mocBzLxw/zS17dti8NW1/WSivDr5+fny+WSScZ+PBbxIJCy2+20mWSKc86UMa219Q4oENB/59zr6yuWR9B24AlAN7A4bOAr+NmMDG0HDl6aSYR4NWvtL37xi+UtGGPb7RbhBieKzns3Ukpw+C6XS4wR3w7kSGzexZgJIahS0RZMAEVMTdtwOXly9P2QUkIuB4AAfHzBmpfIbu1Eztnu1kZP9G2sdCKxOeepZFrr6AOEizElSjHyWXuj6zqpGBKbNpJxQhkhJHddTwiNIQUffUxPL88/2vyo6/uf//M/Vavaef/+q/fDeB/7oWma4/HYdXWeTXGbFgO0gHyMkFGW5ev5RChTWlPOirKsV+u2bYwpNnWxXm+l1Ixaa/vb27213trEuXh7OwkhcybjaNu2q6q6aduiNHW1VkoxNgndMM4Tydi9wD313htTECZSelwKBaRJ7/0wDpxzeKSDbIoWjRGWYpRCbNbrADtb57xzlHMgpWBJ4qFCSsAxXuZbKFSVNNhQE0JUVaG1btsr55xk4pzjfPr7QoCA34LjB8x2gSxwDN2sRI1zFGZFejrbkOB2O+cY50gqCEY4ayEEIRVlPJOcCSWMxZTZrGutZ7ecNOvPK6W6vg8hcMFTSinnwY7OuhTJUuigUUBmkrPPNpnV2HBC2ewJat1kLT4MllGpS73A98AhsEAKxcD9/nfbS1VVOJdSSufz+XI5hei11sfjnlLqfYixVUprbcZxfHt7RQtLKV/67AVHopQiN6BkBLKPz49ZNOB+eKu0bVuvq6IwHAVBXTJCi8JYa0dnEcqxLExmaA63Jszufmp2iMOf4yIAV5iBh4nPKme34zzLbKD4w00EJmF0QSmu/KQPzTlX6j+vP/8h6QEN4+l0AgCy3W6fnp6EEMPQV2VJCVFCKqmC833XtU2DL4mmCcebzFZChBAUsJjlAn6Z9ldjXKwNq6o6X94oJVorTBkoJcvyOsy5lr1rPit14xXobJ3ovc85vbt7RwjZ7/doRbH3oOSk0c05p5wZY6x3T09Ph80hzas39/f3r6+v+/2+LEuMo5HA8V88PYgFywDgS6AJXxboIZivlFLwehd2LM421mpwCIE14zZjpRYxnXMOZr0xxlqP0sN/YUqqtbbDGGNSSkFK+NpMhj/1anW5XBAIUODUdX05t7vdFqHhdDoRkgkhdrTOe3e5MDLhns65i/OZEk6VtTb6SZYSkCDOJ/b+QvBPT88AbcZhuFzIbncENpUzfXh8/PrrH1x/cQkhHg4HSunDw8P5fL6/v99ud9vdoIvyS3Ac+RW7KXh4KOFI/EVVBkK0VpTxmKIPyRQl4/zdzVEzZoyWQpdluv/0FPc5RcaFOhyOM/8krVbrGOPpdBZ8GpzisvR9PwxD07ajc3VdILwiNmG1ys2CaPjWZnZ5Wq1WjLEPHz6Ai4FasijM5XyC5S3+LUIeYaxer0DPwzwD1Pil0cRtwpOJQJnnH7zOjD0qzhmlJBOy6KYslw5XDz/z1GEyJuO/v4SVZ25rmlcllmEh/sIy6I4xJjopMwqlRutCTFtjhNJ4X4RLPMmccx+CCz7GSBhNKbngwxg455wJpRSgV7RQ/AtJomVwiN1yYwxk7a21ZWWwJimEyGTS/AdpHstJ+F8YP1SVYowytqOUGEO9z33ff//99zlHQvPt7e16XZ9OJ2enGcz793dFUSgtc9aMT7QfbGLijhtj9vt9jJM7EBzjl+JsScnoIUB5iiFywZVWWFf01lFKlJJCSfHFTgMqoUVQDl2+mNfopJRpHiGgwmaMgVUfY8AZITPYO89IponRvLc4TeyF4IQQyrLW2rnxD0sM/z/poSrrvhu6YZRan67Xm3fvfcrNMBRlmTKVSg/W+uApZV3fFWVxvp7d7MmHZyXnXJYVpSylLIQkhEqpUoKGrcc1gjir1hp5SCmdUooxV6WhlAgjwYxe1VvUkrh2Ukog7Og0AV6j6WOMSanW65Vz9v37u8fHz9YOnFMh+OHm8PHTp+wc5+JwPFJKr5fu02/ut/9zi4OB1/+90mk2iGezomoISauSEMGoHPqWlEJP4k4skhRT8iEQkpVUSkrO2M3tLZQ7J2zaOWweYe3uy6YSYasqquCCHawd7PPjMzoYkn7nQDvOLhoxxjSvwNR1zbAS2Xfr9doo/afffvt//s//QVWC+ktKKbVgnD6/PK3WtXNjXVdaCuesYNQ5//ryEr3POYWQQspSKzK73W23WztMGPp6s3LRxxhloU1Vnc6npusTpb/5eF8Y+eMflyQzH6IP4eHz591hV67WXdt9+MG3hKmn59NqvcuZD6NnnHz46p3Wuq7L52fuY2JcEZaV0s75cXBSFZvteuiHcr0ui0JpraSMMV4vbdd210tTGFOYutBSKZkTUYa8nc+b3X50NoQgpbq7e3+9XB4fHu/ev2eUfb5/0IZPDBxtuBB56C/NdU7kySiVGU+Eh4gUQlKMIcTSlIUuKPg2Ih8Pm7pc911XVdMe8mpVrtfrFGM/DD7Gm5sbFwLlHPWLMbqu6/v7ewzzsGDf970ULM3W0LiPqPSVks6N2NUaRwewgTGVSM4pMUpH67kQhDFGcYMC51wZkwlJOYeUckpKciG1taPWEoULY4JRDhAVjzSZBeAIIUrxlNk4DkLSlKn3ibKUIiEpBecZodjWFIyTlKIPmdCpsciUc+58kCo550kijLAcs+DCZ88pL3RRFKVUMpHpcRJKkpyV0YyxGNM4DlhxjTE0TQsNCefsw8Pjfr+F/oC1LqVUss1+d6yqtffudH7jXCRCmnlbLQTCeWaMMkaGIXrvYvSE5Bhj17U3xxtCuHNRKbA8u64bhOCMMqWNmRUVQetCqkDihN3WMAzQMkKrwRgzRhGSjVEpRaWkUpOwoNZq2vBVyocgmSCU5hnpRQWGI885z5kGH6NIQgghZIqZEMaZQMmInh7DrbIsX15fKKOZ5kiiC04KKZVi4xhCZJQLrpQ0JPc5Me9izmG9pkpL7z3jjDKijaaMiNl66L+WHqr66elZmuLm9t33n+5fT6d26DNljEshBKHUmGK0VinVdA1KEmwFL6XTjK0LIYAjKyEk9s5CcCj5CSHYdSpmgdy2bZeeN8WcUwo+YKcDfQbaqOXhRj8OyJLA90oJQvJqVa/XqyWprlY1F4KQnFIyRhWmcC7EEEUpL5cLZMuWJIy673g8xhjhA3M8HufJoV+vN+v1+nK+CCGV1IQQSpnzIeYklCCUKKlXdT0OA1TVcJuxnIVK01q73+8h1Iy1fjyFq9WqLmrvfF3VzrqhH1JKpCB1VYeUrLWXy+Xx8RFlr5SyMIZlgt4WIi2Yf3LKFvgSRQRADCmF0DLEMNqh73speTD6/v6TC5M6tDFmWdQc3LjXRz57Tx6Px7e3Nwvn4RxDCpQxypk2hjB2ubb/+P/8fOy7TPjd3d1qs8402+COt8dqtebSFPV6sxnq1ZZkMY6BkGy02m5XVbUSQhyPN0JKQoSQ3BhvLTS7dL1ap0y4lCFn7EN57y/n89gPQzdILrmQlAtdVDHGkPPT68u3/+1Hr+e3l8cXKdRf/dVfGW3u7+8pmWS3tam7rh+GMeUcvT9dLj7FQFK0o/euuL1NlA3ena/Xy+XSNI1Wal0Xu92hLOqUXQjWun6zWVOalBac87JUxojTSeUcuq6zznEhyLzCeT6f8yzPCT1B3JE8axOl2UwGj9x0ZAQrykJpheTBuZDS4G7GlKgQmZCcWc4s5xyDD8HRmQxNUPInCMOVxijGOBaJKSFCzHykEOgXWrM5J6kk58x7yzms44WUIrLsbFBS2nHUSl0vF87YOIyYSYQQCaFcCO99JjSESCmTAoKAQgjhnZdGMsZWm7XWWhszA5KT7jcnzDl3vTYA1pYhPyEEe0ivr2fIWlR14EwyqoqylFp2fT9Y2w99SIlxHlAtpYxiNITkfbheL5fLJaUohLi9fbfZbAUX69VaCIVlo/P5vFqtpNTE+5TIarUCyIEajhCCaINxMZ2XnOf/m6XcEAI7aDKOI2M0hJBiZJR55xllWemUcyQkUVqXJTBbTB0IIejYgk9UcJR5jPKUck5ZCBny5MhJZ/2Fvu+dty74oii0UdgJIzFQzihlfDaPk1KGEGOMzvlMEuOMZ05oiilKJYQQhP4xtqahivHheIPKGtN5a60RAqunQIeANmDYwGZZFYRppOgQIga/i9yeUqosDRAbNFbn85nMSBTGPsAu8Gpm1lLls0gqIB20DkgbOH4zL4V6H7U2XTegZVFKh5Cen59zJsjGy3gAu9DPz89Lo73gNjiueLvb21u0xojLy55633d4pAiZHFxDCCTnnKucEmFcCoH5mJ2VU5fen8zmwEClIRfRj8Pr6UQoWW83Xdc770KKVExrkIgsoHgzxqpyEtymlFrn+mHY7fecczpLNSyTLvCmkOFwGa/X6+Pj53c3NynFTCcwF8tTKDalkqv1Oqb0Yu3Ly8u7m9vlM6SIZRyhlRkHK4QY+rGTcuz78/mslG6a9rvvvivL8k/+5E/u7u6u1xYfZhnYrFZrQvMy2FwmMWQuojfb9X6/IyTDvBYTJjRYMca6rj9+/IhvpNVEBMBEF/uMKWUkeK01nKxyzuv12hgdghvHkQmesLvuJuIHLtTb25sxBr7wVVUxLjmXgGsFV95fHx4etNaSG4wWY0yXywWbm23bz5NAs9D2kJ4xJMPFz7NgDiVkwXYW5F3M2QVPCwoXMtOR6UzhYzPJm1IWY7IWiqcwsmfGaGxycI7213POy2JS3o0xNk2jZ3VoSikhTEpljLbWEUIh++1cIMQXhT4ejx8/flwqLXwpYGIYCz08PAAJSLOPMdpioMd53tZeBhV4DhljXT+iEiezrhxC53QW+r7v+91us9vtYvIpJUa5LpQ2pTZiGG/PlwviiWBQYZg29TByeHh4wAQbnqkII1prKSdhHsRcNksejeOIRmGZOszz1Bxm9xTEjWEY8FgCXUDEA3aHWISRp5pFo5HXtdaMc+e9854wlgkZraWEKaEzpT4GFgIhOcbE2CSohccJdM2maQpT5DE7C8UOmWJmgikplVT41koJzjmlGOTwnJLgnFEaoqvLigv0eX+MtbjtdgtuOJn5UngcETsWeQl4JJnZrUHNgsNCCGj+5BwpJZMqAKfWDiE4DL3xhMEKgnN+c3ODoAZ4J+eMvLIwt+isc0IIAWy39MjgETnnDoeDlMIUZhzH//iP/+i63phJSenl5SWmCN5RWZYhpM1mc3Nzc3Nz03UdaAlYh5azASy+9YLJ4mw37cVaZ52N0WdCnB+L0kBT5HA4kJQ5nXjWJBMuRN/3T09PGGlgFopjiaf28fER4b5pmru7uxDj+XLGDJ9QorVGPQjoLM96KRNuq5Qfp5FXTJN/5NPTk+T89uZ2gXdx0fi8ZwdmVN/3kPW2dpTa4DMcDoeFtYU9KdzHOFO223x1zi0UdSEkY5wxfjqdm8t5u1n99rcfQ4hQeWKMjaMty6pte4jYlGW53W4RIxC/yGyQmyf/d1KVNWPMGF3VBZOqbZu+740xMUwUnf1+j960ruumabQSSDDPz8/OuZ///OfH43G325ZlaQodopdSCMlzTlW9rSqDo8uksM6leQ+AxknsPee83++rqsJK/LVprA3jOMaYY/TX5vLy/Pbtt3+KRwVN4cvLC9Q7sBSWc0Yzh4Y4zLJ3CByg9hVF0fd98HYZNSHOstm+DU871j/xX1AbcIvZLLOzzBoIoYRQIabdN0IostFyZnEk5byWj14EbB89222No82Z9P2wlHeEEKyO4tDhrQGM4A+X8IoDy2c5BzZb4yEI0C/YOGI2LcfADwtG6BuWd8Ffs9YOQ39zczDGUKZzjkKSsqwo59rQpq1TzvjwhJDT6XS9XlFZbjab5+dnrKos001kYhBAVqsqxoByHnvIUkLwmIMMkuZVxyUOgFESZ/MVjD0ADOR59zDNdE0MjZAF8awWRQHoD4ZKmEKnnLtuGK3FoeBiGvMsFLIFpZ/ry8AkCz5qbdZrY4ML3gshVuuKMU4pidHHFGLyOTNMFpGVKaGEkpxyIinF9EdID2gDkc9noQ8mhPjw4QNAefQ+mJ3i0iilYJeBFGKMSTkOo08pFUzH5DNJlNGUJ6JqWZZYjgWcBz1kpFyUsXVdY88FbKLJFjjnYRjAFFqm3KhMURBxLlLMhNDz6Rx83G7Kx6fHnEnf96t6jUf8cDg4F8qyvF6vMd5hQRr8KER2dN9L5FrUFvFYY0I4jAOEulBZ/PibH2/W6+vlyijTSj2nTAkBGcnOzk0Lr3G73XLOwTGF4hh+3t19tdvvu7bdbDbO+5xS23Wvb29VWc6D94y3w2/BYi6Kou06IQQoE260SzGOHL8wIlDpQC/9cund7CYGBggkyjHJ78eBzFNZhFRMFAkhzmFXVkipt1uwQbwSlBD2m998fHs9//o3v66relVv/vVf/22zO7Rd98tf/vLu5hZSd0KIX/7yF5SRP/3Tb6y1lDFr7fPz8353NEZyyuq6KksltUqEDmMXIhVCkOC999vtFv3Hp0+fjDGXy9no/wEE77vvvlvE0r23fZ/GsTdGayM5p1QLpQQezhgjCd55PwxDtV4ZY4yQ4KjE2ZulrmshWFkXH3/7Gf3WOI5t14QQ26YPLuKRQOU0DEPbts66Zc8RsyJkhaIqc87/yasLAYvMA3+EOTYr5YnZmg3bc2xeqcWL89kN1DmXUsyZpoSoSoVgzvmUiHO+LMuUJg4esFOlVNtNwsnohoFJ5pz7bsiE9F2nlA4xxJBypjlRmOKJWRIGbDoyS7cuM21oliC94SkN87Ixdrvw15BOltIYk17UDRA/prNMJ5Rim+aKd1eaE5KkypwzLAQSminNgHTOb2+Q5+n7/vX19Xg8QoyPMQaXdSSeiYvMaVEazplS6rcfP47WEyooIeg5oM+PUICP6tykpQaFR+ccrB4JyQvHCfduIZXgpOCLIC5xKXwML6+v2CahnLng+3Gw3g3DwATXWgsphnGMOVFCwzDo2WVyEdVfrSpvfdu06w/rqijLonp6fKpmJi4yt9ZqHKf5P805hZhzJikRkplgjBD5R5k9AN6ZRqApoR6EciHQt2VYigcX2DqaYjC3+KTSHsS0/OwIISAwwKp7kcHBuGLpPPLsmhlCwMdApwYZAwRrAOUQaieEAPjC8YsxcS4YFZf2QghVSguuUiRt2727vWOMQSwFTzAC0zL9B7sRY3M0DWk27ZNS7na71WqllGyahnMqBMs5xei9t1KJu7u725vb8+k0dFNEHvr+9eXFh0AIwflPs0AmLhcIMH42/kVRORVlnMdZtbRt25zSEqDB7sXaTnDT7hsIxCmnru9LY/7hH/4BLEzUO0VRhBAIpX62nJNcdF0TQkhpKohubm7++q//+v7+HmRTrbXRGiUnIiZOtdFF17fDMBhTcs6qCgYmqVxvUnB912ullVRggnvvv/3RnzFKHx4enj4/5JzP57MQou/7YewozVoXd+/fD4M1VYlCWAjKOeNCU0YTISH486W/vbn77t++s9a+e/cOzyeYJC8vzyg/EYN++MMfjuNYFAXjvCgKbLQeDgeMTzhnnJvn5+eu66r1apwVIcuyzCEuORKFthCCMbJaGa1lF0bOKSEkJ0oIG0fbNg3eEY9NCAEUcM4FMnHOGWs03vu1lCgyFgwNOBL2XfCE59mkExw5PIRgACJDL+wmPPALzkZnrVA2e0/O7cW0ug/acZ7XdJcaaCERIdyP1gJYw7Y8rFPSbOFlrf38+TMOMqIwm90Yxexpg9nhEhBxGBExixCAGiHn4XPiAuJ+4aWA+DPGsIestQ4BoFxOKWktJy4opZlkzhnjLAQfrEfkAcdvsXjJs6YpyEViliytVxVjrF7VhJDtdnM6vaU0KqU4E13XYasA6TylNI6262DsQyFbgFOslLJ2xMkCzR3vW1VVWZbomcw8a1lm2plkLrjzLpNJC8vNou4AmRGUYoycTug3KJS471prwzQGadfrdbEsZIwKAWUEEKUE58zaYEerhYopYomQU8WEgOj7fzU9AI+T8w43hATIXI3yL/Y85wEaX7DjOJsxpRSdG4UoMaEhhKQUGZvYeLgciLzgfS4LEFiCW5YncKVQX+CVoZm83+9Rzjw8PNhZSBl0CEyflFLAY1JKjDJspTnnPn782DSTVO+y54JnHUkbNER8KnwAYAWowe/v75Ey65qBxzk6K+fFOq2UHe3hsH8Koe076xy+KdhQSDz4NcAWHJgQwvV6HawLIfoY6vVKKDkMA2HUB09mcBZ7Ekhs4zj2TQuOY2EMHF5Bon15ebm7u8NpwXtZaw/HY9M0zjljjBISDz2wjqVigkPnbreTWu32+/PlggkQsvjb80vTtCkHHKTD4cA5RzOXUqKUhRD7foCc0ThOkMhXX3313Xff9U277DGFGBnj1tq27au6vl5brtAPBey3kkxCdFSInCchlu+++w7rF9BIX61WYd41Qwr/8z//c0LIP/7jPyqlCqUOh93p9BaCA1CvlJCSwyYa37RgTCnV9F0IodIGKQQBOqXEOaGUxDhhHXjg67q0bnDWYV4qhHh+fsYVQIhEbEUYxUJ1nCX+0+w0iWZ3GIauveKpg+YYmJSQixCzfxEyDZ/V/FFfLIZ1hwAAIABJREFU59kpgTHGueCME0op5d6FtumFEKaQIUwHFudoHEesU5BZU5pOS9eLpWU+nS6Hg+DcAefhnDMmENYBpr1//x5ZAZq+YpKwlmxe7qPzbvaCReP7vr6+Yo+naRrAMjiGdhZ4xsVH2mCzpXzTNOM4gBvCPfV+sk4TSuaMuXH0wfb9oOW0z4h4jU8CLJrNXPYJg52WjWRZGu/jel1vt5umuTbNVStzvV6hPI8Tcb1eEUawNoSxOe5vVVVNc0WE5LPbI2YtbDZrAh8kLRThnDbbzSQiGwMTXBcmEZpSct5Z6+r1KudMUxRCdE0LaH3BPC+XS9d1+902F4XknGUSfdBCCsYTne6p1hPHiVJKKWGT5VJinAzDQClEXP4Y4NJut4NSxevrq3Nuvd5WOaHWAMcGJTyZadriC3ds/FZKGRN1VwfUm03s/sRmSRkccq01grtSCik0fPGztJ/ffvstwhxy+3Kn8QrPz89kXjzGL4DqomqOMcLmGggpwHfok2CIQuedUpQGkD3AugBuMyTj8Qm7rnFuNMZIyYvCeG8ZI8ZoY8xut7WjHYdxzKPWhhDCuUhpBFCONTTsRZdlCQz97u4Oc+Pr9dp2XVHWSpOYpgHPUjjM8ZeiWEYoqcqyLsphHECsGq0llFjnYkxL+4VzODG+GAspttcm5wyTkBSC1gorKTFG/AIJeByG3/z6Ny8vLzh1aGsmjl1lMBZelshyziGkQktItrGZGI5+CO3Ikw+Hw2G9Xv/6179++PzZFGq/355Oz8/Pz9vtHnoJe37IMSldhuilYilP5cjpdEFTj4qBc344HA6HQ9c2gJtSShicvL29ff78+f3794fD/uPHj8Cjq6ryPmHMiMa3aRoz7xCFENRKIRwDJu373vuNEEwIorTQWp1PDaV8vz8e27asdIp+vV4vzqPIgmVVUUIBLyCkgjpxPp/RE7DZM4POdggAu5cp0RJbAQct7eacsfiXs4Q8U/iw35Bn/x88KoznYRi6bkAZh76t7/uUg1JqsdZgs2jrMAyodpdaDaX39Xp9fHxE0XY+nzFwRouA0wREl8yiYYhT+JDglaSU2qHHvwKAiSt2uVw4YWz2mkYIttaihgN2xxj99OnT9Xo5HPeCa84FY4GKUUh5vTbDOJjCUDr1UnkWvUA/BDCAzBJtMzQ6KVwxRrWGELcaR5lSRJlfVdVmswEIeblc2rZ9fT2VZfnu3TtAF2RWNFooZ8u9W0oBXEw+b7G0bWuqknHO5o228/kMFFfPWg+MT9yEGGOmBDkYXdSHDx9ijOv1uipNXVWCD4xyrY0QilFGJ3MIS2lmjEjJ8ZrGKMOLqqrO5zOhCT0upZTzP3Bt+vfSQzc4qUvOhHcphfzV3fvz9dI17eFwNMYMw0ApJyRtt3trLQgLuP3oJBbYJEZCqTC66vsu+BBDYIxF6rRRztuUY9s1UomKlcPYC8mcG4UEK0wQmoRko53WQXGKwPMB0kdojCn44M7nUwgRDuarulofyqcHJxg57rda67DbSCmklNb667UxBt4SOaWgtUw0U8ETzYRTVegYI+GMckYn0RIOoVlorLZtezqdQkh9P+ZMN5ttSvl6bY/H47ra5ECUUEPqheCE5cEN49j37ZVLeXNzc7qcu35cb/J2tVrVu5y4d1kKQzL/8OG9lE+McpLzOIyVKYJ125ub0+tbZQqAmCiLAMdZa6FEvSorrTRN2UhVGtN2HYmJz9qIi5QsDt44DEhFzaXdbfZSmiGMXJoQ0g9+8K21cbPZK1UMwxBCjoG+vb1cL9f1es2paJrOGHP7/g5khMH6olrFTOuqOl+vkWRQvJXSu912GBtKEyHJGJNDbM4XRIqmaUZnX09v3TBQwS9Nv97uCVOmrKty9fDw8Pj0VK3rP1P/re+7sioIIZ8/P55OF04JGurFgmm73Z5Op8IYzuR6VYZ5Z9joclVvlFYhBkr5er0py+p0eisKXVamt/50aQ6Hg1aFEpok0l3aSpd+5WMO1o8h+ZjD6XQqKlNVlZEFEVIUYcVWMQRHxtuvbzllwTouBKFcSB1jzIRRJqpKSsZOYRyGtlqtCMtMMGX0OFrn/DCMhJCqKmL0jDFrh5gSobwo6/1+Dygj50yZiMEJzlDCOzeO48hYyVhKKToXhJCEBCG4c4FzkklKJKOeAFGHEDI6W9dlP/hr0zGuuVDOOUJi23ecU8qJ9aMy0lrbjx0hZFWvUXzA5A6QNzacrXUo27U2ORMo7adMUibeeWNMJrRtWprparV6e3ujYhKwo4S66Oq6fnl5KYqibbrrpbler5Sw29vbru1JpoSzTGlIiVJarVZSqbbvaQhlWcAQImfSNF0I6fX17auvfjA6dzweX+9f/1/W3mxJjuy4Ft3zjikjx6pCFdDdaJEmSocS75PM9N3Xzh/oUSYz6spoOk2R7AFAoYacImPe431YGdGQ3g5JmKytqQYKmRF7uy9fvnz5brcbOzP2Iw0s+BDlz8PbfrLdzPI0kgDC0AcXoh97XIfRGA77bkqjUiJGn6ZpW19ubm7Wy1IpGZw8nZrD60vbtsfD8RBj39ZJkghGTHDN5fz8WTnrCWFdNxSLpXH28/ML53y32cCPD3PU6B69vr4yJvq+L5dLGpgWMjoSHRnaUam43qyHvrfD6I0VQpTFwgzjYrE4n8+g39E4udZqWgfCuq67HA/oxSqtjBmJm8UddGbF01RSFnUiuOSExUgITH3/CumhaWA/4o/HI2PMGFMWi7q6XC4XWIXMvVYhhFISxvQzooFCYBxHrB93zo2jQecQGzqLfNG2bd8NzvoYCKOcUhbj1SrAWmOtISQ2TQ1F2m63++6779A3Rm+D0AitdIy063vOhNZXj6pZ9QEwArQyDEZKTSdHP/QVh2E41xedaCHlQkoIdYZxyPLsdDqhmxdCKMsS7Ws/zbiiFgZiUkoxxoP3z/u9lMKY0ZiRUtoPfdc2YAQYo4vFom36xbJ8+9U7zVPvw9PTc5al33zz/v7hDgaup+PRe39/f08pNeNIp8EFDFejRsbHsNZaY9tLrZRarVbr9Zpz7q0zOmHiuqoFjSIUczCMAgcipVytVhCPSakvl+b777/33jdNq5S21p3P1XK53Kw2gvHFYjGa/nK53N3dSClXqxKLddM0M2Y8nc5d11ln69AMtE8zLSUoVJulRVEsdrtdWZb39/ePj4/guMAR9YO91F1RMKUjCLGXl5fvfv/d+1/8TT8MQoj1el2W5fff/4hiTk0mhtDUobSCsYH9YgUTnFS01kqqsiwXixKrcdbr9du3by/NHyGJLsvy/fv3KB36DrtrGBZ8AsI75z4/P3+VvGdXxdqQpPoqZiPX9XbTPuGrpfEiTwlxouajNUJw7z3h3MWfdZyzgIcLKhUPQcfJVwO2TvOAtLXW+SsJCYJiDhDeu4kEY0pJ565lA7KLmUyWhFCEOEQldARB19zcblEoo1WAcxJJYIzNbjF6Mp303lfVhU6ex3RaLoQyVCq5XC4F530/HA/H3W7ngkuLFE1EnDrcweCvc9HzANBqtQLrBa716ekJJxPdmhgjGoRkWov2+PhCKU/z4vHxEXoWiAVCiNBWzPor3H0xeeOzaSUfmeYY+r6TkhNCAGTRCITKA5ENKLssy7dv3z49PeV5AYzlpy14wzB8fvzMuWCC90NvndOJ7oY+hMBIhFPThw8fqqqilM6M03K5TNOUpOnz8zNnbFmWiGaMMiXVolh457u22213NBLo4Mk0NDN3jMhkRKYnr6o0TZWSkESBpg7TjHe0jnOqteJSGEy6kDg/jb8oPeBFzrpsrTXkgwg0yBl4wZRSxniMBpQim0yGxbSPZW644aTOUx6Qb8IvhTHGKKOUD0M/DIO13vv4+nqIMVLK4Qn6+Pj47bffYq6EMSYkR6QeR5OmSfAEhbmUohuGfhwDIaO1gzFVXXvv8zwDIwyVlJ6sn/quE5PrgGDcGRucF4xXfR1nifp0LGa2FLzK3Gy8XC7/8R//wRhbLstxHJS6asOddWVZ+kiU0lnmy7L89a9//f79+x//+BOYouWyBNWDv2K9XmMIAzFuHrxAw2AcRxgzoJNR1/V2u9VpwqQYnTXeCa2SPFsul0maHo/HWXeEsQYyDeMQQqYxdYUIhc7qv/zLv6CF45wry5JETinZbrfGDrjGXdeNY69UBn3w4+MnfFRCoqc+TfR17klKQmK5LK1xnz9/TtO0LMtvv/32w4cPiKqLxSLJir/5m78pyxIwWWu9Wq0SnXzz1bvPHz/sdjtByR9//11XX4j3nNCvv/4aLGWcZMH7/d5OckNrLRwLkDAYY9ZZxmjfD9ba0+m0222kJJvN5s2bNyAbQXmjd8WZCD6QyKzxjArvYgyUeHJ43RNCsIcOIM45F3yQiRRC5LnCbYQZXJqlZuxATCdJerlcRufSNMMpyvMca9HotFrZ+0AIQe/HOQfCE2oLa22Ifr4+oFUpBXLCWDRG2rl1QwjXiS1UmTO3M7cJ9bQ60FobQ8RN4cz3/TgONs9zwa+6AzCxM1cDvhG0DPDs+XzO8/zd119X9YXEuFqt+q6TUgat27alnNGBMyYYYxh2USoJ4bqaCf2MZHKuRGMMGCVOWljASmd/9m5hk8wphDiO4/Pz883Njdb6eDzOfUE0eMO0lwVOJ0iZuIZ4PogPzpkYidZ6HA1UIVJqrVPir+w0nwZ00Mt0zuNGAP7iMp7binMxmLFcLs+X6uuvvzajqZt67DpIMauqGobh7u4OXBYwLnqc2+0WdPr8fpHMZglomEaj0EsH34uMO/syARsBDZDJD3tm5pGWyixHapmlpyRGN5nd/UXpAckKGBk7uOcoD4KCTUOeIEjDtBtkvkW4e3Pzaj58hJCiKK21jAlCmLVXOT+ldn6Rs3gAzaWnz88vLy/jOH78+BGoBzju8fGxrhshpHc+RiLE1UgLDS6cGzpNomVZ7ibv2Lk7gsOEQ2mtxUo/kL9KCe9tCEg5fBz719fn0+kUI91utzCDQp8NABZ2j8fj8Xg8KCUJid47yliSpE3XMcbevn3LKL+5udE6AU4BVoWAgU0eLGj5AhOh84kTAOIOHxLMZtd1b968AVJAS2qunMAmxWmlQVmW86YBBI6+79G2wYvDc2vbFqeTwsLEhDRNsixLiQacyfN8HHtoTvK8AM1Kp0ldYwzjhFwn7yyjjNIAE5EYI9hbQggm1+BSUJbl5XKZtWSo53DtEVPyPD8ej2mqt9stPnNRFKfT6c2bN9jqcz6fAdBmmOycG4cxBB8jsXY4n8+vr69fffU2RoJyEO8LJwHS5Kenp5ubm5nZ/1nLG37eTT1f5uADvbqbXK2B1OTgj+K4bjtMovFwHRtCjNBaMUZCCP3Q13WdpSVaOLPy4trhNBhcCom1kZJAonMW/wmhZy6RKaXeByEUginmDNBwguIZTQW8aD2Zm+K6YU8JwlA37XInk05/bilhwBDBa5YsIkRonTDG3HTr0Qyd/ysiFxjFPCuenp42m812u31+fr67uwNwptOYJxKYnI3wMDZECIpCPPy6vuSLEhu5z+czourc/vVf6M7xt4t4tZzy00JQXB/GCETzeIkIa6vVKroIUkgIcTqd8DbX6/XxeEIAQS/zqjkOhFL2etgzzpXV33//PRdXk766ruu6xk9AyYWnxxjDVqgwTeDOTgrYBzVMe0QQOREiMAhMJoNnlBH4aWGy+AVdNKf2cVqrbvsBVM18qhn9uZb6v/313zgpFAqIpzARm5MSmo3eezgmOeeEkHHSiUopMXyIt4Vt9SgswCoABYNKhroDY6WUMmc9Z0IKJbiMgSzL1Xaz40zMKgs4XaNmNMacTqf9fl8Uxe5mJ6VCP6ft2sGMPgYXvFAyyVLKWTf0CMH4g/f39xhFEUIUWc4pc8YKxgXjWqoiywVjIQZCI2VkNIOx4+v+5XX/ApwyuyNAwAN/PWCTT58+IRp67xOdbDeb7XbLKAVcLYqiqqo8zwgh+/3eTrvOyeR+M4vzgPvMFztkQCkURbFer1GeU0oJpVKpRVkyzrkQxWIhpFTTUiZgtNVq9fDwsNvtbm9vt9vtzHQLITCph2eCcAmVIRA9rgdADbTkmNiAnON0Op5Op65r7+/v0eaFgMd7v91uYozH4xGeuzjTqEe11rvdDjr3uq6VUjc3N3/84x8JIYvFoihyEgMlIc8SRqN3ZlkWJHprzLztZMYNYEgwMkYpfXl5QTbd7XbGGjhoId9UVVUUxTC47XaHPDS3+gH/jXGUcikxUjt23fD8/Hq5NK9Pz+2lro4n4gOLBNsx4VE4juP5XM2SEj+Z40qlvHdd10MLgBCA/4oYDclZXddYXcUYOx6PCHN4ekyKQEk3DlzJQAjlnHAsudSMia4b6rrtuoEQ5n1kjM/qvuVyOZezUG2ABMMGQMz6aZ1KqZ0LjAnvYwjEGDcMVzsZ5J48z/u+hyYF038xRkxIYGACDJjWihBiJ0Rorc3zglEOurip26Zu96+HDz99nLtfSGOopUCbIK1SStE5n8cgEAHFtMwqSRLrHKh8FLsQSsw6NPbFL6gMflb3dR2Yj6mMtvv98Xg8Hw6ny6W5XBpjHD7FOI6Xy+V4PIZpDHa9Xq9Wq5ubm81ms1qtYMqU5/lut9tut7e3t2VZMkbBU719+3a1XCJmglOllEKtCwSGK4kcpifvZzI5KoLpAnqDSgrfC0JwFJfAzROlr6YEfwX04I3ptGfTTxZESG+gQOUXG4v/r379z7E4fEpMvc8cKPAIKmjUvFLKONlbAuwjtQAI4DzhooqfzR0lIUxKh2MKYGgtjAGM1tG58zhaFH1p6tM0fXl5SadFWm/evIkxpiaJ0eMp393emdGeTpW1VsjrAhzgF5xpGH0DTWCQjU5euHhhkG/PRw15WE4GG4B4hJAsTznjetqX0vf9+/fvpZRt2+KEcc7TNCFI11LSCVTi5mudYLnbHDHjpKQCzwZU+/z8jMA9a6jmeaI8z7fb7el0wln505/+9Mtf/hKhp65rHLKqqkACEEKQgHHfgBDxlRH0cWMRF5Ae0JjRWt/e3p6OF1AWt3e7pmnatt7v9zF6P5lYMMaM8UVRDEPfhZbxqwwUBZB3kfOrD0xZlmma3t7eoqapqupwqshErfzrv/7rb3/721/+8pe7m93D3RtBKSfUjabMi4NUWvysJkQyQ50qpYzT1rb1eg3pJKKYNR2jxLmr82vXdT/88MP7b78GYwNRStd1iDW3t7dN13kfnPMxkhCiMdYYSwgZh9ZhO1hRBO9RdAbvubjaEgPoEFAxzjuKlmDmQvTe922HpI6sZq0dho4xBol2CKGqKkAxlIPzcImUsuvC+XzGH6SUuhgZuW6yQ3xB6RkjYVzMnwH80gyiEapwqAghUsoQIiFRCDmtuGCEUGtd33eYCsTjQgTo+x7SYTNZhNKrl0yXpukwabSE4GYcrbXBk6qqF4uFc5ZSTgjr+9H7+Pvf/95a+/DwQCZh1dz8w8wXikjke0opjWGxWITJYPw6WDfFJT/t4AI4mAsRfEGc56ZpCI24pLjms34pBsKZ8C7g+xpjLlVdlmWqr07mfFrchh87zw/GaRZaKZUkikS6XC6zPCeMWmuTLC3LskiSYRjO5zP234VpkhGbjnD4MY2LCGCtheAFa4gYY6fTCW9zjmDAFhBT4ZPg/cpp52CMV79nMnXFwEqlaXY9LfHqUhzJ1TzqL00Pm80GUpnVajWOI9I1n/ZDIdbYab0RXjlKbKXU8XgEhY26GzUpEiP0+FKms2gVlRSl1NrrqEEIARYdCKMz9fH+/fvHx0dCCBBfJME5wxgVQugk0TpJU2xvp8bZQGI39IMZT9X5eD4NZkxUwihDTMTXMcYcj8ftal13tTFmsVhoIaPzMURGSJYlWZai3/Dy8sQ5XyzyxWJ5PFSbzeZyuWAnGkpjY2zX9XgZ4zg6Z2F8oJXknINK1EpRwiFWQ7zbbDaQz356/IBASaYeI441pggxyemnISM6yXAJpeVq+bJ//fz89PT01Pf9ZrOx3gkp7bQfablcysntByW8n4bSUR+cTqfZzgQVyQ8//DBTmYSQuq7v3tyACDqdjvTntr/JskwIvlqtztVZSqnYNdNsd+vFonh+3nsf0jT9cuqQUrparW5vb5lQ33zzzXq9Nsb85je/CdOqjCxP0kxzQSOhjBOpuE7kOHYQtgKOARumaRq854xhWhNfEJVTmqY0TavqUtddXdfffPMNIpGP2J+cZVn26dOn3W6HOzzHl0n8h6WVFlw2jgq60EopH3yw0+487zGSeXNz03Y0TaQUKsSAW9oO14VCSKj4FzThvxx5w6J1jD4453wMhFAmeKQkEuJjYIJHTyKjzgbBFVecROZsINELKZPJZnxmZQkhp9Npt9vFL5YBQH3n3DWI8Mnoqa7rvu+z7EqbiMn0IkmSLMuU0npadHo8HkEvPz8/p3lGCWGEXs5nShmyiNbXkUDc7vV6jXfqnDudTlrr7Xb7hz/8AdwXDBfm32wn1y8hBAl+HvuAc/4wDE3TpnmB2EcImTUOQI3w10NQAlQ1dsTaMUIIDjy+Xd8N+I50cpFCkhCMkusC1+skR5gk73h3fd+XZXlN5FIfDse+76VSkcQQo1ZqtVp5YzabzTiO6/Ua0IpOexixXQ7JDHUeOF44+uBDrtfreZIR3xFNKT+5t4Lz9NNWOIQLIdiXhQLYpzg5PnDOSaTM+xjjXy09LBYLxvl2t0uz9PHzZ8E5ZdfnhTYgcByuK8orxB2cOQAr/CjENXz6qqrevXtHSMQlB4KGNjzGqy7WT6Mlb97ceR8AtcApATPi+7/uB611mmZlWdqp3pdS6iRp6sZ5Z8YrKGOUimtdJWcZOPr+5/M50wneBPLZ1CyhWZHgMuP+o2JYLpcvzwfUcbjYIFvIFW11MYY0zaTkGDgoi2K5LBfnRd/3gUQpdN/30PDgrKAn3HXd7e0tlEVxajujGmjbFukBJxV1xpWrnUzczufz8/Oz4GKJWydV3/dockKoM3cU6DQ6C821Umq/34OcSZLkq6++wlmcSRvGyOVyefvufjKi8YfDQcqk73shMKvF7+8f2rZ9Gq0bbSYTQrFXdYghuujH0YyjQSgsy4W11pgxy1JxqUH1wom4WCwSrS/1GU8STiSYSsnzfLvddV17OBzv7984x621XddmWaaTRArBGENmZZOKP89zzmiMtO+NUurh4U2a6XE0PjKE+OVyeblclFaEECElmbr9MUbo7oGBmkvFc75cLpEJ8EaGfoiE0mlLwVzLG2O04oQQZ60LdrVawUDXDtev3zQNBi3wXoJnwNEw+gadwjinlKZp6j2kMjHGSAnzwQficfgRgvFNke+RHshki4QkB7hGJo85FBAwIMDVE5OXftu2aZqAfMDBRi5BMAJNP7lYXyeQ27bVUh2Px+B9cA6LBRln2JOKyhVzQre3t58/fwKmBvX69PQEjh6ScQRxKDISrZVSq3JBpsYyICniHR4UpRSJ59oj1Eopaa0NMYTgCQbNjLEOqh6FUoNNAxbWOFDceDLz7QDLj0A8N6Lnxjj6E3IaiwOY8843TSOkjCRSyvCVnXPI9zB/1VplWeZ9PBwOiU7wZa/74f21EEcGitPScoRWMVmwoAOH/mjTNOiOxMnfsG3b1aqcEYCftj065/IkHY2J3gsl2WQaz/5cYet/+2NCJYOxy9WmujTWh8h4NxgbPGOEc8o5jdGPY991Td+3eZExTowd2q45nY6EUCiCYyRgllAZISg/PDzU9QXbBruuCcFZOxISKCWEBkI8pSHLk812lRdZ016sHSnl6/WWc6l1SgjTOrXWMyqqQ93VvRbaDU5xWaRZkWap0mM/EBeiD9vlOpF6WZSZTpCl8OJ3ux00MEopF4PxzgZ/rM6eRBeDJ9H6MLRje+nsYN3o3ejd6BRXisuHuwfJZNcNRVEKqRmX7TD66IzpmubcdY1zow9uNAPnLERGuWZMHQ6nz5+e9q+vwbrf/X//cTi8Wjs6Z+qmutTnWYMbQoChAhAQDq5zTgimtSyKbBz7vm/xxKTkdhgTqcww0hAl5946O5rqfPTe1nU1DB0hoeua//zP3xkzQH9SFMVud6XgCSFJkuBHheC6rum6piiyEJxSYlVmY9+87p8+fvzpcNp3Q1+syqqugdrqujHGFUU5jpZzlS9WOisIVy7IfmBclJ5ona2SJHcumGEk0ZJoFrlg0S6LLEtUW1fn46sdupvtcpHLLOVv7rY6zazzq7Js66pIk6auN7tb7+OP3/9x7JvPnz6eT/sQ7Lk6fXj8uD/uAyEqSSKlQinKuUoSoZSUCaVis9kUy2K1W6dFTqV2hCsl00xF4l2wgREmpMH+LamlTqVOXQgq1TpLbt7spJbrzRrr7MEDIG91fTc6OzorEx0oUWlSrleDNb0ZB2iDGNVaaa21kFpIrRU2iLVt61yIkSqZbda34E/Qj0FcZowJzokPth+LtNBCS6aIo5leDO3goqU8ZkUqtTDOdsPApWy7Lkw+qQAcYNsQPQFFEU0AdMBic86hqY+THKDvh+fnF+g7YiRaJ3leLBal+GLgH/XZarXilGquOeFDO5jBek+sC0JqpSSNbr0sUi2UYENXR284jXBXc86BRQCmDCGkReqiK5cLIfmb25s8Sfq2ceMINInEAHYhy7JyuQRtAK0jTJlijJfLWUimExGjDdFaNxjbGdsBCGktxrHr+yZGJwTVWkol1GRbMDe3x3EcrXMhUi6YkC5E/NP6ILW+NE3TdUKpbhjOl4sLgSux2q7u7m8ZI2bsSfBacDcMoxlVoiMlgYa0SMp1efdwky2SMs+0kG1dKy5QOA5mJIxa69u2T9NcCOVcQCtICME5dc7MYxkY886yNM8XQqhhMM4F76O13lrfdT1gAcjhcTKH7q2r265ue+sjuD4aWXB/DWErCuS2az9++gSuT8x4AAAgAElEQVRYcT6frTPpskQhj6ocK/SSRGdZCncw5x0lkRDqnINwYG5RoJ7AqPqsjELjCNMrwQf4CS/LZVEUbdMeD0fgx2EYlstlXdfPz88zrLDWdW3ftT0wApv2cQssuJay7/ucsSux4ON2u9WTOVee55fL5ebmBhgHqXgujRlji3wVQojxusQ1RpIkKY5s07aIrWryA+gmUBZjlFJZ47I0Pw2nYRiGYURbCTqZh4eHx8dHIYTzFohynEzZUAhfLhdAdULIdrsFRMX/5NOsCSEE3xcfG9v6oFy01hJybd4CR4PKqKpqu73B8DabnEERKSADh3gGldBEg3S4qM/Pzz4G8Pub3RYCocPhoLW+v7+/XOq+Hxhl0Hd677N8Udd1muZSKiHVarkKJIYQhrH3znZNnepkURRa62HolZJfff2WkNi2baL1alm+/+qrcpG+PumH+1vBGfGu75qPP33Y7XY1ubzuX/OiSLNse7OL8TpUPFeW1tphGKKzjJIkSdIk2R8OOkmo4IMZafSUXm20hZSM89EYxvnQjVprHEVrAQAzlzhOCDaNI1Xj28Xws8vFarXCbVdKKcUJJV3fCyG00vGLX1C4YX8f3FnQ5MQnAa4n02pCLa96WUYoF2r0Y/TQfZJIAt6pc/UwjIRgteJ1AfVc0EO8hGsiJudqdMUB0aCkQpiWUoL4BrTErYR97My8QxUCEc5qtRqG0RqHlBNCgAAUhSmcbjnn6/UKoBBzMA8PD3HyYJ7vDoReGIa/v79/c3vHGFNKiWkNDh44isjRXHuTeONwIokxxuA5p4QEY0a0cyBCUUoJwfu+a5oa2iEwalongl87o6gD0B8F/zmTV7MSJ01T0NGgy8ARUUqzLJtHo/FPa23btXXTgBxOU53nWZ5nwQeuZfAeK2YhURHq6lcEQQq6gPiDu90mRj+Lemft3OVSm2k77Ew/cM7btkPxN4uUEGljtF3fM8YWtGSTy68Q/y3O/5npYabR8fkmuZXEUJL3HmwPGDGUM9ZapRISCZyorbVt23jv0O9C/AJtYqZlQXjZMLFqmgYzBugGA0TjvbZtD2LUTx7gy+XSOaez1MXwh+//tNls2qFnjAVK7LRQN8aI5gcYAK3VYrEoyxJ3YLfb/fDDD3jEy+WSEIJNO/jWQog+T30ITdclU0s8yTJMeHrnIN9er9dSypubmySRbhwhsoSaDcQCfGGh+cGOUoiv8jyfDxYqSpSQHz58wJFF3QBjHwQmPC47mUGCDtY6nXXN6CjGGBmjZVmqyaUVP7xt2+PxCHKTEIIyGVcUumRk3LkM//z5sxsNljVB0t40TVbkUkro8IDpMI+CnX0uBjcaY0xeLo7VGUss7Gh8jNY6653Wad/3P3163Gw2Sqt3794dDq/n83m9WXLOrLUh+K6uikyvy8XdzaZI1Ldfvav7gQsFNuDl5WW13eDGJkkCCIyGIVrNUARlWhd5miQJoRSBMpFJXdf5bpvned+PSimhf5ain48Vopv7Ym/oMAx4F3hZwDTOOc6Z4izGqAQXQrAYKE2EEJRFKYUxlnNhrZPSz9o/N+3W9dOCcTr5P8Zpxyf6cLh36McC7kAygLtNpomiGCOYk5lWYpMX2XUSDTpjxvCb0RvY7/dAM2A/wmSQhxAMdutyucwCv5k3wyeEnIEQMgwjMApyBqJ8mqaUXqcl4EsBY5uHhwedZq+vr/ixyFXoiqVFKiY3SS0kjURKyQVvx8GTSDgLjshEcyWDtSrReBSghkBBN01j7HBzc6N1Oo6j9xH1mfdRCBUCSZKk78fLpSnLEgMTMWB/5XUlMO4IEvP8RvDvCG5CiLu7u8vlcjqdcJfxHmcBOqZ0Udb89NNPIACTVAfnlZCcMi54MFcrjpeXl2JZEkKuPQyhEQyhSA5XQ5eFc2ZWo2DWAUp6nJ9ZmBCuqvTru57vApJcCFeZMt4+0EP8q4zF1U0lpWSc+GCVzgkNkXjGBOIIgANy6Xz0Ec0ZY2macS6ggTFmxJGqqmoCWUoIBgEy51RKjk3rzjmlkiRJY4xluRJCEMJQN+ML/9u//Ru6Ty8vL/AhgfwLS04IIcvl0nvfNDUhpCxLUPlIzuiI4qNut1uEFfLFNlcoFjDSgfvpnIN/73yfJ9s+1vdt1zWc00WZL8q85AshWHU8QWtojFsul3Xdeg/fDg3/ODyQ2QwKMjuECfwngEo8/zDNqeJEpqmOMRpjttstlkMIIaDlxS9CCGBpCAHztGhknc9ntKnX6/U81gBOA4I0HDL0u/Ab/LwmhfGiKAZrzudzkqUhhJu72zRJL6fLLE+YcStAK5kqKiSbtm0l5YR4F6KS2oUQCWuaJgaSFiWq0rIsX15etFaQxLRtozgNdkyEFIyuVnnTNe+/+UXbjkmS/PTTT7e3tw8PD//1hz/UdZ3ohIRrS+Z4PN7f38+GnXZynGaM/e53v/uH/+c3uMCovtMsU2mWJhljDI5JoFA4p0ifSKvj5NqPY4ZQC30qmVx4ZzM4ygLnbBiMlMyFYIwJ4bpxAaRKmGbWZnNTQGmAHqicAdvxRyBUm9valFJK2fwGOb9akM5UNYI4LiNgED4qMgpAMfIHzgZy3pfqJuAqjI6ilmWMLRaLT58+4WRO3e8gJNNaG1vg7yU0cEGhz8GvGKOUElNE+DxIWhDR0esmZEEnP3AaYsDWQnLtlxBCcPumAW8PaWnf97MrlJnW86EaRjXDGAO0h6QQyojJvTEqpbuuQ0FQFAVuB14ESqUw7aAETIQQAwqLqqpeX1///u//HueqLEuQdXOPBCTH+XyWSqTTwg9KaXWq8U3P57MnEeL+oijKsvDePz4+1nVFKUUZFKfhNWTTyZ1w2O8PQsjVajXVFhIEtZQCO1Hm+g+PwrkAZgKPBZ/5z8sN/zM9oBqFYAPCSvQzkR4gDsGI1iwai5FM/RxRFCrGKKVo2wa1G/4UUutyVc5QlzJCaCQ06kQLLu/v70+nEx7xOFlXAtGjk/PlFbXeDWbUadIN/fZmBwM4+M+gC41rv9vtcPqHYfjxxx9/9atf6WmOEXIRvAlo/1EtSSnrtiGMJlk636hISd02mUw/fPy4WJavr6/hxx/fffVVXuSXqnp9fcV4Ea4rjPIvlxqXGQgUYHz+615eXhBbrbUvLy9QOwC8o0rAZ27bVikxT0jCxAYG1IQwgA4UHHmeW2spJcZcNxjjeEGs3fcjHBUhTISOllIKP7IQwuPjIzRRCB93uxvgFMQgYPOb2xtvfF3XczPNT6sLyDSRRxmjjOkkGcYxUiY4/+mnDx8ev0oUv1wuaZZbH5JIFotF37er1epwfA0hPD09rdfrur5sluWshF4tVx8+PRljkRcJpQDawfuPHz8uioUzPk1TlPygMa21fQhS5G3bos5DSxnnDa1IvHTI1ZAegKZnKI23kKzXuGA4FaBQ0jSRUtDrDF0Uk0RYKxXi1anUWSfFz+JpJCqEV/A5OAZo/6IoBGhF3jVfTMNCxAEpNqVkbjKjjyoEnyHh3KYGosd5MJNx3vl8Bmb3k1YbfwShEzJWjEogCqO+vNbNSVJVFSwIJ5Ercc6maUKvlv4+xutIF34ycEM6+R4CviAVrVar4/EI3IbosVqt2ktNuWCMCSWJ5E3TJFk6WjOOY6REpwk1RgiBl3U6naC50FqHGIxxSiVSEueCc8F7n2UFY0II9ac//eB9vLu7X6/XSiUxRiiXgBWASqHuK8sS8yvwm8EbwQAdnkCapufz+ebmBukWmW+uuYdhWK1W6+VScu6tdaORjHNC+2Ho++F4qFCrmXEsYowxQoeN5TwIpLNyZBgGQiLuJpw5cOZfX19CuCZ4/DZEqhD8NKmaz2IENKuXy6WdNt/hy86dqr8oPaCraYzZbDZScvxP7y3gP8rD29tbIFytdQguSRIhZAjXGfeyLL23GH6BkgHOKpjQIeTqK6Im02Ml5f/6X//4i1/88re//a330fvY9yNjIkajpt1VSZKAOs+yzFiLSJdl2Tz6fzweAQFQNWPySF7NeK+CDZTqWDKBdIVPK6d9T3joxo2/+MUvXl9fvyz/h2EIxnZdzSVD+5exd21bn84nxgTnMs+T8/lsjJNSL5drxgQ+DKQOkMniqqOsBkmCV4gwASUrqhk09Pq+L4oM7NB2u0UGgi6264brJZkWQsQY1+vlOEYymTmnaYqhNu/jhw8fhmHY7XZzSE3TlJDrqoC6rlHDAtF309Z1YF4pJdon4MFubm6en5/ZtNXSTUoSdCZgIhRjrC6X3WbV9f1vf/vvhHhnTJrkSspiEdHnJITkeR5juFwul8tlHFSe5cdzfaovr4ejUll1ad9NRiZCCOd9lmXL5bIbB601I9cCcYau1lp6Be+BcY41Iai77dVcBGjmqkbHgZ9jq/cEkm4y/aVgHmbqVgiezrZClEXnKReS8RBCJFdHz6n8p1MWuXrjA8XPlY2Ydk/NFRjOwHa7BZxC28kYwxjJYuLddYs9nZyFxvHKQU8Q7VpGz+QDPjlGykFogNEdp3UUwM7gTBAi55QwN06AD3A7jDGcM63zWY6JDIceMjIlEvBsgE/YVaWDHswcwvBwUKwopZy55kj8QDgM4pVJKbu2C5urlBNJvaqqLMvyLCeRYpCWUS64JJHGQIKPh/3xUtXr9brIF5h1gFbKTN5Wi8VidhCglEI9BX4J9x1pAEGPUgrkSiZV1XWnb55DSHI+n73zkgt48/Rdl2fZ+Xja7w9959q2A+pCDwMQ0JnRe6+1fPPmFk+GUvr8/Bks9Lfffrvf7/f7fdd1qK2zLHPOeA8tFjdmmIs/rGTGQcJ5wO0O07wqkN9fh1xC/kEliBiBZjKdpGCI+/idz8/PUnJxnSwbBZegRJTSq9UK2zDwUnGCkyRRSkDwm2UJuCY5LeGD6M1au1wuUQKDep2VBuv1GgL/9Xaj0+Tm7rYoF93Qj+PYj4PzXgmB8h8fFUUPxK9oCcykPKAxIQQ/Fn2Roije3L/58OmDUJJLoZQKJDLGVKIZYzyS1Wolte6HXifJZr35/scfKCVzpwSkJIr6EEJZltCYwkMbaQzdNjb5seADTA3hHo4IeMK4wyjLDoeDUqosS4g3cAPBsOEPgpZNkhQVwzAMs2gaT+/Tp0+EEATleRxBiJ/dLpFjMKqdJ2mcph8iJcMwtJ+7uc+JxzX3KoZhCCTmeQ4fiLpt0jTlUlBPB+tSxV5e9+PQLJdLwdU42tX6umg6TdPD4QC4FEKo2z742LUXTuP53//jm29/ORiXL5acjxh+VlLCsXV4fRnGMVUpCizMx+r/vmHbGjsLhBhjg0Xz5ros1vsB5sn4vsaYNNUoXPw0ezxM2xLBoyZJEiMdTe+8o5RaY4UQlCkao7VOJVfCxFrjvZdSEUJRBBdFgQl5XGCEYyQGNq3Wwe9EZAQMRz1UVdXNzTZO9mWwXcJvFkKG4GcCCi0BcBpkGq1C1woxDl1Q8JN+svTJ8xzAGYkBH3IuKXCQMMiGz5PnWVkW3lvALCHY6TSE4BhLAPDnCc25TkJhgWSA0R9CCHiPLMEko3HGcs59DFwIHwJ2jZyrijK22WzarkVdO5PsODD4OjBBAFTCG5+9BjA5672v6/rTp08xxq+++gpVCOIpHhqwGlo+WBUON+y+79FKcZNbEXqHaCjiN/NpqGizXsMVvG0bjN+2bTsMfYwc2Wi1Xkd23QeFGHs8Hr336M/jVCRJAm787du3cdokijeITj4KYtB3CK2oMnH+gQsBB+O0oB5PGyTnXyE9CM4IiSH4NNHj0C/LheCMMtb3XYwEWJtOy7bGYaQsGccx0UxwgWzmnGPsKulFNYeyoyiKNE2NGVBtoUpyzhXFgnMGNT0W8gClVlW1WOSAFTNTKYTIknRZLvMsN8Z8fvyMYq3Ii/p8ds6VZXk8HlEjY/RMa4U3sd/vAb3xag+Hwzj9usbcNOWcS6WSJFuvt865EOJisciyvL5UCRd937dda41VWo3jeKmqvuuDi8vlkhAqhFgsCsb4+XwCczpb+zLGnp+fcZLwBSEUgTcvMNfUnrLL5fLl5QWFIVgmP8nzyYTaCGFyciZAcNdaZ1n6+vpCCAkh4Fmh2kOtinINHgZiGtjx07isnyxpceyAu5Mk6YbeOXc8nyS/zhngK8zkibGWMKqVkkJmaVouFotFOY6DNT7RiXF9InmaLZwPKkkioT6Etm1++unH27tb71ySJkVR1E1rzOBJ/PjpsalrKeX+cGmHkQtFKfZL54jdSZJILsauT9V1pQQqdFRpXV2DPQiUtF0nhCiy3Fv/4cNH51xZrqy1p/Mly3ImeB9jiF5rLQRXSnHB2CSBk5lEAgZqBlMvBJecUMpiCIfjcbVcLooF4dQ4Dz9WQoj3AfasQANIz3JawjOF8lEIKQTvexiRFWh9QYFWFMXlcoFq5ePHj2W5QAblnBNC4zwHG8PMM0Aw4r2fR15wuuDmhMAHpIyhHFwl/EEkBqA0N2lY8S+4d2gYoMBC6J8nvecgiw8zd+ARSQB6UODiYMcYL3UthODs6vmBEmq0RgRurSX66kyHweb9fp/nOXaYq8lIY5Zp0Ul2Af8SxMfqUnVtp6YpdERMay0a73hQCDJIXdbaJNE++KapjTVSyf3+9d27d+NkNUYIybLs3bt35/P5xx9/1FofDofn52el1OvrK0wnsyxb3y+RvaqqOh7Pi6LcH46MssViUddN33dM8L4fzucT5TTPc2NHH5xUknM+joSLrKqq29tbwK+ZaYRzj7WW0DgMPRdCKlHXtZJqtVqFECmlUIQaY8AfOudghhiv+yc0/g/P5y9NDzxQybniSkRGqbCR8cioj87DtLWTUhJCKeXGOCa46wYuk8E47/xyuc7S9PDy6mxQKmnbnnNpjM2yQuuUUl6uVk9Pn8+XOk1TyoXUiY+D1DrLtXMj5yTLEkA2Mzl8cc5R1qEUeHx8RLZsu24Y+iRJv/36m9fX1/ZSSyGds58/f0YDCkxi3/dKKE4oNDnlavn08qzz1DvfdyPxxJPoXMiLsljQm7tbqdNluS3ylTVhiMNqmWmtszSvz+2pqsfRVdWFc14kZXNuT8/nru+SRGxW72KMzvTb9bJt27693L159/nzZ3L1M7gKZ9GMQSr6chsM/eIXNorUU5jr+1HrNE1zY9w4WmNM03SMiTxPhcAoFaM0Omek5PuXZ2/HoW9DjC74NE1724/jKIi6u7tDtIJULE3TGMM49ij7gL+USjiXjIlzfUH5wjnPktQYo7msTmetU/B72+0WBTghJIaQqCT6QEKwo0mULovCp2ndNJHE4FjdmXK56Ibuv376eHt3W3/68O7rt89Pn70dQgyUJE07UCY3y2WRZnd394Swruuc999+/c6PPWLBu7fvrLVj13eXmjh/eHnRUjPGovdFlu02m+VicTwe+2FcUcFV1IJaO8YYvfFN1VkbsqwAFgshlEVGgi9STSi33o2DI/S60tm5UNft3e5mKcRoTNM0XdedTqfNZhNCSKbatO9HJoa1J5SyGFlbd5wK65z1nklhgulN7waH3rVzDrnWQfMdrXVWSJqkwrto7TjVkQmDhZaQhNCu6yllnEuVFKO9jIORnnTD6COJEIMI2dcNl+MwDEW5PB6P3TASQtu2k1IOw2iMMeY6RR/jFZbNoqnNZlNVFSV86HszOi44pSQEb0wvBMO/53nmveOcY7CRcyYVz/IkxuicwWR7JD6ESCnDKjfOBefC+977YLzhQiR5MroxKdJ27HSm27ErWC6FZJTHQFyIhHEbovEhoVRwHij1zqVJYowhMT5++nSpLqvViglepMlgxhiISnTftGmSdG1rrXXBZ0V+vlTdOBjjqqoOISwWrVInBNDHx6ebm5um6fp+TJJMCFWWq6qq6rqpmypNVd1U/+f//O7u7o4Leq6OzptL1aA0QYWN+HOuTiG6NNMhxLzI26Y35sjY+e7unoVIuBis95QdL82i3I1mzMoiqVMzjlKLcn3T9z2jkUR/OZ3zJEVlWffD/f19VzemH5VM27a/uWFJknEuL5cL55JSviiWi8WqbdvgyWq5sdaFQLfbXd/3Zbk6HA5S0uPxvNlsYqRQA9Z17ZzZ7yudyHK5+HOHpv9HeuB8Vv5SSqUQztok0WimjVe3jBGn2DgbKbqGkQjinENcC8ETIrAZzXt/OBzA7bRNi+YMwAikx+BkKGFY/Ocna6DVaqX11a4SxdQwDN999x3AS6I1Z1cDnNPpdLlcAClOpxMqLD65zHrjYSUNby8ohZRSmuuu7ajgdFKMAaQvFqWUknMxjqYoCq0T51xdN27qfwoh0iRNdBJjvNndxOgTneV5/vHD4/710DRN1w7GGh/C3E7Erxl6L5dLIFPgdBTOZLLnQ1UBLI8hAzCMGPqNMXZdG6N3zsE9H59/HMdx6IN3jPO+66yzUkpORdd1y1yDe4E8CXHf++DD1cjeuasTGXZNPz5+xGcA9wIYhSoEBcR6vT4cDmRakqyn1QUoEMfJOfJ4PuV5Tiix1sVImrZ1j49DU5d5FsxozfDu3dsYibPu6eX1H/7ub1er1eFwgDZXa80Zb5o2EnZ7e4vHqKR0ziVa7w8HrDCjlOJMggBp2w60WJbpcRyPxzMmIrXSjBOMU+V5DsMYa62LxHpnnY3dtQ4DsYnhEvjCOueenp4gfxCc45lA/IN/2e/34PdjCGVZEkoiJYzzuj7N1ofQnhJCTiczjiOlBKwtmXq54PfnwjpMXqTDMAilsLtCTC5DcITEDwSQQiOtbVspBNaNzXUAaCuAaDFZ3eEXpRQujSGERbZo2wZXWE1OTeO0fgBHlE6OWzOZjpogxjBfDXAay+VSKcWCsNixGqMUV0u0GCPWJsPfVCmFnSW4EfhUzrmXlxcooP72V3/7w/c/hkkqhhZXkiTRXaeg0yxTWuVFQTnb7/fb7RaLzdu2fX19xd7Ajx8//vrXv16tVmhhHo9H0DVPT5+7/mqE9Z//+Z+n0ynP87/7u7+TUmfp1exkbh0rpWAoYIypqgslVEpxPldv377d7/fffPMN45wLLoTo+4ExliSplGocjPdB6wR9EXgXpmn25s0bIcThcLxc6u32ai1KCIG4GfInPDHG2Gq1/kJ/fPWyxe0DFw3+nE+m1DN53vUA9JGQv4aphgshxigZG7F7IUbCWF4uvlR6UHL1cJdaSa2890PXAyWhRUwIybIMNTUEG+DUmuZSLhcNF8f94fb2th+6GCPfbJ+fn4f+ahIH1wcca8h+0aRaLBbY1tD3PQIcTvD5fIbXk3OW82s3FecbmhYo3988PJyrMzReq9XKGmtHE2NcLZeMMWz7m9o76vn5GX6c6/UacxiEUaEVJquTPMvLRW/Gcr16eLg3ZkzyzJMYGcUMbTcOr6+vSZLiSKFDAKb+crngGoMqyfMc/Qb4/fovFhSDi0SFDi2W1rosS601aHHwHuBbUIuY0ZIY8zy3zl3qC+M8zTLkV/TDu67b7/dox9V1HcJINSVT4xQkAJrMuOozPYLov1gswZlCLBinRdyzGp0xhuiJaAUS0oerQ5dzLgRpXfjxx582ZUFiSNO8G4YkTft+6PtRSp0kGaWckEpKPY62aZqb2zeoBdFmQGXZ1E2aXWZHjUnuHY7H4y9+8TdJoqEUwuwhISRJkmJxfRRzSlZKjU1j+s70HdU6SZKmOo9d+/z46Ve//MXc14GIAH0jbAQ6HA7o8z88PHjvITJRSl3q2se42W5Ga4ZhAL838+Az9ur6S5omWFEFzwJISxH+4qS+z7IMs2ZIG+CX0C4+nU6oSCB01lqDYr1cLnmans9n2LHIaYuAmxz1yaS1jdPa6jgNQKDFghbC8XiU8uomTSfvoKvFU6CcScYYo6IfR2eDu65+YVCsxsmK1XvfNo1QMlDqrF2vVm3buqnopJPyZ04JhBDBOAlRSck5f3v/8Pvf/55G8k//9E9udMiFXddpqa42dpFIIXa7nVSKSxFjXBRFnud+9PA2xoVyzt3d3aVpend3h6s3R1XOeVHknNOu75rmkVLuffz++x9/9au/JySibyEnL0tkbu/84fAaIxnH8fVlDydg9DxgHAJFA2Zab29vEYWqqnrz5s12u8VsSpIkUmvCmEqSxXL5//7v/53mubU2TtZJVVUhTAEsQlmL1wGnHzqp7XEycZ3Bv82uxkopzinEDjhgf4X0gEMwy4pAMtBprS7aOJCNxxjTLNNpMgxDWzcxhCEOZhitMYtlOY+8McbgQIfZMXhncs5BASEkffjwoWv7JEkIYXVdl+VqGv+7soeQnaFbgO4ims96WngCoU7T1Gpa5Sal3Gw2dJrJ2Gw3Xd/hiHRD35K2SIrL+Yyvtt3tgJXatn379iusHEGfmU01iuIM9Cu8aIAoGePex6bpsGo0BCKlphRdxC5Oti1Jkrx79265XJ5OJ4h96VUXGGbmAa0toLB52gVRDGMfyJHgZIsi49MiYjE5sZBJME4pVVrHGPuhN8ZQzvM8f/PmzX6/b9sWw8BSCmsdHtdiUV4uF3Qv0fdzzuGvQyRFNQMt0PPz869//Ws8E5SAkMyDt0HmQHCM9Dr1AjzYdV2iVdd2XZIqxmIIfd9/9/v/Wq1vRmNHY9quU1ovV6vtbmeM+f7776vLJUT6/v17QkhRFN9//31VVVVVQX2Ebid8ovDGcaIYZZSLsiyPx3Nd1yFcrejQaBFCQPmO79VNy3PMZFY4S85gk4WZcJhxFnnunEP5C4xGJ0lijPF0PussFUK0XRdjTPS1UQzG/HQ6xRiTJPXBIMlBj08plVLNPQMgGPQDcEgAxtFlmaVu6DRAL1CWJRokbdtySqGzQGsa32IW8mdZhn3OeO/WWs7kdIq4v3r1p5fLhTGBhifCKP55PB4XixznTU5LI7z3jHGtlZSCcw4A5L1j7Or2j/kkNmm+yVR0YtjtfD6jGOKTtBpsGHzIIahbrsqrspmdRXsAACAASURBVIiEvu9D9KNxkUQhRLFYeO91kozj2LRtkiQiEUilEH1CUIvbwadfAK91XXsfqqpK0sQ5h/t+c3OTZdn5fH497c/n8/39vRACF9AY47yNMUqpiqJw1qdpKoSEWByB6OPHj2/evLm5uZkusttub+q6/fjxUUoNpkgI9d13/1WWK+/jcrnOsuJ4PC8WixAImyYiZ4CLrhIGxaEXjdPeDkLIdru9VgldB+EJDuTsmW+dQR35/PxMyP2fkR7+28TEcr3yMYzWMMGzIu/HgTC63m4grZ1XGuCy0WmcL89zxnmMUSoJzDLXsKCkwIowQqLzeZJKxqPzgrLNcvXm5hboEsiOfLFDHK/ETmaKsMOdPwM+D3QCgCTz4ZtRDyIsntr8jnFPhLy+9aIoUN/gx+IzICnOwq08z+um8SGs1uthHLu+50KsN5tFWQopQ4xplkmlur4fxjHEiIEDREy8ue12W1XVDz/8APi2XC5h5f3x48eqqmaAD6mfmFbiTHNwcrVaoWM5DMPcCRRCgMwxxpxOp7Zp2rb1IQAA7vf7y+Wy2WxQwDHG3r59K6VUSiVJstlshLx+NVB8s/JvuVxuNhvwMPirt9vtZrMFZoEzPnaUzpwYm+Y5ACMgyYARLPzU8L4oY9bYvh8SnTLKluUqBno8np1zbdufThWlHPcH7PDhcPr3f/93vDL0V7H0exhHNAOgZQqTrz1YBca5967ve8y7dF2nEw258+wdgnCZaEmiV5JbMzR1RUmgJDAarbVwT0N2RyLErDgYrYeHh5ubm2EYjscjEPfz8zNsFr33WZ71X7itzAoC773WKkk05HOIYshG+I5zGxz/An0zrjfQgJ62N1ZVda3gpYTCFVBXa/3u3btZwQlciemWYRg+ffp0PB6RX2EojSRhJudUoA2UfYiwGI9AQKeU9v0YI7XWx0iNcWma39+/Xa2WZVnkeVoUmdZSa8kYUUoopYLzdXWpTueuadu6ccY6YyEZCtP0wIxvhrYjPgTr7DB2daO4uL+9y7SmwdPgo7ObZblbr7TgWnAl4MDIGGPBe6015hgAs4qiAKjHt5sjyc3NDcIL0AClTOtEq2S1XN/dvnHWF/mCREridYAAyiVrbVVVSPBQEGmt797c/eM//ubh4QEsRVVVi8UCs1awGwEzeXNz8+2330J2iGhTFMVqsx7MeGnqw+n4D7/5Ry7FaM35UoFU59OoHee8qqqnpycwnM/Pz+D8GWNt2768vMzU5Wq1AmTBJZ1VmtAgzJ3CvzQ94LyCf0C4nMTvCtPIqLhRvIBYEFzEGBHNlVR62oI5k6p22gnMCGWEBucZoYJxb12idJakkgtIJPFtEVLBJmOMDsoH8OaAori0c0ZBPoBtJCLgHLbEtIoEYglCyGKxuLm5kUIiLq/Xa2BP+LDOLkm4qPjhuHVff/31crnEdAJOAOc8SbLj8dz3oxCqaTrvo5TaGAu+BYwHpOKASKicQggwSuv7/nw+o5zEFyTTomm4XgNBYDQBP3Oek8QXgXoySZIQ42KxWBRFCIGz66jkt99+C2c0QghUFhgMTNMsTRIAWxBZk5Ajubu7e/v2LVA23kjXdSDuME8OfS0hZKKMAp75rLjF8bWT7QeuKGQe9w8Pi0U5DOOnT5+d82maXS4XrZNL03z3h//6/PI8OluuV0+vL1VTP+9fY4zffffd6+vr09MTDmRd1zEEQq62j5hIf3l5QWsBVDjYksWiuI7aEgq48PDwgLSNX4qLseu7uonO901LQyQ+ILhAzHp7e4sVUoAU6JNtNhvsucMPcc5tNhskYBJjjJFRhjUyAFXzKbJfUCsz+QbwjhOLK6MmNz0csGHakYAbB+Ul5/zp6QlUhjEG+8iMMXFysQZKQxWCf4Hkuq5rCPmQgQChcNpBcEEKzKYN52ZaGAWGahjGYRi7rjfGCiHbtsNc0fxJINMEg4f/J64kIQQH9e7ubi6pZ5U8smmidPTBDGOiNCM01cnhdd9capRUSsoYQgxBCpFojWk7XPyu71FWoic033qQw/GqZsaUco1vB5ZGKVUUJYp+IdRyuTbGXS6NtR5xDNoEFPqfP3/+9OnTOI6E0ClfXnEkXnSMcbfbtW0L8zTGmLWeEFaWq8ulqaq6quqPHx/3+2Ois+en1/Pp0nfjdnOzLNeCq9Vyw6e9uYyxpmleX1+rqsISgf1+/8c//vH19RXhQkp5OBwA0fD72TSKhOc8NyfI1GT689LD/+g9eOudcdZ73/bd+VKVqyVlLExtzBijswHQRmvNBA/eJ0lSX2prLb/OdjMgVlCfr6+vZHILIJPuHtjkcrnU0Jiaq1EdIURNxughONxAQghC0nx/5vGfl5cXVHb4eIi/QAqAnBgFGMexqqokS7XWXArn3Ol0dM5ty5uiKKrLpSiKNM8IIXXdwtkJ5ScO8Xq9vlzOeZ4X5QKMFiHk6ssWYjH0TdMkaTqacb3dSK3IZNQONh9Zfbfboc6F4hCEmLhaSrR8mumbZj44OupQOuI/zSAOVA+CCJoTWmvBlrARLsvycDpKKb/55huwUp8/f4aDNPahhhC0Vt6vEcSxTxjMHt4L1GJIHviyOLVoLfzwww+wJcD9RFxGiAQkXK1WkRBNIiEEFkygayilm/VWUEKck1J++vQYQhBCvr7uj/tD2zb//M//HDxp6q5t+tOxssa70QA9UUqxjpRSent3C1oJLAH2IKEgw9qJYUDHa/njDx/9tAsB/DV4TkopIXF3s3331QNSLxZOeO+LIqOUYu4EX/zx8REhG88fI6ly2hmFvK6U2u/3q+3GGKPT5Pb2ltgrPALdx6/bZuClI8fRWmsZvbJJUsrlUoGMBbeAeC2kJNPUN84G2AaULMjNmP94fn7Osmy9XFJCUC6cz+fNZiOEAE8IAAsiDjwVfhRIJKTDOZrQ6yTg1U0Hj11KiUXZwDrIdt57753zkbIYohtNj0ZLiERxQQhJpFotyiLNrLWcUMWvTTUAHURSQHX81WJyiwK3ea5O1o5KqRCU9957ezxWjDGpUu/Hp+dnSimUIHXbxBg1vw5sHY/Htm2R8Oq6/vDhQ9M02LIJ+0uU4zi3RVHc398TQl5eXoQQ/z9rb9ojWZJdB9pu9lZfYs+lu6pYra01gARowL8vSPNRGGokkU2KbNaWmZER4evbn63z4fjzLszHYQUKiaqoSA93e2Z27z3n3HN/+umnw2EXlrnIWP/NZuM8CpFZaxPCJegCWS2KIs9zzJs6Ho9IOhFZt9vtdTxR13WM86Ztz82ZfH3GCEUfA2E0Lwok6EAFjDFvb2+vr6+4PDFl8qoBAeAB2BCnEjINFByI+tM0Ccmdc22LU/zwLw0PSkrBRQyBUcYEQ3HQnls7D3CSkVLauZ+mSQhujMFVSylljOZFWeRFcM56H1MihKDvkS1NN23bcM78MiYlxtg05/P51Pd923fzZAkhVbWilOZ5Ps+TEGKeJ2ttlhk8CVQYp9PhdDrFFKuy0loTEmP0zlmTXbTYSB+U1iHG6COEpNM0FVXpnIskaa2LolzXK58uHaGUUsmlc5bSxDnlHEaPF7WPMSaE4lo2xmVQc9N0ry+vXdcfDofb29vc5GVeDv1o/cwYnabpeDyCmEIbBKZE4MLCdYmXSil98803aEoghISltwCbAPTD4XBA5t73vdYbpTQ8Nq5MFNU6xmSyXGuTmUxp9f7dh24YcMUAKcKFBfFM2/bee+f8arXqukuIijGtVhvOL1ce7IAWqpn/+c9/xhWJUgNVHVK/a2yWUjjnKGNSX9TW02hDSFpnnIn1as1Jevnyebb+b//u7+8eHvtxpIRGErkUkSTCaNO1Uisfg8kzwenm5m53ODrnZhcCSc659aq+v912zd03v/vQD0NzOtGUnj9/Hrvu+fOnd0/3l1pWaUrJOPbDIKXi14YjJIZNczZG/v73vwfVjChY1xX0Kqh3ocC5tgWgZeTKWHrvN9stzvM4Tf/uj3/c3t0yRrTWRZGP3WWyKe4ItnjhOeezLHduiDG6ECilSsHf/yJwyosiXkFOY/pxRIRAXoL3cLUwAkDEli9cHGizwl4CkQtWw3tvrZvnaRgGxvjX59dvfv9dSskHjxQ7xiilyrICjTLqYluQlmrAee/RJ49UT18MuEgIKQQruCCEkcTgQcSyCIrbaD3Pc5YZzgWl1GjtF3suCPOklFVREO9RyyJiAdAbhp4KOjubhUAoEVL4EKa+W9WCcznDHdaY7GLQ5KdhRP2KZzRN0+l0wkw6qMZRQsWljaNpTtoYY9ThsMckqJ9//pExAmIZgAQKgqY9f/fdX8VIpmmQQnPGm6a5v38AcogwvNls/vCHP6A2oosLEcR+QKhSSs7Ofd+nGEMI59PZGJMZQyltz41kFBYmANLBKuMihQTrfD5vt1t0EN/c3IzjRCkBi3ZFMsNiF3ahqXwghAX/m1DTUrNEhraHiUK0/vX5K9R+wUdIlSmlq1XFOTse9y74eZoEFybTZVkoKWWR/fDjj1wI5y6xYRz7p6eHosi605ExrhTnnCcSrJtW6xshyDj2hESwKIyRYejyPO+6lnM6zzMUFq+vX+u6tNYSGrr+PIyt1tr5yWRSG6GNoDTQGIQW2igXvFBaZ3k/TcrkSpnj8RwjOZ/aLMse3z1577XRIQQ7WjhPcMayLBtIyjI521FKWRQZOtGstZwJQlhZ1vf39ygIGEshJDvNQ9sNXVcXRQlbvRglY2Ex+MNBdc5BdJRSwsJCfIIAACjgyjdc7Q6xzuPYT9MEbPd0OqUUtFZt29/ehhDSfn88HE5lWRZFNfmQac2labs2hsAoJ0SmJI3JSboIbIZhOJ/PIYQQJKOKEkoJ8Y5olcVwKRfqatO2TfB0miznaOmiztmiqMVi6YOjxRah/vv377Msa9szY+t5HiklQpphGBkTIQTGhFZFVW5ubm5YCpQSZnIeyWvTZxsvtFGEchoJIW9vbzBvn6bZe//wcE8Ft4TwsuyalmbZh3dPKQaa3MfbemxWdamDn243dZomOk75+0dG0tB2JBDFFSHEZJpLejofCb3YcYMXQbYxDKOUoihyKUVZFlIK732e5w4MPyFKyrqqNuu1lDI3xk4zoawqyn4chJRcykCS9bPmyRRmfbu6u10xQoTglBJeZtg8QJmuCDhnMvgUA+n7EUY6Ss2U0nM73NzcREK0Mc45nWXe+6wo7DSddrvz4ZBnJi8Lk+nZWUYZcjUEb+/909NT27YpxTzPCElZZlarOqVkjH57e4303TjPkRAmZKF0SHyY3Gq77adhclPN6kQpYdz6EImjgmeZFoKF4Ly3lJKUgnNzImmYbG4957IbJibU2/6T9XEOnnHqrRNCSCZIpEVWKqGFvPjEaSVQxSKqOeec9bO1RWaU4M7F3GhKiY/h5u7We8/lZcRv0zTaGJ+84DwxTghJjGVlFQiNMYV5ur+/3x8PJEZBGU3EW8c45ZwyRherGE4pV8pMk1XKWAtnXLNeb0MI5/NJCDKM/ape9cNQFLkPLsb48PBQljXygMNxF1P48PHdbifv7+6+/fb7v/mbv+n7cb1ed13fdf+MygxxSCmVZRkyOWjPGCPn89EYJQTb7WYoaNw8W2vv7u6klNF5whgJMXj/6dMnFO7zPMNG4evXryHE06kpy3qaprrOp8mC+Hl+flnYfk8plVIwRhhP0TvBKad0gkTCmKtbzL80PACpRCYCrTe0BMfjEc7JAEygfVScAQjOVtlFv8zYPE1CCLIUpyhjAQSlxX0MUQ44bNu11lqhJFTMyM4Qfud5RB0K9QU0JNbaREJdA/VjAHAJIYyylBK8rjBtoqyqoihooLk2L7u37XZLGQND+8///M8AeQAiEUKA0gbvASwiOcIo6a7rbjY3PnqAWoh50Ki1XYtnCVw4LL46fd9zzgghePGXl5d5nt+/f08IgU0Q6Ack3ahdkF2CV4fw+YpLuGU2GSAmQsjxeAZJhRTvQu2EKJWaxhEeA5SxLMt8TLe3t1171ou5PMBZBDkwTE3T0KWR+yqnw6+G0Nt7Z+0cArnqcTGRArqmPM+22zVjzHs0edr9fleUa8ovvopo+0LOKxk9n48wQwYihOSxLKvZuuPpzGiTEsmyrOv7tu24EmU1Ttb3/VBk+ne//93QHv08kxgUZ5ymFFxmhFGcVHm2WnOlkDne3t6qzIQQQozjMKAkN8uwMFAUsI4QyyAEoCuci0QvU9YJIYyxiyaE/aVTnXMeYkRJEUmilFPGOBNGCcSVEIlzFzMYAP14WMMwNM0ZxQq+A7dEqAkwhoQvHcvgRbXWTdO8fP1aVqVQF5BH8AsqBTkfZnYCf7+aFDw9PaFZ+urNBZIDbcNIsQkhaM7Ay4KNcNZ3fR9iLIoixNg0TSKk7TpC2MePH8H8gYMkhPzX//pf/+r77373zcdpnN49Pc3jhPW8wlNiaWm+4pPbLX95eUGJhgoYRTljbLvdHo9H7D3sf6UUi0xIoaSMMTrvpZRKyuijd66qqm7ox3m+4mBDPwgutdaHw6Ft29vbe+RhkHTDR2SapoeHB9x1WVbFvmOM51n++PAEr5qqqqbJQm5OUoJ2qKpK1NMQcRyPRzROV1V1LRCllDc3N0qp4/GIA4uCDxoNQLu73c77CKcDsLzA2/H1ww8/ADK11m42m6enp7e3tyuIhNOH5wUASggxz6gvYwiB0MQIxX0SQjidTpjT89sIW3Hgvffn8xmyGdwauNf04oAG6nh7e0NgRJsS9iu4h77v6TJ4BDQv7lO2zFvP81xIBvql7/txtgWv8DGwua9XLdhvsnDmQGact9Z6rQmliRDmXJBSxxRDCEJi7EYVQuSc393d9edeS6WV5pw770Hw4gaHJONqKaGUSikkG2OI3rppGN1sxzQE56dpovxCJOI5UUrP5zNG+CJluPJvdV0PYyeXVr7rOf/973+/Wq32+z1CCw4GcPmqquBZiwQE0pRxHLNMhxDmeT4cDjhO6PVDsY/V4JxDdmWyTAphrU2YXaF013Vt14MexL1wJSQRJ0Dz8GX0Lop9svQHpMU2R0rBGAWsIYTA1LmiKIQQ3ru//uv/s67rn376Geag1lpKed8Pw2ItCUi0bduPHz8KRhhjCAxQ4KGHIBLivCeUbrfbt7ddSsmHoIyZ7WSyjAnfDcM8z4xxrfXr22tK0WhZZHrIVFVk57e3oXWMJJDkcZ67vkuMZllmsqzvutPpBKYH5P84jimpEC7iBWxLPMEYYyQXb3Pc7LheOWOJUmstE1wbAyHWMI3zNLYxuXkmhBKCmViEMcL5RfYG7RbURFdklS8DP9zy9fb29ld/9Vd8MTUBapdS4ko555x356bJq5JQwijjjF8FY9DyA4l+fX1FFoU3j8+FFH6cRlzrmNH98vKC6Y1//OO/u+oDcX2HEKZ5bLpeZzkTsukHKeU42yzLUU/j9odeo+u6u/u7oiiUlHVd83qV6Qvaqc2FOEF2iI92xUyu7AICCV2k3kibcO3EGCklgjEtpBIypSQYhx9ie26lkpQSzvlmveZCEEY55+M0oCsIE8UXfIyghkMIvKoGCCHWjkiw5tk2TYvPBb4wy7KiyFer8vOXz4Skh4d7ziRj8ubmZr/fA/h9eXn59OmTMWa1WgHZo5Te398XRXE8Hq9kYdM02AyEkCzL6nq9Xq8pJZwz5Arr9YoQQmkqy3K/33/48GGapizLPnz4sF6vQ0iHw965iwMu5wxKLTzQK/ceY2SJxpiukQk5LjzffoPwkBbf6SsdD6UU9mv41SxsyLl8DFIIjK/J89w7d4nJxiBslmX57t27uq5hd4EL0do5XTzdBGPMekxfKOd5nqYJnYBZlnl/8YNUy+TYEALnbJqDd54zDhS8yItpnqZhcM5JpXOts7w4n1tCyXXTCykgsFmv17iOg3VYVtTmu92Ocx6Czwpzdeay1rZtTxJxzkUXQwhVVblfdTULziHOQa8A1Cld13VtV9WVWrz+w7UDdpnJhR43sOhXggEqVWDc17Jmt9uBqABDi/+rdTYMw+fPn29ubq6ya6Z0WVUsRUqZlFJr9fLyMjtvhBCcoiixyxwL8itXfcQq0N0Q+4bFmW45sZcPiN2PUhrHmBAfov3w8d3r21fvrXN+HCcpZUx0nmfcRBC0OOdOp6MWHKkr6gakcpvNJibCuQwhxUS4kDEE72dAuoJzHxLnIgVPUlqv16+ffu66bp6mGANnNEVvjKzq7P7+LnJJCBnGAVURrmZ8NBQKyANCCPM8HY8HQLcppeucK0JoCJcCTi1+sSklzkVkIaXknSecUUoZZ9Za50NeVCQRa10iJIbEGPU+7vcHIQS0tsirxNI1hqW7yvCGYYAy8spLTcv4imsuYrTp+/50OimlQwwxRC4ufdFpmQwK1Sbg9bBMDseNP8/zn//856enJ7YYmKMchzDh9fX1hx9+QCABEIE0Yhwn55yUQUjZ9T2jQsmLBv3Kun/77beJxBhjWVZ4n1JI59z1B673FEKLECKmi8gFZAlyYRxV4PjOOZR6xhhCqHMOw6gRTe08K6WqsqSEME6tm7uxK8uKcaa0UEpBlgpPzDzP0aKE9VTLkCVEUCklR+HHmNbZ4u6spESDNDudTpiaZYx2zsnckERwM+CCRniGKwTWHLm/Uurr16/DMEAdoJTCR+v7HqgR5tJnmdFaxRgRLYahu7u7w3xGFCW3t7d4qz/99BPgMuAHlNJ5tki/hLgSwMwYIygy3YtVV1xGyPwG4eGaPwLHEMvAdzxaRAg8eMhPhZLwxvBLZu29F1KCB0Ml8c0331x95bCI8zxrKrMsy3ODpAb/yxhjZ4dFJL/yQ4YSA75aWpuUaIpUCMU5q6s1Y/TLl2c8Y6nkOI7f/+FfOeersgoxcMkLnXElz+dzVddXhWV3bnBR4o3hN0opskwH56cQpdRlUTSnFnX37GZKKXzNrsLt2VpknSgSL8WNc3FR+rpl9ifSZwgY8OwBAQFEKsvy+fkZqju8mRhjVVX/4T/8h7/927/9X//rf0HkyhhbrVbjOL579w59heiXviwplynGcRryIj+fTzd3d13X8kt3LknL1PKu665+2jiBkNUiy8PnQpKLmUKn0+nh4QGwEm5MWG9ezCwpkZIKQfu+a9vmeDxfVMTK4F5DMIZnctO0D7fbabEiR6Xo4KlXFFLr0/FYWRdTciE0Xbe52eZG73Z7NPOTlMZprOpaCHk4Nk0//vL5y+zmuqoIo+vNqqjL2SUIasdx5OoynB0XAUhCpNW4rc7HHehKpEQoAa119MLi2nlxXSWEWDsneGSOIyM8pqSXVjtnXVXVXdePo1MC/Nn4d3/3d0oZwGjodoZ0Crc/gOm6rkFX4tFc9dxwfTifz4fD4V999x1jbLVej/N0Pp832y20OtM44UmhALoq38Bb3t/fX+Gsq/sIUEFrbVVV6/Wac8Y5xd9FVQph9/l8DjEJqcdp7vs+zzMuZIwpLbgcdgtSn67rijLP80zyywAVxhisLyi90N3oVUqLI72QWiwyfJSwQogLbescfgxBCHffPAbBOeKMFEJJFUOkWg197727v7/dHQ5KicnZ6z12f38/TZZSCn89ay0ODnYC5OYQJU6TX63W5/O5LAtjtPeBEJoSga5kmiZrJ5PpeZ7yPOva1rmLa/1qteq67unpCf0NkIwjcbzgqFI2TXM6nXDetdZfv369NJ+nqLWG5WKMqSgutgjr9RoDna4CCmwMKYW1s/euLMuUMNhDci4IoTFGQugwDHl+8Sbg4tIDCBoD6vnfpnpALgwO7Xw+r1YrABcIfbjl8fkvynGjvfdXI9/z6bTb7civdLiYd4Exy/gZv4wNQQAgiVRVZfIMVS2+CUv3a42JbfcrQMbkeYFquijKGCNICEIIo0xpVVXVer2pV7VzbmguQ63D4h/Jl0Yksrg24uaFbiLPNGdsnud5coxR75w2hhNKl84moK445NZenE0JITioKLOwkmTpDED6ho2CyY4oLHCAUW5D44S+dLYM1/348ePz83Nd14Ayl6+/6Blw9VyasaVijDVNi0NYV/XL6yESRrRGgoOthtsEzxGBrSgKYL77/R4fEIXtMHZKqcNht91u8SYR55AcgKiY50YbJQT13uVFdjicQoiEeNv3RVXiYrry7V3XFka7xSdqtVodDgcwkFxqRqkPcZrtPFutjRCYphfHyXIpE+Vj3z8/f313v82KYn9uXQzT257QlJgYvL1Zb2ZrExFoX2rblnD2zTff9MOALPW65sBtURkAiEc9gYuPUiq4vIKQfwFGUprGcZ7n2dq8LJz36mLBGxkVmclO4/Tly2uemXkaT6fjf/kv/9f333+PhtDdbgezr3fv3iFxRmxWSsHfsCiK3/3udwjz+/3+avQP3Onm5maz2fzw04+jnZ2zSmsueHNu0GKJ/gwhRNu2mVZo4YaVKcSveZ6HaQIIGeNlvCVjjFKCRnGl1FUqg5K6H8Zpmo3JQoja5IxLIZX3/qeffmKMgQaHJXgIgXJCCEkg4bjgjHHGU0qEXGZkIa2cl4EQ2lwEGmhFuub18FHmy2gK7DGSEvqlSEwxxixTuTF2mrgUKTPWzlyKLDOUUcYodiYljFIK61Pn3N3d3Y8//ghtCFtcwlDO4opnjB+PJ63NPNthGO7v772/SCurqvrnH17OTaKUSCkeHt5hOjo6WNu21VrjCV53EVnwNBwlu7iLIgM+n8/GmDyHHwxp22aapqL4XQguhEuGDadYXCOUUsaolCzPszzPiiLDknLOUyLOAWlwiHn4E+wIlgJST1ykv0F4wKO6Ejg4RajUgFxDqnUtVaSU4zhlJoNVw+vra9M00zxLpQBD4y6+vuMrTQdSRUreDz2hAuz/PM9K6hgvY7sZI6iywyL0xqpJqVF0A4oBvUFJ4IxIJbUxRZ5rrUgiq9XKCNMcT1VRdF2H9AFn1Vp7bYmEvMQ5F4L/cVuG0wAAIABJREFU/OnTerURjEupOedFnhdFkRdFpS/m73QZZtJ1HUmEMoqLBncrpOh0AegQG1DSKjjNcZ5SwuxuYNDQpcHWAlJ0tPxA/wBd3TWx4pzHePktV2E+Qq8UEqAQklylFeO8qso8y+dpgH0/Ai1Sm7fXPdR+2TJhFID429vbtMyEwPuXUhqTeZ/wELHjkY1qQ4xRUkmleJ7neZ5TyillIf7lkAA2YYzt9wfFWVkWcGymlF6gp+BfXl/QAs04L6vqfD4TSiJJQoiYSN/3XGpr7TD0IayKoiJcCUmlkYnEdhgjFS5EQWkM8ccff7x7uE8pwRtcaw25BITRyD/U4pSOfAjb4Ir/kJRub28xZJguNrqMMvQwtn1XjXXX9+vNpu27EMgwTrc3t5TR5y/PL1+/pBRTisCjCSHw8IDMNMaYZRlaZ5HJAhcOIXz77bfTNIFqxq7ebrec82urJrQGRhvCqOACVex2u729vX16eoIlzDVHhvxhmqbtdts0TVFXUF4iRRNCPT4+Ukra9uSWIUW73Q7HP6RosiwRkiip1yullZDCZBmJ6f7+/suXL8iNoMQry7Ifu2matFRzjEpIugyeEZJB2g4NCBafELLZcszZvdaOVwnc4XC4v79HdEfMkEJ04wAwB8UcYjYosXmeClUVZW6tJTZaO1dVZWfXNI0QCpx/WsZ0IyVCW7VYJjLxZVoOWNxrq2PXNRfNCGVKX9ygy6Iuihq1NTD2tMwTBOhyHYeHF0GEuB5/ay1mo8UYxmnKMuO9w6fIi7wsy9OpQTyIMRmj59mSy5ipIASnjM525oxpo1OM8zxaG9q2dW6mlEJYnhdZP3aYv4RrAZv8/19s+P+GB0JiZow2+nQ6xiRCTD5GhAK2eNvigBljfAycC2e9yTKTZT5GF2I3TpQSqVUiF7US2P8Uo8kzbQxlbLIzY/LDx99JKT59+kQ4UyrjXJSFtNbFmDjnlJLD4YDjhB0MplpKSVJklBKSnLPn5lxVFZeccaGUSokKLvOsGPohz3JOKGOsH4dbeX9FM95eXvViaQKmZFhmTQ/DEFOiaBY1hjLKJc2L/PHx/tw1Q0oxeClE17Z2mpydtdTez5jIPdspxXjJN9NlqhSe7pXxQ9stcv9roRBCoJRM05hlJs+zxcw1YMQVaDSAwmiSpIueOiyjzaBzCN72XUNSOBxORVFu1jfHfcsI7fvG2plSstQr2ujs9vb285evQ9+HGBlnMKjJjGnO53EcGKPOTzEGzqmQLBFP2WXITIzxy5cvqAullEIQO5HjoYuBUyqMyTB1hxAmhEwpDn0Xsyx4FxiNwQPux/C70+mERcjz3IdECJFCtE0jhCApaaX7pjNCZnk2WycY84y/vR2KLNsdm8cnIjibhnkc+9PpuK7rh+1dSpQrxaS8ub1T2sTdm1K67XohRIox+tC27TCOjNKHx0dtzDxbpJzTNOE0WutjTJIRJZWdrOTSzpZz7rlPkSTGTVEwJTln83H6+vWZ0ESJEFLRlFKI3vv/+2/+u5TyP/2n/7Rer4Tg3vvz+VRVpZSiqsp5nqq6zoqCMCaU4lwwIRTjnPMi0RC8nWdGaQh+vaqnaZJC6CwjjIUYuRCEUq31OE4A4nH14LkQQrTWhDLGxWq9xcWUiH169+F//s//qbOirleEXFoNKGXzPGmt2zYeD4d+6LM8o4yezo1Q0s7z9uaGMzb0A2dsXa9ITM5aRtjH7z/O84xHDyTAORdDJJR4H4L3bdtOUqZwqdJCDJnJEmHH0+W2TTElwtbrDdLzvu/n2Sql1usCpimPj0/eB2t7770QXJQll8rF6FPyKQ3TTEjiUuZ51veddRgOSGMINCXFOZeaU84YG/qGM+rsSAiJwVpvx2nUmX7/4b3zLoRAGJntJARv26Ysi7ZtnUPX99j3/TD0lNKyLB4fn/I8+/ryVXDufABkivq773v0J/d9p7VijB6PB6T/fd+P4+C8f3l9SYnkRVGW5W5/gGEMpWgFkwCWP3/+PI5jSvSqjkVijfs2xnhuztbOWungg4uOUCaECC6QFClJKUQuhVaaM25HRxJtm26eZyGFUir46KLngv8G4UFLIYQIKUolQ0whxlPTIDBCxkeWvj4p5e5wmKaZcU4om2bLOSecBZJSDIRdBmwD7BZCSCEoYzrPhnkapsnkOeWiH2cfiZbSzjb4AJKDMRYCw+Ctq5EqmuC3221RFEPXZVmWZSamEKIfpiHRxKWMiRLCCWF9P3jnKaHzNMeYuBCAki+jIFYr4LMIsHicV6oqcqqMnp2d7bzdbimjyoh+aJQQgtHgnDHmfDrO89w1DSlLQmLTnCilMTpc98aYfugQtHGEAAqjV1MuDlHADYrLgNVpGHq8AQBW2P2UUoxqIIQ8PDxAm1iWlV+cMAghoPXe3t7yvs3zLMZECJumue9GzrifrJDJApRAt3kkwzDmeck4S4wyztJ1eFaMMYREeEwhxtC2Z22UlDxGT0iIkSINQYcOEDApixDE3//pn//H//iTMcb7mOf6fD5rYdbr6nW/6/uOcSYEnZ1NKXgvrtcKikjALISwaZoE4+d+YIwppfRlQAJhlEcXprkv8iKEaD3pxikRsl5tzqfj66mNLszD5KwnxJ8OJy6VUKrtBzs7a/00Tsk7vdnEX7nSbu/vpDZCGDuPRaFi9Bi+No2zc2GzyVNIbdNCrRd9dLOz1E92JoRqred5SikOQ1+WheDMWds2Z/DAT0/vhmFYrdZ1fSxLGKGTsiy8d1orSmlIiTKOfxKlAG0IY5IySpLRymjVdV0Moe+6vCgma4WU0TmtDSGEEuask+JipXeVAyFLeP76UterEFPXD/f39+emlUqXVT1Nc1XVRVFeD9Q8TyH43evr+6fHzWad59lqvd7vj6VWeVlWeZlSYolmWRZ94JIpLgF9ICyhDsa2iTEyRhNJjPEQk59mVJbA68b5rJQaZ4vwX68rLSUMMIQQGKsHwC2EOE0z2Bq6zCAax6msKs55TCQROllb13UkdHbOeseFiClKKhllCT+RiOAM+tQYXN816/VaazHNY17mk52011LKrMguNUpMb2+vqCbzPOu6VghOyAVONyarqjqlxKgwJtfKXA3KEGiRpY3jcDjs/+N//I8//fRj2zbTNP7yyy/H47GsV8fTiQvZdG3V18/PX6WUSpuuG/pupJTv93uts9OpnSaHMuuXX37BeMSrqkVrvd/v0RDHmEgpDP2olFJKpphgmghKQ3I5z3PbdsCBQxeuPHZyv0XfA9LDYbbzPFPGCfGo8shiOywWKbdSKjeZd05w4aytinK28zgMJCbUVFqraZwIpb/88otRilK6Ksv1eo0hUPDkwC656mSumj8QvOpXU58AeeHBoFSEYSoiudKqDW1wAZoBFLNXtwC0uSOdxxcUzeCRpJSYAYtigtJEltnx1z4Ma633kzEGEAH6MAgheZ5Bbgx6BvdanuenUxOXKSK49FE+g4gGFYxGbmwv4Bu4vpH3gZN/fX211qJpXi1e//FXhjaozdOl2y6llBACN5vN+XTSWrdtp42C5CkuYy0Qg1f1qu26q5zDOUdTMsac21PTNJxfpOtApWCyiY8JvgRlOHCM69gAsP15ngsqQTzgV3PO4xxTitenCUocaihrLZiGa9hAahmXmckQSt3d3UEYNo6jdc6HYPLMZKYs8ucvnwlJf/Wv/+3u7W228/OXL33Xnk+H43Efg/fOoZKDNQuKUSmElJIzAnEWnuPhcDidmhASGEIQBohn/TgywbPMpBQZY5vNhnMGoI8Q6r3b73fH43GzWWutOGdX5bRf5ktf/oWzFEL0zqaYYkwpCS5ijFTw685HAhFjFItDFzBrJBxYdhAqyF4BPGLz7/d7/C5cu5ASAPG45kPAYz99+gWGHEIboHAgPKB9uhJOeFj39/efP3+Gy1vbtlVV4XQYYzADGagpsBrAoXGZdgnOH85OIYR56axKi9Mw/GkwoQgyQuz2PM/BxGL34tTgpmKUUMqc81pTQijnoq5XlFLrIkRrEAq2bQtZIAnEaO19kFyMw2i0me3UNW1uMkr/olmH+DXLsisvBf41hIDXZEs/ASh3tFVzLp6fn7///vtvv/32hx9+gMQxxui8t9a1zfnDh/eUBimZ4OR02gc/o+0ZywXPXeTcaRmaiaWDKRyUIJDhwSIohCCVYpSV3scUBRdt21LOhZQYggB51ZXAQIr5Lw0P2BwX5moZygrFDm4ibNy+75HhdkM/+sEY9f7du6Y5D31HSCqLoigKmkjf9Xzx1o4xJkpxvYKgJwu5CtgKB+PKaVNKp2l03gouGKeznTeb9Xq9KoqcLLcGdKVccJNlxhieXVT2u90O1DpoEmunpjlnWdZ1nDHadS26/BijIfiyLKHrx38yRgHjUEqhJb2/v395ecmyYrPZnM9nMGygGbwP3lsoUiBUp8vcUyA/yO+urP5qtYL9OL6uR1prDXkMyG2UBbe3t//5P//naZpwSmG5cW3nQX4EXQB+hjHiPeecF0XOOT+dzuAYwmLHhudLEosxSiVvbm8+ffmMA6mlOhwORZZprayd+r7FaXfOCyE55ykSILPkV77zOP/gh1BiwsJBKfXh8aOQvGjO50ZN08Q4UUpRmrS8qAPwlKFXYYwpJYYhDUNPKbF2cs6iaQijC65D7vB8+2F4fXvLi7yuSsoY5ex0PocYHt5/HPrOGPPTjz9sNpuh7w+7t6Zp3z094XgDFL2mO3VdB2+dn6QUlBJjzH73/E//9E9tN9ze3WmlumFYETJOUwjBh7Bdr4oijzEIKUzQWquua8+nVkrNOWvb5nQ6Pjw8lGXBF/t3nHPsc855VVVKSu8sJQka/Bgjno53HpsH3AOQcbIoyDEoKSyunFh5PCN4fUN5td/vtTZo8ijL8vX19XQ6bTYboLIobX+VC86QCzJ5Udnmec6WKIUM7KolgRhpnqenp0dKyWpVLwpROk3UuQtKjrQG2RVdJHM4C+ByrbVd0wgh4HHkFr9ORCZku3jlsMjocSPBAA3sl/eeMZoS8SFifQilMDQrufTeQ6wBFQaOUtMN84V58nVVZkaPQ++9G0eChAbmfXBmnaYJPphIfUIIT09P0I9cpZvA9GDY1/e90vL17aUqq812E7yvqpJzdjydYgyH/ZvW8ttvfr9aFcH7aQ5XlzDE4NVq9ec//xluGdjk6HPE0kFZA6u0eZ611uDAnXM2uLwsEACsd5RRyi8UAJRLoNaQuf4G4YEuRiicc0KZcwHpM2CNeZ7ruh7HEetCKEkxCsFISsFbkgJjtCqL9x8/lFX1/PnLNE1KXFpjyMJYIKH+NZOB7EBKiYkCKJa99/vDTmslpey61tqJkMQFY5yFEJz3fd9P8xxTnGertL5GshgjBBuw4QPR//nzp9vb27ZtYozD0Dtngbc4Z+u6Sik2zTnGKKUAk4y/ixf8wx/+8PPPP5/PLc4nHhJyXmvnaRqRzeFJTNNUVVXfj4QQdCPjqKeUQB5e5TE4TmiexJ+oDEDw4p2jTLkeOWPMZrM5HI7WOrKMdMeH1VrP8yiEjDFlWY4EP8Y4DGNVm6uNYIwxxdQ0zfFwhKsrbmqEW2vtPE+JecYpoREFO6PcuzgMo3Ok73vwpSiEUUOUZYkmary9q2DGuhnxLy+KmLzUytop2Is69krICyE2m808z+M4xBiMMcB5+r5brWrG/jLiCRcc+Mmu7/7u7//04f27569f3z09+hiss//4v/+BUPbu6XHeboZhmMaha9u//9PflkXx8PDw9vYmtMKVgde8ubk5HnbaCK1VjKEoqjw/393dZUVxbs7v3r8PMQ7T1LSN0cZ5hzcjhNFaWTcboyklQkjnAqVpe7Phgq7qKqbU9S1yYWSvqGKhC5qn0c6TlFJIMY1TSsm52c6Oi0vnLXhsvENr7ewuduIomhF93WIGjmoVK1OWZQgXr/+6rj9+/GiMQSMLdh2QTJxEzvnd3T2J/irlgoi5qMqUkh0mNEgfj0e5TFtKKX7+8ktZlFwwLmiRZz54IRkhCePlfx29rqEINStOAVoyx3HEnLibm5ul/gjGGNjhXdldYAAICRBNAGApy/JwOBzPDQq7QMipaYXg949PxhjBOF4BjRR6meBbFHmWmbZtjof9/f0do0kpUWSZ9xELi0vPOYdoAaYECh0IKYGkYRgMIQS3waUrJVN1XfZ9a+2c53lKcpqm9WZFUpqU/Nf/5t/c3m6rumDUD+NAKTsdG0AaDw8PSLauOZNY2sjo0n2M6wJaBrFMGFRKKa2HoaeMee8SIdroEGOWZ26aoQKgS4cNqsnfIDygGIHuc7XeCBEiSfgmJHTQF+ELFLxRhlJyPB7hM4W0BYsbQiBC4sIFrKGWuY/GmP1+//z8DJN0mJdhXAlqQ4z7WMKpiDEwRsdxeH194YRzxrquq8pyfzyM40gZ1VInlq5GEQgPfd+PY//09M7aOUbf912WZVqrYSBC8Lou27br+44QMgx9vCjnsqtuB5cyYhsavNHADImIUspaF+E6kFJcZpcDUrgsrhD483A4wHeac47WVuQFCCd4nGjNQ/mvlHp8fPzy5cu16hTLQIgY99jEYMZQ867Xa8yOzrJMSkUpa21/zR1QKdf1WkrZtQOl9PXtlUuBbKCua07ZMAyH3c7audoYaycplZTcGNN1vVLK+wD9DCEEqBHgmjzPh2GAHvHjx49XF0mkVIgTUkrGZVbkzqn+3AshUO7gFkNjhPfuu+++HceRcyYE6/thGPppGrGMuF+wDt77mFIgqTmdUgyvry8P93fW2dD6vKjev3//9PjgQ2CUdm3z7unxp59+hEIGpwXbe7/fa6V4IlLKEB2lBJegtZZJobWa7WytHaeRnmnXdad4klI+PT1N0xyC894lEjH+tqqqrusJJXmeee+EvBxspIFwPrhyAzFGTkny3gbPaT72nVJKF4Wb7VU67JyDYQkynrYfTqfT7e0t8hLsjWmanp6emqY5Ho+Pj484cUVR5HnWdT10gNfpkrvdDkEdRwNBfen7k865Mi+KohinGYXO+XxeFRUe3AUYjNEYwxgFGnxJPlLEpROWxl38mec5xg9cJaR2mdcEcBjVNtJN/DuldLut9/sCnxpdZnTp86LLWIjdbnd1RQw+cC5TitPktM5iDIlQCNx/rV7FhowxpuB3u533/nQ6KSUBEM3znJkSLXiQqME7DwQVPG6Px+PHjx+//fZb5xxQINzakLTixT9+/FjXl65eY3SM0XuX55niIs/M4+Pd4+OdlIIm1/XyfGK/+N04jvv9HqLYuq4/fPhACLnGQrKUa03TdF1HlzZ7iFxgbuRjwLh56xxjsNKgs7UXHkJKLNoV9f0NwoO1tm1bGy6vSCklKV1xqyuMDlEdYywF76apjaEwpm+aaejLPINf2N3dXdM0JCYAKUopJi6TIZDYAnx/fHzE4ccpSksHssk0YxHypBC8UjLGMI59iulmc0sZO55OoEOQs6eQkkxXhRxEosMwCMGtnR8fH77//vvD4aCUFIITQoQAsWbnmd/f3+/3O8C7UqrrZC5M9UF2g+QIsCwBLiwEFAt4kPv9HmcYuV5YfD1xZvRS36A0AWGgFo/u8/nslgEs+Ph0kfMiRIUQwNbgasZvt9ZiLtXpdDLG1HWZZabvB7TM2NllWaaULit9jXZ1XTsbCCEvX18e3j0aY+DJWhVlVVX7t7e2bYWJIXigXpvNZhgmpbJhuMx0wmmH1YxdpnpUVfX+/XtjzHq9Rgt3sGGeZjBGSDvSMOS5uSLO0GsOw9B1XVEU5/PRGJ3n2bUyIyTN83StpnFU0L2BH4gptV0XY3zZvfkYhBTeW60lLPayLH95+fDu3eP/8e//GCLFBTfaSyf229ubFGJdVmgKh9EW4tbxcMjKSmiVGJVa+xi4FC/Pz965jx8/cs6GoXfecU4pJSGELMuPx0Pfo31STBNmbmtnByHEarUCSoCMQUrJoucK30nGKOd8jE5IHt3FoQCrCmwnJYJLCt3FCB7yMoJFYsOQhRdkjBVF2baX4SL4vl+m2cDLGqwVMqeqKjebbdd1o3XTNAmplFKJks1mk1w4HA5Y7RACjKGKIo/JE5IYI9M0KCUoZUJwqeRVtIq7vq7r3W4XFw8xRCZK6Waz6fsePWXX7m5CSNu2+/0ZGTSiCLKWhZ0ep2nCiF/Uptba87kxeU4olYLmZTFNk3UuLwrg/ig1AHtCnTxNo7fzOI5D1wZnx74jIazr2geqtb67uwsh/PLLL9ba0+mEPXx3dwe87k9/+tN2u316esKw5Cv6t1qtTqcTlIcpBa2lc3MITggZY+CcPbx7zDJTFJkxSimxvam1oYyRIs+/hnA6nY7HY1VV2+325uYGsPM8z2hFQgqOBUQfJR4f8nIppXV2mEZjDOXMQdiptRAikMSX+Rzs0s7lfxtwCQlUVlYgCfI8V0bjugRtwjnHFLNpmoRgbiYY8mztzBg1WudZVteVMuZ4OAohrhKLsix//PlnvwwXxLZDVQFg9Pb2tqqqL1++wKNYCMk5uNbeGN22jfdOKbXZbt49vZvn+fXtDZG267tESX1fZzrDEiPfWSgN4r0rq/Ld+ydjdF7kPrgQvPduf9jtD7uyLMuqcN6mFBH5cUdfgV2EupQowBxCCOpQ6C7IMtoaGvCrTgkRBQwBHjlmiVy5QWRegALmebxyzlc8CuUCmEYAox8+fGCMPTzcU7q7KrhDCEjet9ttUeRKtVIqa11zxrCKyyau65pzOc8z3hvay/M8P+4PaDMWQpRleT4fncMwmQAu0ZgMSdy1KfJ0OuEzIsPC3nh4eGiaBtyDc25Odnuz5a3s+sY5F6ILKQrxF4shpBqoF6WU7949hRiac0No2mzX5+Yk1UWZjowJmS8QYSklPHaUVn1Pmqbxzo8xnuip6zqUAkLwm5stIUQpyYVBeGiHvtZaCCG0iiEoIdvzwHjCsCMAgJcmyhinaSKUzNZicALedtM0fd9RSpr27L2rqhIcu7UWXikh+O12MwxDihwNJdcNY601RkdHYgwwLkXOYa2V2iTyFwMbvQxgiClxIRFisQJXK0ayDJi6Bh6xTEpAqIMQ3HuPjqUr23/t0yaEFEV+Pp+G2eZ5ro0xxlrvlFLEx7TMeoIXZ9M0lLIYYt93QnBARkIISgnuL9gHYBAI+VXPplsMNp6fn5EFq6JA3vP6+np/f8+WOYDjOMLuExJwLBEaOAC1X4H+PM+lVh74O+f1agWdC2OMhsCW5s0rhMsYy/MshBWUc1KIFKOUIoYghEH1gGVv2xbLhTwYnMTLy8uf/vSnv/7rv4Y5yu3tLbaltXa73XZdN80d6FjcxVVVzbMRgud5llIIwUvBKU2MpRh9TJ5x9uHDhxDC6+srjjaAOIS0l5cXULPXyIqwsbQ9RXwiIaXvu2mew9LiGmIkIdBfJfRXSh9J7b80PCQqhMqKvAyebNbb9XoTUzztDilGRimwPJ2ZelUfm/PkbOSUEBoZOzXtZnMzTS4EUpg8kpR8GPreaDPGWFUVZWzsOx9CCr5pmu+++SZTmbk3eVkUdcWlqlbr1ar6+vLsvCXReztznlKM3kdCmM5KobLJhXGa1zfr0+kUojPZykd1PLjoOWWkG7u2azNjZkuGsVWCM+IFU1qqu9t7kigXijJZlKt+mIWwbdP03RgDaesuRcq5JIRRmgiJ0zRIqeu6DiFClzSOfVFkztmiyJ2zUnL4BqNbB4eBUjrP86dPn6RUZVnBcgDRHvYvTXMYx9baQSlGaYoxUkoYSyE45+a+70NIjDHngtY8xvT2tvcenRgXR/EYY1FUd3eXhjKcRoQcxkS92iYqUiJc+LwcCKMhhZiE8yQvVkop7+L53NhpHPuOUkpTmsfJu5kWGRckr7LVpnbe/vs//puffvqpKtfgySEpUkqgVdA5671ljI1jr/T7rh+Gab5/fIqJNm2vjQ6R7M5HY7SPQSrVtu00DVIpI0TX9c4FIaS1zrlACCnLmlJ+d/cQQpTCnM7HPCuUVJSwsioGO4RTzHIze2eDt4M7nRvOqWz70+lEtlsmssOxW28flNLz0HZDj6ERXIr1diOUFEpKoQ7n03q10loLzsdhiDGeTqe6KF5eXj58+EApz7Ly6/PrarUKhD69e/f8/Aybk7gMuD0dj29vL9M07XY7lD5aq4f7x7ra3G7vnXOSa5pEma+261sl+7b9uRs7oKM0USaYFjqkwLhQShNChGSECS5UiMH7oKVx3kshGaGUUJpICpEtgzeQ70MwCjkAHARQPUOk55yjJKQwM6I5jSR6wWlZZEWmh3GcRzH2LYJEiEFLRVPs53myvm07owtGeJGVw26ncx2oc84RkoTgh+Pb49Od95bQJIRyLjTnTkjZd3vGWJbl82KrxxfLdEKIECKE5JzX2pzPDf4zy8x+v+NcfHj/Psuzw6nxIaUQ3r58VTrb3mxO5/PxdJqsJYxVZRlSioRY6/q+NyZbrdac867rldI0peQdRUYVg+L8+fl5bFtCYlEUwzDC7YkxTinx3ma6EkzW1arIiuCi915pdW6bh8diVRVKifP5XK1qQrkNYX843W7LzeYmz3Mh+LfffpdlhnNW11XTnhIJSgtrx74nRZEpJeaZn099vRLOkjaMmali5MEHJngI8XW3y8uirishy5gGa6lRxk1BG62lHvuxb/pMZcmnMi9ppH3ft6mt61pQEWmknPkQYiR5XgLastZb2zLGtNDOOUYYWP26qkMIPqQQIqWUMWJ9oJSSRORv0jUthGJM5Hl5PJ4zk5VFcRmEaYySEqWKlFIqlUgilGZFTimdRzvOcx2idUHM3lo3TuN+v7ezzUwGwufcNDHGoW+nyZLFnyvPc6O1lDKki2oNmcI4Ds7Z9x+e4EaptU6UxZCGYdJ6dM5WVZnnmXO277q+73Blh5g4Zy444Wnft2a7rlcFJXKaxqJdDiPSAAAgAElEQVQorPVt1wmh7x+elM765iyEmCbEZCmEhP0qWuK999vtbUppvd78+OOPq9UKhOo4jkJway8df85ZY24wlAoQGbnkTReDbqCfwIVCCMMwIzkC4M4Y+/r1a4yBsss8VKUuxRYSc0BSnHPUFoQQKHSNMYA+IavAjzkfmJBFWb28vFhrrXeE0fV6vb25++XTpxAuDOFFo+XdPM3eOyhnCCHgyxOJ3seiqG5u7t69e/f29maMadu2qsoQQkoxz8tpongpwNAPVf329hZipJyN8/Ty9mqdHcZBakkZ44QTQvq+z5xPdamUnuc5BIE8EUj6jz/+iAHum82N9x7ej0KIalX3w6C18t5ro4WUlHLrHPNkP9sQQt+Pd3d3TdP5SHNlnJ2adtwfm360t4lLXXz6/JkJ00/jbrcTQkCUgijunBsoQZv6YX9OkaGu1Up56531lDBKmJ1nxhhnQkr56dOnt7c359zXr18hxAwhCS6KopymyTkfI+z1c0pZURT9OKplUh5bZp7TsGRzKXHoKQLUeowkAu8y79w4jkVRMEpjiMBd4WjZNM3DwwMSZGCDP//8M5wZy7LMMx2CGcf+fD62bVsUJXAqKThjFJOcvPc0kK5rd7sdk2IcJsBZFzW5lFVVnU6Hi6RNyxCCc1YpmVK0s7fWzszyxSLUmJEQguwbwra4yN+HYXLOQ32OsjvGuF5vlFKJEKSz0zxLKdfr9fPXr19fnm9vb4WUjDFw18aYFGN7PhNCfvjhh7Ist9st+O3gffDOGNN3TYohyzJK4svLMwCxtm0XMvkyq3GeLcSBxhhUKkIIRihJFzV2XddciiIfsyIfx4lGxplUUmuj6rq2djbG3N7eoF+Vc4aODedclmUp0cPh9PqyA2K2359Qn03jrLTkXPy3//Y3f/zjvyuK0lnSt9boLBaL4442EIXCygmr5KybxothBIkxkgR07ip+BWbIGKOECiG00thgKGvoMiCAL6ZBv031wDlHPMBeQcGolJrHC7cO3Bz7AHuUElpVVd/1zjnOuHPuH//xH7ngkNNcb8lrPw4Iht1uRxLTWgshD8cDBi9bO12RMjRMSSnH2XHOpTZ2drOzZVnV9Xq3282zS4lmWZFlRVnWMUTGeUwpOR8l995nWWbnmaS/TLC5isOQFINGM8Y8Pj6CDPTen85jDBE/IIS4vb39p3/6p4eHB0g7rL2MD7p2AOCZPTw8qMXBUCkFXxCU5E3TwKu8KIoYvdYKrwwnD2h7rOPqYu+qt9ut1hkE/ggDeDTwkHl5eYFYZRgGWL6ADDidTll+MQ4hhECXDa3tdedlmWlO52EY6qrSWk/zDI3Q29tbURTeW0jF69rAYff19RWtoYhM+KUAQ4BQAQ/FSgIW/4d/+IdvvvlmnmfKGKNMcB5I3Gw2fd8yQqEGhvUbTik4TKzh8/MzluUq2SQxMUpJTEPXex/ncVrVm+hDotQFRyk9HA7fffcdouk8z+Ps2TD99//nb/Mi//jhIyEETieZ1kM//f2f/rfz7vb2drPe2Nnv9jt6c3M8nvt+nKYpywoUNNubm3PTALPGZ1+QffGy21Eh53Eqci1NttreCG3GcWZUCK6m0Vrr86zkTBa5WK+2p+Yn1L51vYLDfEpp6voQAmN88RMg3kdCImf0yiohdYgxDm1LuRCL37VSCsg4ejIYY9vtFhAo6lfn/MPDw353/OWXX4RQuIuXVMZdURecshgjI5dpyeDqYrx06mDWy+Gwv7u7hbShbZu2aTlX2ITwtYSLHPYDMCK3jFk9nU7WXgSmSinkeUhlIFDknK/XawirfvzxR0rpy8szWl+Rg76+vm6327Q0T8BfNqUEEmW73YYQGBPOBWt9CENdr0NI2f/L2nssWZZcV6Ku/cgrQ1ZVlgCBNsOAnLwf5L89sx7RntHAJvjIAtBVlTLE1Ue57sG6fhDsIVAxSEsRGfccF1usvfbaZc2YOJ3O1vqqKqepbxolBFdSUUqXyyXis/kdaZ7z2rYtl0Krsm6bvuv78wCnwjhdLBaXyxmUgZS1SvGyRVGUZdl1PXA/4Csozh2PR62l2R+LomBC/+sf/uPrr7+WUlqffIqqLFJKMiXG+WiNP8cQfByHsiyZFEIpKjghSWiF9ZR5QCEenmT1Q+Rq7I2kZsi6T/Mu4xt+BfeAz+BZ/2S1Wnnvz+czZwwGcRxHXRaoRnLOhODWWK0K51xVld55RighREoJpR1oSIDiDZYYY9dmcdg+a23X9VkPJyqllFJlU7OsCgnqmy4rrYqnl+e7u7uyLNfrdV3X0IV+fX2VUiZC4FqnySdQpwl1zl3Op5ubO7CAcXO+fPkipfTe4RlmNU0hhPcuEU0SAVxDKYWeNnwySBd3d3efP38GUgm0Gl1+4J9Bl4kxDgoW6qgySyV7b2NMi8UKRr9tl5vNjfehaZpClx8/fowx3t/fD8M0i58fj8e+7+/v72dhSCllXdcYlksIgRgkSFAonqNDG6gXUBRA0tYaYNaAR/fHHXrfUGdCVx3JGHeMEVVchKufPn1CVw7u6n6/B+cqpgh9b/TgvL6+ol8hxXS+nEtdKKXMNDHGlLh23QNIndt2pmmCsynL8scff2yaarPZoOdOCPGwfjCTFUL++c9/0VqfT90wDNaYaRqwd/jvCK65lFTql+MpvO4+Pr0g9y2KggYnhDieT13X+RAmY8qy7PqBkDRa00/j+Xy+zg4iqSxr46K1vu/H/f5ojMEkVMaZzuNsldY3NzdfffXVdrsttCaUCmOsd9M0hRStd1rroq44F1DbL8tqtpuX4wlmhVCKmZqg9wdGKKXn7mKMmayJJI1m4pzjX7Fo2F9wXSC655y7vb09Ho/w0NZZzvli0VprKeUI40IIRaF9lnIjuZcCW0wJm0nPzl11ClAUHcfhcNgXRbHZbF5fX3b7vZJFdiQRsS0qLujVGobh9fUV/Oabm5vzuRuGYU53cKiGYXh8fATFbrPZgMSIuOpPf/oTJAtfXl6apkGcWmiNtlDIvgGI77pOa4V7B/oW3quua/R8nE7nlMjMEKGUVk2NGLSsK6kVYVQoidJuVVX7/b6qmpJXSildFk3bPH9+6fu+risdJdyw955zwZigNEqpOZfnc3c6XaTUKRFkLaikWmvRa9L34+l0lnIoikJw9vHDk5SSUiblVRMFZDYU/1gePYKAKcMJ0BgXsMxz9jAbfWQVSArxhSuGUAPLErJC1N/rHuAMZwoaamJt29ppgG1CAhFiwOFodFXoghIM3RYwLtZaH64q9ohZ+r4/n89jlunGy+uFVkr5EKSUutCUUsY4+nQKJZ1z19itqMqyfPjq68u52x323vsY0tOXZ84ESVRJLbiUQtVtY5z10hFSIES6XC5SXpto/vCHP8RIhJLWWvBKf/8/fmeM2e12l8sFDqZpGmunGEOMCVTId+/eHQ4nQsh8l1CdxnbCk6H8hSD95uYGQTEhEeSHlBLU2798+dJ1nXPGe4+jBkiKQ3nROiGkEGIYJoRIEK5AxwkMLoqBqJiVZXk4HLBEqD2sViup9G63u7u7A411DjmRSG23W0rJ/nW32Wz681WlnDEqpWQcUwJ90zTGjKh9gVDRNM3LywsCFmQPDw8P5/P5hx9++Nd//deHh4fL5bJer/11nMNRa304HGDIuGDGmKpawqUtF8vJDDOTVSk1DEPf9zMrjHN+d3c3DB3N4rhPT0//zzffHONp2S5+/M8fRz8QwiB3CEttjPn8+fNfmQiJEi4W63XXdafLZXI+pTTYy82i0WVl/XOizPrQj5P1YbR2mibGxMvrvu/7190BwWxMxDqHPizYHcbY6XTabNbF1+Uvv/yyXK3vHx7bxaKs6qKsGGWqoMb5RBnlgjBunCeMex8p5X3fC6Eo5ZRGSjnndHJ+HCeutPe+H0aSM3XIqE3OUkqZFMHZxGhiNLoAi7BYLCilp9MJzSVIPcEXJ3mQV/DeGON9EAIl2OssJjQSzSFkmvsxq9K7ADKr1tr76HN3PdAqIfgwDk9PT8ZMnz59urt9QHMfaqczN4/knrjdbvfu3TuUW6fJNk0D/HB2bCmPt4R5+fOf/6y1vr+/77qrkFzXdbe3t5RS4KhKa5JNGwIjZELjOBJCF4vlarU+Ho9IYqqqppSdTqebm9uUklI6JeJ90FpTwQlng5lGa2ShuZLTNMlC4xNDCMaYqqk555Gk25vb1y87hGjDMCglx3FkjILB9fr6ClY3rDZjbLFYIL+BxZumCajsODlCRSKcUGGslVJNxoUQhGDYdxWC1jrE6EPwxjDGQMuljFnnQIiQIUDMAofc5aZ6OICUxTplniuDPDLkttzZi/wK7gG1e1TkZgzEz3oPKV2jUUKUUlVVrtaLjx8/CSaEEH3fv7y8aKlUoZRWMXeNomgGCxtjJIThZFBKnXNFWZrdi1SSSRajhyfEoWeOzIBajLEsy/P5/PNPPzdFY4wxxmFUHCEsJWqt88Fba9u26c5nay2mtSDWmKyllC9WS2stl1c5LZBnEA0BTQohTmbSSkOuADGs1hq+F6gU3APPqu6I4OAnEFqmlEDtmJmpqKnEGLUuhQhFURVFIYTSugwhSSmsvTaTV5XXWsd4lRlAVGWMORwOaJXAkd3tdvNcdaC0nPMQ4jB1sLM+q1NM0/Ttux+UEtM0HY+HT58+oYG5qqrhw2fkE5vF6ng8ck455zc3Ny8vu/1+f3NzgyMIsB5EKUza+t3vfgfcKYTgne+6DroF+B4Q6k/Hc9PWzliknmVZ3t3dxeSHyb/Nt87nM2hdYPcOw7BcLhBYzNcP64BpejGCzsFC4CnP9fzmm2+enp5eX19vH+8LJQolaF06M1ISu75LKen7O8bEzc3dhw8fDofT6XS5v79nlBtrnHNFUX7+/IVzPk3mcukOh+Pz67Fpmsul17qUUvZ9b4zbbreMMpQHwWLo+/Hz56fgveCCMcaYOBxeGBOfPz+9e/cu+Ng2i5fn174byB2lhCkpKSUkUe9CDIkkOvTjlehlroxbmFpg91f6sukRDAKCQ7ACvrxS6nA4rNdrfm1xdzkatafTaZrsarUehmG73dJrz6OBXSaEoBxSNrWzHtUsPACKJSl3MMTopJTTZIahJ4nMtECEsbNoK1bpfD5zztFCiwOP6ASjlpbL5c3NzRwd41Ng7MDPrqoKoOjT0xNe6qoZ7hwwW5x/oA6UMDO54+EspV60K5IYSYwSbo23xguhQggpUin0+dRtt1eBXgwuxclBjDWnQViiwBLljFBijNFanc6nui4B4xgzvX//vmmWhBDvg/eeMeF9HIZJ66tKK+zAbMQIuTLiEOnSPCF1plQhhqNvhgXA4CBvgA8OIXSXy5RH5iC2fruMqFjgh+CPkG+YY3HQn34F94BEhuQCzpcvX641FvQG52kB0zTpQjNMfo5RKDGOYwrEOVcV5TRNSitoqXPO5/JI9A6NgUhQ4HtCjKfTSZcFiYnShNERWEdIuxBCzufznfd9N3jv66ZGkQCOB6mZtbZuaxccdghvMQ7jOA7OBUBAzoVECe4YkgAcFJ/7CQFQSMXhcnMgYNC2xt9M3MRaIwAnWVhptVptNht6FTzQ6KyZaexYW5IJiIQQwMdwPEJKznhRFON4wn07HA4wCnOkgBQbk+xm4Gi73aK2XBRFTMTHAKgXyDW+bbfbCcEgiKKUqurCjhOye2uv06eFEOPYn06nqiq22y0mL+G9+r7nuSG2KIoPHz4Ao8D79uOEzm1CyKdPn5RS3377LWPs08fPRVFILqZpSDEClDscd4vVDaiBWHMp5fl8xs3HcYLOBHDh8/n88uXJWjdNkzdWUHbuu3EcGWfL9RKAxpcvXx4eHtq23e/3i6ZRgk9DPwxDf7kqu03jeDgcbm5uQoohRR+8m9zCTJO1xtrT6cSlTJSeLhcfI+Vcao15v999990vv/yyWq0Am4yjKZoipNiPw+t+N0yjUDKSdDmfx2FUSm02m5fdqwv+3F0iSfvXl81mI6V0WZ8KtiOk6GMYzVRVVTf0MB/DNBJOXLjKVxhny7KUWr09AHMiiJODvNxnaRZCyH6/f3769Nvf/oASOHqYdrvder2e48q5WwKQb9M0ZrLABpVSwCrQega6+YcPX+qmWi4X1k5lVQJmAYg/99khQZ8ptkqp5+fnh4eH0+kyDAMsRlVVmJt2uVzQpoNDa4z505/+tNls0O2FkhvEteDqpmkaug7ZMIaPQl4l+OichzA4yyoX4M1DgBndYdvtdhiGoigZZz4OkSTKmbOWMSaUbNoWD0z6/nK5XPpOCu2C3+12r6+vSsnFoq3rMsaYUpRSWuu6rkPtAZAsMnWtJbx1eDPgAQuSKPHeS61Op1MtRUix6zqlRFmWiZJ+HOq6JoyCbue978fBxxBJQqmcUsocnUNkuG341NmjwJj4NzKXUx4kbvO4wF+nND2ZCVuCj/zpp5/wx7rUIQRQH5i4hg/W2skMnItxHKWQ8ASMMU4wnE+THDgjF1NKlaUOIcFeU8KRPDrvhRBDPzRNBdJeUxZCiNPJWGPLmggh+76fJkMIQY8M1gIJCqW067rleokEdru9+fTxAxoCxnHUqhZCWO+VUqvVyru4O+wRX7PrNEEOZDNHEI4zgRsIQjScP2MEgnpA/GejPMtaILssy/L5+VnrEtkDcgtEHzg9dV2j4XmG1ymlaB50ziGVYUzMhUfgJ0opdBHilKBfjzGG10QhxPmY0mCtRREPzsBa+/79L1999Xi5XBij4zhub9Z2nI7HIxzVarVSWgghnp4+13WNxjqEkD5PecSzIWwEp9Pk0S7TOGH1oDoOHMxam1L03jd1LSWnhJzPQQghhJzLkvA3c7I8N8eicIKGsrqqhBBS6qIo7271x4+fuu7Kiwfe7Zzb7XaHw2Gz2ZRleTocjiSmRIL3RVleLufVahW9x+D7oR9ijOjlBvEhCFEUBSSUYU2MsWVZ4dOh34cwmVJqrI1D9JCc5Rz89N1u150vSOnqukbgzDl/fn4e+x7Vo77vn56ecHXhLSil4zCisGeMAeCGYCiEgF3GNyNbfcsZZVcFeArOK2P0cDygWqaUqqtqGAYlC3ad6qEwRfV4PMD7orMMkbu1VldV8AhCoS1BQXaAZUFDGSU0JTIMo1Kqrmuw/mMePvP58+eUEvBSlPERXaWUlssl2vGOxyNGXSGm3u/3SLvRd/np0ye0/Qshfvvb32JGIXkj2jqTvkju2GeM2ehfX66FLqw/jMOlu3AhALI573ENj6dju2pzBXRQShpj0OvunEMoFiOJIZnJYVLJZrPx3hVlOY4juRKHkpTS+wjLg1ZwWDmTBypQyjgXIQRjLOeMi8QEpySlFMaxJynElAi5igbhJIMPAluEjr+Z2QjYhnI+g0hzTDzXcnwWa8GbIp/DwBjUKmDofh33EFIqyoJJXhTF+XI5911RFrJQIVEfCU1RCMI551TY0cXoGafRw/Mxa00iMZAg5DU/AJpk81y2FEOMkUnpnKvahnPVLtqqqp6Pu6qqSEx1XXeXMxRXpmnqOkOpdjZyzqD12LRl09T7/R54NyHkeDziWDhjCyXbqvru3bs//q//xQRzzkldFWWdKJ+moWma4HzwXjJalYW1nhBmDPjdJKVASKSMFKJMiRJyjWTv7u5++eUXyCuiBPry8sIYQ3KAfQUiDNwf9Fw0oAEEwOY1TTPlr9Vqtd/v5+Coqr5LiT09PUupVqvV4XCC4xFCNE21Wi12uxcp+W73cnd3t9u9CMGIILrSXAnCWVk3qqgIE1LR7ebm6enJWc+ZOB3PbM3bZuGtWzaNtUYWitCIWQtd1+lCrDfLstJQebpcLloXVdV03Ugp51wa44qiSomez5cYI+EiEno6dw/390oVj/enpmyeyA7QLcuStIvF4v3791Lwse9u1qtp7MuyXCxW/WSkrhQXLBH49RBjWVXL9SolAtIaY2S1Wr28vFRVo3V5d3dHWGGd2+2Od7e3j19/6yOFO2+ahjIaI1GqiDG9e/eu7y/7/ZEQyjkjhDMqu8vIqBrHgTP9+Ptvnp6eCl1Gn477Q3e+pOCXy/UwmJSSMT6E0LbtOBrOFQr45/N5tVoBwCmKYhrHGFSw0Y5upGPXdcvVqiyKzx8/OOeGqmrr+r/+//+8v3+4v3047I7WTpSCYWJ+/vl9WZbopBGU0RiUFnYcx64jlPhCl1o9Pz83ZaWl5ITFRA67/VePj8Nko4vDZaiq6nK8NI+NCUZQkSKOa6KMeG/hocuqlOor9GZypf/yl78kRsumet69RHedOTrPSaaUDsO4itS56FyKMTkXYoxAilKiMRJjXF23IURKZIqcEZlCYIQIxsa+D86llO5vbxkT3hPGVF0vP358krK4vb2Tsgxh2O+PDw8P42is9db65bLebG66bvjLX34qiuLDh09CiLpuKeXOhe327nzutK68d9a6cexub8k02tPhbE0IIWDYUQyUKaUk6/uxHydZlJ7Q0bqiLI1zg/WcicVi7bybfIwxjqMNMXTDpLUKgZjBkUAZk26aWBI2pP3xUhQlpdRMxnvPhSiVFrc3yEg4594lSpOQ3LkIdSmkKeAISCmNsYxy70gIQQgefKBUBJ8oDVKISCmnTHERnNdaR+eVlDElOxkhRPQhEioYL4uy513f98mH4PyqXfSXTiyXMQQ0/QGEnOFungdCAP5FPoESFP4JQRhi01/HPTjvmGPudKqcSymVVUkp7fs+uQC0AVkkIYQzRonou+sY0VkNhuYJaPANgCDnBDmEkBKJMSqtpdSUs6IqV6tVVVXBeZDwLpeLHYf1erNebdu25UJ6jzEdE3hd4JAhY0W1Fn+jlHTGl0UFxjqjXMmiKCrESnMihsgF0xPxT23bAn5JKWpd4FMg+QB13NVqhdfB3yODQ+bRti1iQ+fc09MTCobYRefc/f39f/3Xf8Gv4INQl0MijAQC3Z4xprd5OvJuEFLx4gh7L5fLx8+fEieMcUIJ5zxJogvBOQRLbFFqQtNi2Z7Ox/PldFfeBW9j8i+vzwCL/+Vf/uWf/umfoO8EjEIIAagXWWPMQvNd10FgCkRVIBKXy+U3P/wA1bNpmoL3mEIKqHqz2aDsXGg1jSM0M0RWGcMqAZgGsldV1WSN9/60P/ks14EHwEoyypu66C591w1aB6QXOEVaKp4bhsuyLMsqhL0QMiVKKY0xTZP1/hBC+M0PX//www8hhF/e/wRY43K5xOiFwG5UgCjnrb+/vyeEoGC2Xq+Bw6C/PYZwPf+cn0+naRzHceoHqEzTGOPr6w6TBMexRyiHbY0xnk6ny+XcX87OWUpoKH2KkTEWnBdCSC6dcUqouqzOzgfro48xXiFmgJZAa3HzrTdoD5JSFoWOMaQUqqqSQjHKldSLdknnL05SSl3Xl2VlreNcEBIIIefzxblrmAnSHZimM6MmJSqExJSn1WojGGVU3N7c//nPf06RPj09vXv3blE2qJChT/Ny6W5ublFIAFSFxOX19fUtGoaCPEwevc4lLLuuKwrSNA2l0+Vy+fz5S1WWzlgITTLGi6KUUuHqwclJKZ33lDEQvdrF0hgjlCyqkhCSUjpfLtZaJXjfD8aY8/mCPWWMpRQJIc55ra/2AY8UY5RCiyzmH7J+rZQiBA/gC00nh8OhrmsI1bisUIl7hM4PQgiqazjYOL2UUkYIfh9DiJQSQoa+Jyn1XQdC5kdCyrI8ZU4aHgymYLPZzJRWGFWSy4So/wsh3iJRczX773UPp9MJYzWBXcwSuzBwh8Ph9vZW58nm43iF13ENZmcLzBrcUKwIliw4Y62RRfn4+PjVV1+9vu7natgE3XlxdXRgXzRNTQhBPSrSiOPLOU80eO+BDwA6n6YJ1AJg/bAvVVVhboExFrxVmrtFlFKUcmhyxejbthWiVEpNJry9z865pmlQzZvJA8659XqN2AH4A2ohML5IGLEfu90ORYuUxzngbsM/cc73+73I81XwG3w0MIrT6eT9AyHEe397ewtRP+SYX758eXz8CmYLssckgT5wLWwCHsW1h1QJ2B04+qASwVOC4whE1eeRfGjiQ70UKjfGGNANQwjDOF4ul6IsjDE+BJYSTD9eEA29GOxBKQUFGaCZ914yDqoC6tJoLMeqLpdLbBAYtDPPHdXpOfvOJdO0XC4JTeHF932HEjdSaViceTgE59w4q8siUZIIUVrf3t+N46i1RNcLgixsEKjr9/f3wKNAHhVCoBsL24q6V1EUM0cLUxkOh/3Nzc3//t8///LLL957QiJj7OXl5ebmBlKjP/3003K5cGYyZsIuAC4ghDw+PlZNI7WmnE/WGud8jNZ7yjnhbLSmDD5SwqTgQrhpooxLCZF2RiknJDgXOPdKEpdFLNBRD/ElxDS1tYSxRKn1XgjBpaKUZdeoAaiez2ccThgdIGawNVIpIdhl6FerVdnUT09Pk7On7lLUdVlWkAmabQU+GporMF5fvnwhV/qsg8VANIlEGfrbCEoopXUW3kiEyEIHklRZREIIZ0pJpZSdjNZ6t9tvttsr54VRznlT1TQRJSQ4RTFFyYVL1uc+sp9++gl1cmB3uCY+K2/iGsYYSbpCWwC1MkRMfbiSRxG54z9m/JnNFX5cH8qu8swoTc8/cI4OpZQ4TnNwtl6v9/s9IQSl367rYMrwGEII6MoAkJzLziE3DwHymmvgsEt/c9MDIeS//U8AiOj+AJ4OyUmdR7uA/4Co8HLpsBbr9frrr7/GqZJSKi19cMvV4uZ2W9UlZSSRWJSakL9K5r2+voYQjscj6PM+jxAgs1ytUqDYXzshyhKVNND7YChVnoXEGCOEOhfKsjbGUcqFUF999c12e0sIdc5BS93nHmakKSFPmMAS40S6rGeJuvFckYaJhOkhue6UCXakqioEyNvtVmtNSIK5hLmheYQObssMHwsh8rSpkeaWyHl6hMv0WQCOWBmEu8DfZH4T1XEAACAASURBVJ4wA0kfEObKUtd1KSUvCsU5lZKv12scXBh6mrX55geAP4ZfhymfzwBU9nCRIEdTliXyXMb+SnCEah6ioZeXl9PppLRGeIEKB5JxCOjj5qCUAksBW4kTfzqdbm5u1us1dgqHCgEH4rumab799tv1evVm45qmaThn88Pgks9OxTlXVRWecLlcrlaru7u7x8dH1Jaw5mVZQmcQv+JCfvnyJYRQVRXkrWyew0MIQbFkGAbY+rIsn59f5riky3OWMDJkv99fLpfT6bTfH1IWAP748SOCXxgIACzWemOclLooqmmyKV01XGMmd87BCskz/lBhAqejLCvnwjBMKVEhFCFMSs2YWC7XKdG2XRLCGBOcy7ZdYphlyjpOKN3ZrKyM6z9f+dyhycD0p5S37bLrhqKo5olv6/UaNmTKmseoxqETYhiG5+fnl5cXnyXFZmLkDKyj0osGBSwyjjq+csiPrmAKif79fp9SAtjL89il+X6FTJve7XafPn1CsQpCaij85A6PcX4ShGgpqyjO2QMyCZNFM8dxxE0MubXCZk02BLWIRxEg4kv8VYL6es1nauJ8IwAsIzACl2+/3yOEQsmBZ5ry7M/w5bKAOY4KrjnwCXjov+3r/046wCcjhADAwVvZYfJvxnajrhVjSInUdf34+Pj09DQPHmCM7fevd3c3dV0aMzLMg/BXHi5eklK6WLTDNCLfRM0WKAH6e5umDp5wLrgQcwZDWPLen89nUDiQc8zJvsrK5tjvu7s75xw6hiC8ijo+z3J4eBGgHDgNQogUKcwlmGSHw4FkRcy5joJ4FqsPy9U0DdoA8QBCXNEkVKQRRKeUYJ3RhRDyRLxZGAMNpdD2wmWAsjFczu3tLWxZokRK4ZwjhAqe8J19N1BGUMEGCwV5DA7KcrlExRhZAtIaOAncTJLJUXB7IBThe5bL5fl89iF47z99/Pjbf/jtxw8fGGNSCGPMzXb7/tMn3A2tNTw6Uua6rl9fX2Gysba4wG3bsunqhrXWlNG2bc3QhxBgRB4eHqBvMU2TVjXLQylwwDjnQkilEG1QOAZ090tZrNcbhOqZkGoeHh601lyKqqnP3aVuG1WgrG0013MMSLJ2G/YaYUGM8fX19eHhAT2JuBHPz89N02A3AYxwwREKcM4JSUVRfP/999M0ABt8eXkBO2MYhsWiRc6NaTDjOCJaX63WkzVlWYYUy6Lq+365hqyTXkjxut9N1kzWWO8oZ6rQIcZC6hDCarlWUq9Xm6enp+DjNJn5AMCMIuOhnLXLBRbQx0AIiSRJrVK4kmghnYI9Am8NRD6cLpzzEKMkJKZUVlXTtta51Xrd9X1ZXqUwpZQoR8P/wVziFoACh3ai29tbdFBdewwzbw2g62632+/3//AP/4CLwxgLMTLOJ2M45wxd3yGADns8n5xzjHMktW3bphjHfiiUppSKhkcfGKFaquNhh/MPNCJk0nneNTJj9DARKV5DcpYFquGT5ssC1RN4LJIIGG6wxcg/UhY5x8rMoT1yekSZ4BbiniKZxq1HJxalFHn8vI+EEATTMB1wCXO6gBcxWbwPKw/f9jd7iP9bsRX5kbpqPEit9evrq58sNAWv8MIw5ALyAL+HABDr0nUXWFiEUdgAGFb4NNQw27b9+PkTaCqAs9q2JemqMGOta+plSoQKXpalKhVjzLgJrhuBtrzO5IHSSxqG6eef31vrQ8ByMCEUXgrbgAQQ6Tb+O1Lgvu/7/qKUEpJLIefoHsQkFDl2u93c2YgfGDNjGkgoZjNZayHvigNBr91nDEQFQggyR5wDQO0QwUaMTCkF5tC2bd/3z8/PyNnRvI3/SBgdhpGxc1lWWhXoQmeM+uAOxxfnXEy+rHS7qMtKl5WWVCqlQIrF0YEEL14BuVHf9zMwSLNiPsAcnAGEkxM6MBYLjJrozpc612AQ1eI/cs4l0yAFzYUipGvXAIcS6Bns9/t2uUC8jzATmv7qjRQEply8vLzgdnVdVxTHtl0wRp2zZVmG4H/88cdxHJfLtdbX+DfGeHt7yxj7zW9+sz+cMWsP9aRE/tpEOgeJIXcgIo+Ep7y5uQHHjOTmtRgj2tTbtj0ej8AHhOBSSmt9WZZCyJeXl7Ztod8AHBJ0IGNMjInEIIRAzIuoXyk1ThNnKkXKqBgH012GQleUcGdDUSrEIjPbhzFGUuJcci5jnIYBEi/cudB1w+XSgxvKOa/rlhDCuaCEc8aUTJQ4WvK+76fRaq0ZIygoYgsIIShCAPkE9AeszxgjpfIhGmN++eXDOBrGxGKxSimN41QUGmZIXNUB/Pv371EUxI5TStFm+OnTJ+CHsCSoRak8JE5r/fz8rJQ6Ho8zvgSTihzrrxgspTTG+/uH0/kEU9sP/fl81lzDq2l9JZcDkcccMyCcLveR4S4DMsGRQCSeUkIlCQYNL5JSCvE6sXmaJuzgzc2NlNK7GDNrC7r3qLlSFNGlnGuKsCEox6ZMDsT3s8xXBsMF30DzrAs8IVYAZ4Dm3hEYw9kNqDw4K2RJ7Pnt/l73gKZEPBYaPeDT2rLGo8s8Omr2VzOYiCG0p9PJeUNIoDS1bU3Izfl8HoaBc6rK0jlX1E2GLDTii6vPrButpZlG59zUd3XdVWVLCMFnwd8wwTjjJFeBZq7bnA2ALI9dh/udG19pHiSHxWqaBcisAD1xVmIMyDBgpvGXCBXxWeiVwzGacUnUqYqieP/+/Wq1appGKQ31NLAVY4y73a6uazwnlB6QOAPTAJSBywDYB9lGjB737e7uDl5ZCDFMIywLWvPxXghAEIvNPxnJCkIzwKMwyqcTRuNJvCCA0bm5pixLeBGcLWSQSqm7u7vT8ZRSojnppnlyKsrRyDaQWkXngBE754DP4JQjqSd5zOrlcuFSOOe2qyXuM3j92O4QQoqwqhEcAX3VTEb4ctUxVkr+/PPPSom2Xc4ggNb6+++/3263Dw8Pu8Mf/+2P/344HFRZEM4oIS4ExqgPAQOseB54DqgHaNJyudxut4fDAdpWcCRzRZQxtlqtOOdCUKlE27TGWOfcarXEykCsZbVagVaEvxyGQUu22WwgB4KQwntvrCU8RZI8JuUNfW0aSilJyQVfVCWXYrFa6rLQZUGMcZOZgyTcfFxDpQoAg0DVEepZ50RKTHAmOCcpkgQlTSY4SwQBEzJgVJ6w/kgmcAyu09MY8zGBrR9SdMEzwaWUXPC3tBmSgQdYLux1URTv3r2rqurTp09fvnxZLK6zSGd0d551A6UNay064IqyJIz6GFzwKSVGeEoxRSK5sMZwznAlu75HO8Xd5i5674wZ+x5wIolx7Hu4B4wzulwu6ODLkeWVHorjSjAnw8FOXKejkytexOEqgHPiiM5Mobc1ACTrPlgEEHCZb60Wfg+oA//x5uYGixDzMB/kKLgvNDe+xDyJD38zA19zlkazFhzea0ZrfgX3oLVmlOqyhKPzzjVNo6SsdPmXv/yFELJarQCnZoIzn6vkqPQC5tNaLBYLmGaRG68W9QJj+bz3zjvGeFFoRsk4DDGmuqzIG3mQ4/G0aDcppRgTIVc0UKtyGAeSMTtcgNnWUHodzg6/lfHTcbm8iqfDwwM9j3miAxAPKWVZVt5bY640YdhZmAz4diklWuSwqcgJCCEArNl1mmNYr9dal4Ad4MyRN+g8322328EMwbAiE0T0DYQXN3a73Vo7VfkLF6aua+f9ol0IIaUQSinOBSXIyWhdV/I6k3Wq6xpzOPzkhZAk99nPpBEUNuD18UZA+WLmdMN9AsS/2W7v7+/7rscrAFoplA4xaqVIItZaZ12MMXjPGJdS4LNgZZCGF0VR1bULfrIWFh+YPmOsaVuEmdCbE0Kcz+fvvvvu9fXoY9BadX2PFgGpFCEJg56MBblLIIbFdAq8gtZ6vV5zLjabm8VySa5V1j6EQEjkXBRFQf115lJKSUiJLcOrgR+C5PLTp0842EprmAzgIfAZ1o6MsrpuDsfT0A+bzZYQijDrdDpBrHexXJ5Pp7Zti0IrgZpN+x//8UcpeFkWzvlpHBP3UipCCNDOmbAUsiKvUook0nVdDDGFiCbQGP8qH/k2wESQB+PrnYspSUrwDUBlM/jJ5+QGeEC8TnLmfd9xypRSwzAs2iXjnFDqrGOMBR+CjymS4IN3XjAKz00zaxFVOpLHyON5Yoyr1eq7776bpglpJepesIZYYbwRmmMopWaynEvCCOfcTMY6yxn33lHGWFGmlCYzlVWVCGkpNdbudzs/+aIoAI0i9lJKjeNw7s8zuosZEuiVwXWAXZ4Nt/ce/fkhhEIXUilCKaNUKklIwpmBd7kGNFTMkBHWU0pJGVVCuzy/Er4HtgU2h+XJuBB/pJSClAyPMlca8FRYScRtuIBv/Q1EKFiW6oNlS28aqn8d98Ai07Js25YSIqismgqQGbq05iwPeCXnPIRYVa3W5W638z4457jglFJjnFKFMW65XBLC/u3f/v13v/vd9u62qCsIEDln/vEf//F2s5yMmYZBCiEYlVyQRMui6vthVTc+RcaZnaYQggpKi4IEJt70jsNywXMoIVKMJnrvbaJRFbKf+vNwripNSBzHHsexKCoptRAKcDnoMW3bNk19Pp+t9Ug7WB7SBMkKGBelFEwqy3N0jbFKFYxBqItKqXe7A+dXWAnThmlmwYJJBWcDhwE5JkLIN998kzK7CUeWEIIwKsaoVEEI8X6C2j6jvCrqoiyFEJJxiSEwpXYhFEXNGa+rhXdHa0JVtvv9npLEGFFKINKPgXSXQUppJldXJAZSlQ1uAmfSGi+4T5GmSElignNnoVIgovOFUofdDkUbpbX1nk+GRiIos8PknUueM8YoSYSxEGJZXJvFgk9FocuillXNUvR0oIxGxpY3N5zzGDxJkdKEWrqU3DnTdee7+7vXw5mkVNSVHPrFZsUoY5xxwZ2fpKq8x4g9q3XpvZ+MpZzpsiyqRhUFF1ood+nH25sbHxwndNWuAPpba2jkTHAf0+S81kWcTCA0JOpjSu56aYFxnU6n/X5fN03dNqvtZhxGKYXSuhXCOuedY5z5RO4fv97tdvVioaS03q+KkvkwTGZ9c0sJ3d7eMcabph7Hrh+H5aIl0fWXg1DFuR9vH75SqqJEMCqcDQCapNbeOze6SlXoFaeJucmPwwDXi8onYyyEtFisvPcpRC3kfhynsqSMlVUVUkqMEkpgiOfcF+4h5NkMMDQ4t8GnyU2M+BhMdOV4Lt0mRUbcNColYwjTOBCS7DTylKx13dmTtgWGgxi2aZqnpyfQ1W5vb+FQYYjv7u4AH4EJhrIH0C0kx+fzeblYk+RioNb6sqQODeeJSsaDdyl4EokZCWeSROqNV0wSmrbLjavaw/FEKWmXy8EMPnnC0+SG0fbT5OZMuu/74/G43W7Rqo0Uf+4FRm2DcgHNVJeCFkIwmWIilJGY6qrquq6u2g8fPnAmy69qkqKzVtR113XI8BKnqixC8ikGRyInKYUAO436NmJB0JDmyigQoTlMSSlhJBd5MxpoztIoTTG6GFNRVCEESrizzvvoXVwupPeeUSE4ieGaefwK7oFSqqTUmRbpvX99fZ2mab1eIhEGngUDN46TlGqazM8/v+/7S1WVTVOfL2dKyWKxBND08PCARAmEkw8fPiAkJ4SA1tlKeT528D0hz92mhCLBnBWNCCEhxKKQqI7McBAhBEUbIQUjlHKWUuKM6aI4HA/jOBZlsVqupJQhWCCSLHNaUMouigI1ksPhGGNU6prWYUFh2QkhABwYY/DtaFeOMc2ySyg8wBmQLIaFW4E8huQBkPgNWEzozb67u3t6ekISk2GuKHJjBwIT1OEJIff391dmBaX4acilmHOMX6F/ra8Eu6IoBCUppf1+fzgcGGOCQ0eMY7sR0RRFMV/Otm1R/gJDDDXGEIMuJEAbrBisSSIJiVrKbGDcMYSoc4Uc2KNSylgDCEUJVTcNZcw5F5zjQipVcC6F6DiXRVEpdWac66yIhS24xqGY9uccUDX8/Csbm3OllZRaSsk4X61Wxthpmrhgq9UK9UOE9iFExjhESlCJTSklkihj3l5HwDrnHh8fp2n6/Pnz9uZmu92u1+vNZoNwAXg66sCovYF0iwPDOF8vFpAbmoEpQkldNzSFu/v7h/v727vb0YTBuMVy6XzywcUUAAsoKgGw492B3pCUYhbeme8F4n14C05o0zRiJ/u+P55OnPPb+zuXFf/xnXiYvFNXhAeVOaAxKYVh7KqqKMs6OXo8n8dpkKWIKTCmrbUxYAZRX+rCmilMyeV8GrH/6+vr7HV2ux0h5ObmBlcb9GVCSN/3fd+jFMdzCw6oE2h6Ha5jpaM11llbFNqlQK9MRZqyHPJc7LXW1lUVE4kx3N3dTWYYhgFDD0Nw53MPPBCNOCEEqKAvFos5iUFSOE1TVVU+Jp+7rIC4hhBITDEGrCHuL+pkIQt7zFmFKjRjzDvPOIsp+RAkv8orpTzNjTE2zz3FEcLPBO0NKSN2FvgqwlM8lZQypZDr20UIgTPmnDPGMkaRDCHVwPP/OtkDwErMtgXHHDcBNHbQvwCMoEafEuCgMJdxzDSlRJtmgZiacxlCCiHd3z8KoZQqKOXGuMVCXS69lJIQqvKkBLwDSOXr9RoHlxAyV3vUVXvuWk1KmdvKGCuUDt4na1NKEuWgEIUQdVUDQ8eLYGoCFh0pGPaeMZaVvQtUsJfLpVIKSQNScvKmPjlz4OZKDOyyz60DGe+i81qBPDqXlcQb8d4ZOsSrIdkHSQxAB4wRLhXiCGCgJIs4xRgpY8QQSqkScqKs6y9mnFJKl2nQWiOtllIqqQEcpYT+DwrLi1t6Op2+/vprmudWgrk7NyRmpxiBGYYQKKE8c3/hAt/+TJhOFFrQZJDAHCeUJVIWBY2pD9FTRhIBSQOUQWDTKdHyqmpAENiiLjI5F72HfNtc1QTLPvNwMF54opROk4VTjzGCuRvzDGQ4b1w/VC+R75txJIRA96lt28fHxz/84Q/A35bLZcxTnvBz0O2FIAAObAYtIdSBhUq5SDuZnhPinJO6aBerwqdTN/gYCGHmzdg1WD1KCFwgz8rMqNDOvEk0kMNJxBi9D1KI9WbtvbfOffnypahKiPDAH8CymFkBkIT5OpPM4HLepeSN7b/55luWxGX4Mo4dlSWXnHOupDLTBBvHOCOEUEZxKXBTiqJAp+cMQb/ldMxcDO+v0yCen59hvxBgoXUUgCS2AxeNc+4DBYTCGCWEgQ+HCo0UwoVAGIOSWFVVuTLEsPLIYlGBx12AhIHL4k4qz2dljCVCmaDw0zNNQyllJ2MnM9cAvv/+e7RAbjebqHXITFPnHCaewqy9LbYF70Pui5rdDJIJa+1+v8dwchyeGVkCTojHwF4bY5QSQvAYk/eBc5IiNcYIIYG1zkgddgFv9/e6BzwQnBWIvagXzSw0bDZ8qfdOCOAwru97rRVj1GYJHZVboLHom81mRiGbpuGcYwVvb28P+yOc9tvqaNM0nMvz+Yx4fD5e85lzWcQGoWtT1WaaIOVfVtX5coYnWywWTVMXRdn346zXiLeIeU4RrAYsL2PEvpHPpbkbBbcLthh2h2QyHM9i6+lNPx3uIb4ZxxEBEeIdnpmaKNvC5oI8jnecsvgrSHh44JlpN2OLuJZlnjmM5UVzE9KvlBJnBKHEdruVUq6Wmxhj07TWmikPMIGxU0oBbUNRZLfbYfFTSlpraObg0EPkru97IQQhV5o8HANeRwiBgHEGFoDMcsZDpLCnVxPDmBDC+zBNaACuTqcnKbXWJSX07u5uPpaITuAJguPwW/jhnPPVagWZs9lVD0N/OByNcev1sh+6y+WCs4f6JG7dDDdjv2AoMc56uVzij7jkEFn64YcfIEK+2WwQqeBuo9qRrh2dV4INaHs49rAIUkoplZsmEqNznlJGWSKMhhAoIc4ZWDQhJOc0xmsvbnoj4wp8GZaOZV4NCBecc5a7lJfLZdf3v/zyy3q7wcVBJw18LVJn730Mcc6VYYDw+7qupmn0wYyXi5lGH1xMKvk0jmNZFMGLrutiCN55znmKMb2ZULbf70E+xKWQUqLvEo6ZZKZ1Sul4PIYQPn/+jMzvfD5//fXXRVF4F5HoI4iOkQ9DTyj0oBwhlDFkEXQ+G3PNcvasMDJlCeH3cR4zh51C4t513ePjI470fMfruo4phTz0ZQ4KUReMPsBYQTlqu90uFgsphCzLRCmGdIFpMkfuMUYf0xT/OqQBbglekJBrYASbDrVagDRIuLEvKQ+/wa/wLlprjDR2zhVa0CuR8tq5CST8baPM3+secLERqwLKSG8UYnEBwAHACcNCWCsh6fr4+DBOA6O0KmvvgrPeTFYptVysgo+ccyW1c26z3oYQfAqLdkkSRdAKUbOUEtq4pJQYvDyjK7gYWFZ0zWAv4X5oIuMwAC+GxaGMNk0j2xb4HYQxZjI1QnXn3M3NDe5t27Zd11GacA+xHzhGc6guMnl/RgxIRrfcdaxgmjcGcd9ut3t5eZlzhTk+AsMEWCTCDUIIPgusuLeJGk4PXgS2D97rimPkHvXgPer4PHu2aZrQMlkURV23uQBTVFVtzLWvBSRjPB58wMzFgs4gYpkZqtZa39zcIK8SQsaQpJTr9Xqece+9R7keOOTcI2KMkSTFEGgi0QcSUwqxUJpTFmNEBwB2EzDjbrfb3t4BrJsLtkIIrrUSLdqF5ouNaBpjCnEz+6zBx/lVbwCWHQYdHuLz589oksDhxxfyVDhaRHMPDw8xJa31brcD1fvx8XHmpMMlYD4VoImUByvhX0OWvokxRkp1Vdmx//a7HxKhISXKBBOKhCilRLoMZI8x5qwVmWk6t8LBumVbcJ0+ixUgmbmPhfr8+fP3v/kBRxEroLKaL06Rc1ciBn4g3Axj7HQ6wMTERAglZjItrSkl8OhKqRhCURQ0kbIsfQxjVgcAi71pmo8fP+JdkNLBpLJMGccZ++qrr87n84cPHxBCXS6X3/3ud845WuDZXNsunLOoBlNChJTGTJQSSvlkrOdEKeXjFQmknMWUVObCbLfby+Vi7VV9pygalSV/0P1QVdX79++h+DvzWa4ZA+eY7jf3EFzrFoTy3PuF8O6rr75CwieEuHRdCMF5B1UPinnh6FqwLrHonKM5gUOQh4o08qSZUIPMBvEHPj1kBjBIUIg54CyFkIRQbCXnnJArARJWyGVVKJ91qf8u94DUHtEiUKYZ8AHCDvIMvELbtlVVo/MAdmexWByOB6WKENL53C2X67puvfdal9NkN5sNxLmEUIT4shRKFULIqqpeX19nVvt2uwUWPE3XnppZ4Ai2BhcbthvrRQiZxpEQogudUlJaLcVSSEEoTd7D2MVIcF5jJvvPWCH8eYZNrjzRpmlgu5G8Y29wlxBEhxCwGQhzMuXjaukQbsOBgVGKvccqkVwwCLmFGKkDydEiUAt0YiNGJoQgtEf3EIwFrDZCFa01Mgac5hlSwxFHOmitHegAH4DoGImnlPLm5gZE/mmasAXPz89wToCDsc4A3FBUVEoppb27kn/mVAAPid9jMdPcjcgozvoMoOOiSi6ccylSMzmSWF21hJBhmMynTyrL1Mx0DiGElGIOHmHKsXR1ob33nIfZYnrvYPWAIOHYzPcQxhTnKuYvbDquMZhaj4+Pp/MZYqUIx15fX8uynJmR6DbHDkLIb05A8fopa2pSIYUQSenVenM6n5q2ZfyFc2FsT2g0dixKFb33waaUQvJoWIuZ3k1ycxLOGxwGInTGWN931tpTd8E42O12i34R1EsQx/A8HGIuk+J0YeOQ/1FKP6ty93oqZCVFQQhLkTF+1TVp6sZ5XxAihOCMkUDbTDxD3JbesJIQK2CnsOk46j630UFxB1ERvnkcjJSSUhYjZg1dSUFclOlKKvXjaGOacBp57j4jhFB5nbUwq3T0fe+9rWuOkUFIzRH9wKQg38JVSrmEmzKfEFAeMniaYWHEiy63rF8bo4UwxoR0bbR0zqny2mIZyJUIMKOFs41FPsc5xwAxyAeg+Sllxur8giy32TrnpLz+JZpwtSo559Y65IUw1zTT5X8dcAkyA7BxuOQovaIMglfKKKfHGgkhyrIQQiASYYwtl6umblFIgT1F2Hs6ndB/CHMPW4ZqJ/I7mFSU+Bhjp9PP8FWArZGRYaWQJ2J9p2laLBZlUSgpCaMhRiXlGMLxdKqqSuTNJuSq7oIXwQGFafDeo19GCOGcQWUVjwHL4nLX3vPzM4qZIcus930PLYF51+cThn2SuWkOTgUHejaRyFIRj4hMHndZAwe2Hmcavg0fjasIfwZDllIiMXljx753QgjKXIh2nLrT2UUnhNhut+v1NoQghW7b1jmPkGc2OjD3q9UKgC8KMGhfstbWdb1eL5EZAMuCETHT1QOJLEqMxEtKiaQHv+L6jeOowScO0UULuKMsS0YozgkMH4ZjI0bGzAOs5xubHnEUkeMioJ6JlTMzHdx2Y6wQXGkFOHE+A9gFOGwsJs+NpngRzC9DTPDtt9++7nZFVSI6wc7SN6xTfCJYxWggAPIGtgJKJsYYxpk3kXMqhXLMXLruZX9USjPBGaNCKOiIzKo21lgqrgQEQEPz7aBZiy3kURmwPq+vr4OZYozDMNzd3Q3TOOSsevYEMNycc5jIGOOVp+89pZQzWWjy3bfff/r484fnL+3ifnNbNc1SKi6YuJzP6E7QSkUfKCFMCC448H0MDMYnns/nGCNUgwA3xUyrxcnH1f7973//n//5n3jf5+fn29tbzpBFoYorKI2MaWBV5Mp995RzZ1x37EczIZEl7ArTz3NqEXwURdH3jmRkGEcLCR+l9PPnz3Pp8e3lxVGHJ8DKeO9RkMABaJrmp59+IoQMw5BCEEIIKTnn02hmCCGxhKjXGFPqIoRAUsKpI3lWD07j/f09XPU4jiDd+jy/AdmP9/50OsHW4S3OZyulZUm0QQAAIABJREFU4FxM08i5yKBCslkhG3EVTrX4WyX5/pvmUoZHr1wj2HGEz7gziHxTSjBz4HJge5yzxhhGWVmUCBWhA1pV1cPDA8m8nZTlK9AVieK7zP3Ps59EYo5zDwwXZ/p0Og3TSPLcU6gwlWVZ6KuLQuVAa10oXWjN8wAsmiWrcEmwmoQQKKFiz3A+tNaYT4KaBMobJCvAbDYbWAScPPw+J7AFy9NfsTHwfIwxMF5yV5efYwHEm+AqwBMTQkBSAnKFO4xnk1Jut1tguLjeOE8hM3nmzB2+ZL1ew5oURbVeb+u6vr29zYGPnWvycxTPGLu5uYFdDiG8e/fu4eHh66+/Xi6XqGOjGRgTIq/ltWut9Jq6znAcgn3gOTYPvERSjIsH4aZZ3oMxHmPq+4FStlgsm6ZtmhaVLZF3EDklIUQICXG95WpVZAWqqq6FFEibIM6YSKSUlqUO0c+4Std16/V6vV6vViuEaXh4BI94vJiS0nq1XocYhZSE0qqu4T7Rx7Ber8/nMwzN+XzGO7qsr+dycxkom0hikFuEEAXnQz+klChhddX88d//qKUiMQnBKE1CMOdMjN57a8w4mYkLQVBNEWKcpnGaxnF03lNKbebRE0q5EIyx0Uwuhi9fviilfvvb32IXhBCYfA7sO2UShDEGgTPMR5FnfHLOGROLxVqIcrLBON8P/TQaEklKiXHedR122VjDOa+qEkcO1h8TJuBZkUWh6januSmXfHD4b29v3717d3t727btjz/+CDU2ce3rkUWhrmfGOWstY9w513X96/PL+XiKIVpjzqfTNE4kJhKvmB74eKhOw2iO4wiEHJnNdrvlnN/c3EzT9Pz8/PHjR1RJ8zFOM9Y022jBOeblxRjxQzjnP/74I2Ps+fX16eUFQ4RIplHgJiLJAHQcMak4JUQVUsrNZrNer2EDgfre398rpWbBNxCZ5l5d5F5Xll0iIUSbhYIQ1mBOJeApZGkzsve3uQf+z//8z/Mf/t//+f8JIbKIuYHNappmHAc8xHK5nE/Y3d3der0i5IrNTdO0Xm+kVPe3d1VRFlqXuiApmWmihKQYKb8eylkGDpXPGNPhcORcVFXNGGeME0LHcXp+fkawhgDn2pRLiXE2pAjnMU2TEpImYr2brMUMu7IoKSGb9YZT1jZtjOnaBkUpYzQET0gqywrPMPvFzDgSRVE655XSsFAoIcaYNput98EYK4TEc6ZEWJ6fhSQRQA0TtKjLoioSoboomJBlVY/GeutjDOhqmfdsjlPSVckqhhBWqxXkzFarFbxvTiT/OjuMUgpjh2sfU0qUcClCjFVTJ0qU1okQRpWUeru9VUoXuqQsccGcs1wwXaiyKoTgjFMphXOWMaqUDNGfzyddKCF5TCGRSBlFhge3dzgc4LCttcEHkgWscFj7vhdCppQul85nnbW6rs/nC9yY9x6gDfLFpmliTECiAJ1dWUCSxxScs0Ccq6rknAnB67oKMYzThB4IpXVRloSSQIgL3nrHpdCFJpT46LlkMUWpZVEVXHIueKJJaaULHUNilEmptC7W680wjNY6KWW7aMqqjCn6ENabNeM8EVJWpVJFVdWE0OVydTyeOBeci67rXl9f4AmQmlRVpbU+nE9Sq0SJ9U4VOqRovYspcUIoId56xuWnT19eX3a3q41gjFJuredcpkTHwQiurPUkUec9obQoSyFlIolxNlkjlSSUhhiU1pMxiRDK6O6w50Lsj4f3Hz6sNuvlaiWVfHh4QN6AQQ6InHjWB8UgP3gRcIg55zGFstQfPn4eRnc4XdpFKyWry0IpRaXgUjjvvHNaF1JKqRSyN1goAOU6iyHix86BZspFRJ8HCuH7EXeyRIauX65Xy9WirEvGaSTRB8coSSlKKRAqeR9Op3NVVcG7GLzgvCy0s8ZZo5SUSjBOGadwBs75EKIx1z52hC+IzxDb4RgDeIRB997rQkeWIokxRUZJCoGmSAlhjOKyEJoITc7Z4/FASBqmgQr25flLoolxVje11iqlyKkwk4k+EkJDjIwLxvlkrRLXYB+Oh+ZuBpGFNGbuhpSScR4p6bqeEBKu0E3w1oaYKGVaF9NkYoxlVQjJU4pcsBiDD857F2MgJEEU9ff/49u/wT38t6QDVVMgXypv/JySM8aQuCEnWK/XiJJInqyZayyJ8Sv6dr5cfB40Jgu93++R1s3dD4joy7JEbwgO0MwaknlmDo5RSmmypm6buqpIIgRTthn13qviWpPAouNEzmSGGXNImYGecpGN5s51lun/gJKXyyX7712LdV3P+gp4TjRCz6w+YJSw75FErbVqiqIo2naRErHGcSGFvAIXoIGp3JvNch8Guj3hGxD0hSzeN0de+BSl1KwIBFoOXgdsbvylEKJu/g9rb9IzyZVdCb75mZmbD98cX0SQFDOVXYmWhNpoVWig/7HQi/oJXasGJHQLUmlIMZnJIeKbfLbpjb04dh+dCW2Kki8IMujh7mZ23x3OPffcdhwHTkMSxpgyjMOJU1v+JV0IDhcoD5cPzSJcXakJvAvQ9pimiXGmpBJS5py77sxYPhz2bbtgLGutck5VNSMzfd8/Pj6iZESuwDng0cxYrmtQ+hjnQshZOQAOBae3MPZgsUDJcs6Ms9JSQhYiiBlZDICSLDAd8mq9xHukEiF6LhgXvG0XMJ4CtHLOgXVkkvNEqwaV6DAMnz9/BiYJIdgCB+HHY3waZiwy01p7H9q2fXl5Qb2YUhaM4WShJ5RIKgMcFc45qjo84kDz7YVzBeNEETOO4+9//3sg+3CpjNAGFG2RGIBY1DherDBhtEIA1TysDjE7piTQAE+zHaI4ANSGS8YfltZCEei9BNMvjcrQLDFgZMhZt20rpFyv1/vjQUnlSLML1w7sDg7dkzRFYYVE4ogDGQa9x9CACG4CIAT8gM1mg1ynIMPFoaWUckpo/OacWc7wSzSJPXz48GG32z0/P6/Wa+jRPj4+FrmLTKx0fCM0Qrz3h8Nh3S4ULZMAg794PNDBA82RGGNSzkzwq6tNd+5Ox6NPyWqDnjyaZHxerzDAJPAQHU1r44Wn858QHnAYLK3aKPhDYUYj4ys5CHxiXdfH4/Hl5aVpmpBizAl73kOKIacYcwihIVY7EmekG3iuimTWFdGBgT4lEtGEGaFGA2CSU9ZaRx+maUohQvY300J2SfNigH0TDUng20vgKQBuouVKhUkGL4nbUhwoaC34AfDpeDaCelZlSl4oaYwx2lhrQ0g4aYvKxjQrJGfa4Z5zRigC76WUkMXCTqfT7e0tbAiPH80SuOnSTcVNQ/KOJjYAUOdTTOncdzFHY4yQ0lgrh2EYhuVyWY5oIh0tnGGwd4oCOU4mDBr+5enp6f7+XqnAhQwhxJyssUKIqq4giQo2CDzdcrXUSl9fX51OZ0WSy+v1GvkEOsCcc8ZZ8KEA4lwKzhhqbdCsFc3w48m6WRFPzTAuF5HlEFPy8wTGFFNbN5U2OWctpPceCycYY24YUfxVVZVSLOO7ABMQTcuovLVWCBfCPDqAaQOl1H6/d25GsdHfBp52e3vb1s1ut7PWisx4ym3diMzO5/OQWentFW5JVVVCy5G0miONv8Czc85BcxBC7HY7TXsYSxcQ5AXUZF9++eU333xzdXX15Zdfchp5wzwTfCun/fVCiKquY0ohxsxYyrldLs/nc9f3gMgh6MtIqC7FODPlpGCM9X0vKX8qwawcClS6wFcRzgEwlIoBhx3JO9Ks/X6PZPR8Pld1jdlMKUSJKCV6IcNAzIN4ZXGLmXrUcK/4VRhhgZELIVC2ItG5uroehmG73R4OB6UgdbPMmWHOPGeWGf2DdgfgYuFtfvOb3/zt3/5ts1gsdOuc67oOWixwbgBsz+czBsUxddE0zX6/h9YctnI1TQMtmUjLU3Hr0PAPMYYUj5Oz1vL1+nw6wQi1MSz4khRm6sjCb5Roh0eArZr/0fAA3AOmVmZ8ShcOYRN+DSgHfgc21ZT8lwthSCkMP7GEGVh2SX4LoIZ7B2JryWXwFAHalFJ02S5Wq5WSMviglHLjxBlTRnIx81gYdd6Q+pUltLCYUuV4H/G0OOfff/89fjw8L3Iu2HfJ41D0IGoymmItnWd487u7O/wVF5gyOsbomStNoYQ9P2rmwgpa/gODRo2PB1zyO1RO+/0e5MvivBhj4Giez2eMIEG9A62a/X7/7t07tBbGcdwfdk3TZCJfKSWMtUVBoXAxcX8i6W5aa+u6xrWDB4KncznifzgcVqu1NpYxhlkBmGNKSWmRcmiXzTRNq3V7dbXp+z7EbGi+Gm735uYG3pDxlHIQggvJUg6CsZST4jql+bTj/Sg4SlGIhFeTipQkbUv8ntvbW2tsTFEIkWISQgguQgxaKS5ECEFKiHgzpSTnWQjWNNXMUeEciWSpHYsTxJ8AH99sNkpJ3ChkS6BgjOO4WCyen59/9atfIYoDt8E7x3EU7Cda6nK5HKYxp7l6w6PBJeMwwnLQy4FjcrScAE8QjxKdraqqMB/+8ePHp6enctZg+Zxec2eLCyGF0rrvOnxgykkIcToez+fz9fX1ZrN5enpCGE45O+cwUbxoFp5GGyADAxJHceWB1u+ArVv2jJ3P581mU86O9x76S8/PzzC/ZtniQlKe+1gis1KCVFWFJj/yoUzaaDgOgujmcCO4V4kECCTph5ZKOqXU1IuvvvxaK3u1uY4h7XdHloWQyiwsY4wLzhOILQwRosRLuL5379791V/91cvrKzwDeOeArYZh8D6Wzn+hLMcYr6+vkSJXVYVGEbgwhoZ88XQydUwlV6vVahrHiKkvzjnj3TDEFIt7xInAGUEwLqPy8IG/7PWz1nTJBdDZGIYBw2KHwwHWU1ocOJBQSC1kZ7RrUkqJ5cm7zFnMCQr1yuhCqkMDXNIqVPTEgFdK4o9yzu/u7tDRhdZpCYZCCHEhZ4hfC7GwRDMj5ApVIb8aWhoxAxEEiGmtm6YBTx9dNUFDsJc+CJmOlBI708svyTljeToONo70xSgDl1JO0zizNaYJZWa5vaB5gGVhrYVnYTRuguCBHBOdBkUL0VJKIMtCtbi0bYsLKOTRuq4RD2JO9aKxdZVyMpVFtxaJXsFk0UUchgGCZagAYK8wwXIb5xCVEpcCrY5+HKDiyaXwfmbuKxrmghcriGom8BegotZaKYleMU4IYxkzwiklVI2MVjLgDpRcUtFapBRiClEJKbmw2vTnbhyG5aJVQlpjJBfWGMG44CKFGH2YYUaWueD90FeVbduW858kzAA2Us44r6xBHJI/qZ6pL774ApdZ0IzFYtE2i+50HrreKN2fu+P+EJyXfFZkw6HFYyrHmJMyMaNJvUJwyLTxBiaK2h3t30g6QtfX17DM3/zmN7vdDnI4pVtejIrRXIXWmkuRGbOVDSnuj4fJu67vRzfv9kG1wWgpQpneglHhTJUKG2gG3DQuEMylywFPXC+GQ3FRqOY1LedYr9fY9JlSGocRtwh//RKrgJVCChrGaS34eH6aJuyDQfKHy0etqUlj35DkdQEwwMjARMswDC8vL0DqOEl5cpJHxXPJxK9RSr179+5Xv/rVZrP59a9/jWOIpwYjx9ch5cX9Xy6XSOaws7YcPUaj7/CNuBYYNv4iOK+Q9lGzHKeEFArydfgxuC9JNOIQAih8vyw8/Kx6AMEgxng6nXAm0WiaeV1CILXJF4smIqkpAC7H48tZZ5Zzys77pmkSeNZ5RvD5BYEB0agiGWFAsZzoupzz6+vr19dX5Aj4LvytSNxQa6vKWC64IsRT0/INRwukCu5c8ibkWXgndiqIi0m3TPry4N0WQCwTe72YCBw93CioLIwo3tM0kTDGWFXVcilen54wV1GSl9Ibj0SyRJUgL2YXIo28wi+MtIor0jQ1UmZPqmr4atjZarU6dTNfKKbkg2csK62bxULmeVi/6zrwGpHmwMRx7QXRwlUjuAKOu7q6stYmcEwrW8pvxnnd1EKKlAKlrkpKvljUUkopdIyx7FEB2pByTCkaoznPEObLOXrvpJLKzJ0wfFQi4pOglyPtipwzjQGKRA8p54yZXjdNjPOqqmIISqmIuTPJYkwpBSk5Fn6kHJpFLfg8mQwrgv1g8LOg2MUAkMdgY2uZy+Wc407+4Q9/ePfwME0TluTUda2tOZ/PQ9f/8MMP6DNN05Q505WNSeSft38USWqjVoNjhQ1DvKisTcS5QJ74/v373/3udwhvUIEsmV+p3uAZQ4yTc8jqjsdj3dRcCD+OIjOQ99Anwy5xW9kxeO+9FEKBPRyCIHV3gH7Ym4QRVIBFhcnGaJgA9gkDOB6P4CbUdX3Y7mCQSilj7TCOXHDvvKDwA1T2eDwyYscg1mIQJxM1HG8uZSWnMVhGuEK5fM55ikAvLCLi8/NLziyEgFHW25sbyQT+jqRyP5GAAm5pVVU+hMk7eSF1c5lUaZrZQiqM9tLb2xuqOvQkCrJ0yT8uNl8cfU6pqqppGPu+t3UNBAzIBy4cHwWwMZFQOcLJf0J44KQxkEiEBJ+LyU/kLIF2AQL0xI8TNFcCN8SlwDNZrpZzlSBnIXIsyRI0NAjpRGTKuLnlq4G34PJQMjPGJu+MMTlllGNunJSUSsiQIr8wAjQ5xCwulOfsMiWEJVwp3CJCEd6cifyHp4WNH3jM4zjaWcDZKVLRSaQKXrCyREsgYgwhRNRbyKe0MkpJQ3uWsAIozOvP5k1Sl7htQTPQW1O0MAe+CW5FEQUCgQS/DZAlJ8EypZSurK3rybuQkpRcKCmTFIkx2t8H+cycM9Q1StsTTSZscm5o3yf60mKeEFQuBqGUtjbG6GM4dWcmhFCYLgycCymF914ILqVcLmtYracxXcZYCN5agzE7rdX5fPY+pBQV/2mrCaryMi2Pi4V54Jni3ikp8fuVUphQGYZBUE8lAyJHcocgIThmrxADhOBCmGn0aGxq2rKSUgph3gUiaOYg0Y4X5DcfP35EBjd7BCEqa/f7vf3ySyXlOAxaKc65D2Ecx9fnFyzRQh+7H0dpZgZnQa5LSxk/oDxuXCAn2a6SkYBzuN1uSxtfEo8okvw4I9olPsrFMEzjMI1ZcJ9iGgZ873DuUB9Yax8fHyFpfnV97boTuIJ1awW3XQjTNGHqDVVdwZktaX6Uw4L/hOdCxgORKwRdSexE29Rt20JAW2ktpIjOO+cgwq8utuggYUIpg4TveDwampGUF4spI0luZBpChvPNiRlT5Twrbwohlsulc97W9nA+OuekkFZpzpgi2oslPYVIJHJrrTamG/rdbleqtJRS0zQxzlMI0DMtzkEIcXt7W5JRnGskmuZCxBNe0XvPpVBKeSHOXQfUd6JFvLA3XBHMBncVIARu0S8LDP9OeMgspxRjikrJGMVcY+ef+D8TVvoJAZfHOQc/FVeFVDenpIRUQtZVXdf15CZEl8r8NCIAyBKjDxiFQwMQPnHWQPVu0SxYzot24Z3nnAvOK2NzypJzl1JGve+9NSZllknwtbQKpZTIC1LKjEGxJOecnfMxRiRWMBdU6CGE4IMQIvLIObdV5Z3ztI4Dr2KgjApPKaWUgnNWDL1pmpDSfrdDOsk5C8HXdWUro0iXP5PmXdM0jIuUcso8Mx5i4kJyIc/dvD734fGRcT5OU9d3iWX0/PFLxAVHsODagNTwC38CIuo6RM9y9j5woDcpxZxDSpNzE5RCCIWA3WO7liRFKSAM5VnLWWqND0M/TaMUIqWIfTNSiHEYZd1AZxvlCORoIayE6BVJmjgEbyvJWE4J3R0GhcYYnDZVilFwbrQOXLjJccYA0WCvn5JScC65SDGGEDmJlEDVB/dH0RqiRHo7UsoQY2JRGx3nlaus77uqqurKcD5XbyjUqEpD7RgZyyEE52bV5ZSY1caHkDPbbK5sNU+6tO3y9vb2bbs9nU5105zP5xNGL43uzt3z83OMEYRAH3xMkfNZIQa7/HCLiqkgdUCEYLTdNpLGCVwSIhyuznv/zTff1HW92+0wV1zXNZj7UggtlRKSpezGqes7pLQsZ+8cY7xaLjfX10pKH4M7ey5Eyvn17e369lZJOY1jljLGlGKEh9LGZMb0NHHO66YZ+j6GYNoWPBfkLhNJjZYiGxECSdLXX3/98PDww3ffZzYPWgtou7FGCZnFrDredR1ukZTy/fsPP/74I8ZlUs4Y87S2SikyPOucU4zWWuhYOGoiMsa0Npxz9Js5F4xlzmXOvO/H9foKUhFG66Hvu9Np4Lypm6RUigk5bslcEzEvjLWMsTJUjyOTZtA/j+NojF4u2xiTtWa5XEaaBEL+Af7SYrEAFw7PVJCWH+c8eKekDMF3fZdzElworSINNiJHKQlxuNgoh0v2/4Fd0z+be/h//r9/ZILHGLTRnLEYQ/Ke5SyUtNbe3NyUlK1gpribyFXxIev1+mqz4YxXoD8r3dS1oY4CFH4AkCHLAB5XYgP+V8qJccYE55yfTicPkOfcCcFDCIILN827UIQQjMZlS2xAdhlCSCA4xaiUBu6ZUmaMl1wbpVKBFyXnUggmODzI5Jy2Rmkd/UzGLREi51xVFjs5ck5KqfP5ZK0BI5NxcTqdL4H1tl0oEjcF1wKwbFXXXKjEGBdinCYhhAtBCOlDDCFxKaytFu3idfsWosssc5EjLS+EGSFBQHu8qirwoEoTYhyHGP1mvdJKKilySiyzFBNTEn0CIcQK4FKMmG3WWhfeETwRpIcMzWEiNQMAFYLnjHGWtVLW6BSjNaY//7TON87K5wvOhVa2lJ4gevV9P7kJHSVQ6adpypl572LM1liWs+CCMx6877suhiA4t9YKJJhKCT5vMzifz4yEYQxJbyF2opgAWAxIZxzHECLnIsaslGkXy7fXLct8vd4U2wZGAdDcOae1EgLCChBP9TGGHDPPPDMWfDDWMim5ECmznNPV9fV2tzv3XUzJVpWQMnM2nLthGLrzGa6kqipjLTgx8OPhQnMbBwSXg3te7ByhC1r0RRw054ykZxxHrElHtwz/0tQNyzl476YpTM45xziPIRitETZSTOWcYhzPeR9i6Ic+5Xx9fW2NncYRt1FIGWIQSjrvhZTduZNSumnq+37ZLDbrzak7w2AusQcpZZEfBtXn7e3t5ubm9vY2xDiM48zzBozh3Pl8LtV5wWpyzsZWUmvG+dXNTYgx5ZxyFlKGFBifKd1aKaVUjolllliMcS4mOBM5M++h+5C01pyDqxWFYFprYw3nbL/bdafz+XQO3httgnNSSE/6BYh5M/AgZdd38AnOufv7ezDXkRfCnDC1s1y2q9WyMrOcRuFQ4c7AV4Cxhj/h0AhIMcUoBD8djyGEZtFIJeu6Aau7FLhICoEvFQcLE+Kc/8Vv/+wXhIc/xaRKwMEXc5G11tk7/G5GpB2Uacj9UYQCvwbcVEAAlEgwBdi9J+Gj0kgIpDoHqAGVhJAyRR9CEGqG8xwN/aNbjs8B/7KiPST4UjwkN+/S8oqWVWXieyEkMMaOx2OM8fb2VhLZDqBYoA2pQF0ELVwEqIL7jlFnOCD8sJRS13V1XUttYkpFIeqSagL/K4SAgtvhcGgWi7pZHE+n4/EAsmlK0XlnjNm+bZWWNzc3nPMYotIyxqisYWxexSWoXwL4tYBCoAKX4iaxXNAqBHgQ5vCwUPZ57zOt24Wh468jzcnEvi0pOSt7qThDhwa/Z64/GPO0kA5uDkallWWE3eFhjeOYUjQGGkrcGJtSFkLW9QJGCCYxomm+GBAFVQY/GK4TMRKufLvdIp6BCwCWJO7G4XAAcG+MTTnBf5XqHrR6SAUjJS+RBt+CbocgdTnBRUopc5ZSqm2z3KzBRfn06fXu7u7m5gbqGjg++/0eLCZH84N3d3c5Z0bU9ZIzFTIo7gBOOC6nRAs8C0uUbvgsXAgKYk7bR/AX8Y0sZ+99mBwTXFqD51JMVNGyaNjADPhYiw1u4MUCIAKulWmvMuNMCIHd2sfj8fHxEXev1LKSROBLvYvKsu/7T58+oamGI3YJ0+ecu77POdd1LXL23t/d3e33e0YZarxQCtBaey+stW6cEIcCrRmWUnkfy/vrqnl4d49kUWkpJJ+m7KZ+TomstdZu1utxGF9eXoa+10qxlIdxFEouFovD4WCtLbSFYRjc5I7Hg1JaCukm17Ztdz4LkuiHzaNEqKoq+YAZiIIDI2VEh7nkoHKeRY9IpzStFAPJ2/uQcy76j+C5YBgQGRiwKbz5P6c1DbNA+V/bKoSQQ2SMbbQCmvb4+Ij34P3yQtqs9AkKfspIChtKLOlipr/0THDZ6LDhPgIkiSmNbpIqmFYLIZSZsQ5bWR8C5nQyiSkyKrHRQ7a0+pxzvlyuUZWXVj5MECZVVCTJhqSfZk8HL1NVFRbqaiELayvThMEwjFrP2iYISPi0qm6k1uj6go/osLVbzB1ptD1wf4zWIYSuO8cYvXfomXnvMBZnqxmo1Vr7MGmNFbvzJgl4AWBNJR6UGJ9oEUViP+2KKplpTHMdEB2UKtQ4DLhG5NcYE0F1Uiqz0hKE+Q7j6OPcFAWSdtntT9QDZDMZieGuSiKthRDmZiPPTfOu6/q7u3shpJRSKZ3zzABG+w6QI9CzxWIB7N45B0IEClP44tLUKZ4R/1eS7ul+vxdCBO8YpbGoilBfosKAwZfGlaARUdwTwKrGGJ+9EkpbAzm267vbRHPvGM2DPSOtWy6Xb88vsDSIPsECuRBSzXsRCiaAUwZMD70fTts6SxTntB+bEdcWRxLFBCZgwOF5fX3tbS+lfHt97brOSJX5fPjx26DEDBeTqPeOJ7jf7x8eHqZpgvJEyYcKEShj80pmy+Xyw4cPx92+JECXBbokXiLslhE7A5vYcWwdKSKHEEAfYpwDqKmqarvdhhibxQILiDIxaMv0bkoB2P0wDCugW3hbLK+QUrq+vvrLv/zt8/Pr0/MrJjqN1c7LMAUhRQjeRy+khBTSH779Vkt1c30626qlAAAgAElEQVSttE4pIfdNKWFVg/feed/U9R++/fb9+/fTNGXIPop5dgqmhbYrQOAx9XhkyKvwmUiGClwMk6bjNtNhQFLC9SpaqAz6ABKjgvZzomXiVhRg7T8UHiAFpaWSUrKU67rmKXvvE58PNqNkLZCYjKAZVE5dF0nDB4xknOGp8deRh8KjlSq+xDpCtAXjPOUkmZzLfO8RKper1R+//86QAGHp6GZarYOzUXyQcxPSItgc3I0ieihudEm3lVKeATpzwzBYzGex7Jxbbq52u11pggnqs61W7eUTxT9P57OxFvaNIVjEtrae20Tw6fDCPgRtkxA8Zx6Cd27q+67rOsbydrtdrtrSAnU+a61zSjmBLztFWgaZc0bNjuqhQBOIvjHP6rYFno4xjn52rCEEP7lxHDPVTHCXMFz8WvCUCsW2dIMBajualSs9wPJQDK1355xbaytbIwIBQwi0mHYYBmyy+OMfv4O3gtoK4HdcEeccbTf8O4B79EVge5w08jhNO5bi1dPWlxij1hoJeObcVuZwOKCAwKPBasmU0tvbm1Lqw4cPGGKCN4fLQxmBFDi6mcWIgVBNC0jQGIfBQwYczDTspUBBk3P+/vvvHx8fbVXJpLDv+nQ6YRqxEG9w/sFHUCRTJki7G4+jom2UKMGLHO9qtYJGzuvr648//oh1F4fD4Xq9cd63Um6uNpHoPeAoIpdC0YAmUwgBuL8gOfHSZVWkCeEzA0DTtq1kMx0Ri8VQ/EnqtydagIFYW3A/7z30JbuuKzpjpTIbxuH6+ma5WvVD3zSLSBkevCoAA9gAknpE1pxz5oimmnOss56E4FWtV+tqsXjsx/H86YwAzTjLbB65Lwz4jx8//tvvfvfy+qqkXOY8eRdjXC6XLy8vDw8PcAWMZqr/8O23f/EXfzGPQ2KhExF/L0lNuAkYFL++vi4ZaggB96og5FKiJzdL/wKKQDmLvjeMAe6lsOqBdsDICyD8y8LDz+YeSllQAhf+EMdVYMibQC5OQufwdIkmnBnp2SbiqHhSfCvgBhyKozVqpcKCY0opYRcKECdBQ0nOuUxTHoXKli9WtmUafEc5vNlspJxHH8Q8CTXzQNLFfDxjDLM2KD7wg+FK8MsdTYwjpXI0mVXKwEjrwGZQmCZ3DocD4HXArCXfL6gOnnTXn8/dyQdnrB6ngfGcc1qv19M0wn1vt9s4Cz3yfDGxAVyIEzMHv6cgXexCSACXXMy0wGtFHgOGDtUQWHahCJeUtlSHMAzcGdzYMtNfbgVcDG4sdSnmutNfKKvjK7wLfT8Ybb0LMUBHQGIXSPFEBeVDYQ7LwS9HOXXJZcw5AxhUxJC+vDmKhjpxftBgAK6CT8ZzfHp6AqRmaK8cwnDp/JcvFZzPuqGZAScp4AkqEuTdVVU9Pj7e3Nw0TbNerzebDaAVcMO22+0wDLBDP8uu/VRSF9sexxFpNdqSBUeCsyg2iSuVNGkEABafjI5FTBG7YC2tUwX+I2hTb6ZVYNfX1xBigRhfqZXxxPFMI50ORgN9xWag04PqrTDWIC+P5TnX19egzMFucepnbErr1XodYjTWZpYhNuW8KxkAWm64XY4UjuFMcTDheWJILM/PPeXo/SQ4M7VctI1QXCgRopNaGGuEFEIKpVRlrfe+rusvvvhimsZpmsZp5ohjDSp+JH7JOI7Be4xASSHcNHnnMkkE4g14TIFoe3i4hfVbSkC4e0i3Ouecm71rueQ8690mTFAhXCHfwt1DQoNREjwj90vXAf37Y3GRBKszUcHwh6VKSKRoWIpZOLtSWMATzTgGzdCXirhA1ZlYaJ62YuEeBWpsJmpG4Xhc5oAFqI0XkqWRWHQw8VKS42HAg8sLzYzT6XQ6nUZanIkPRwCQNJ1wOX/HiXLDGON8nh8uElIEdnF2MTuGrwDnDygqLh9nj89ROccYvHfWmqapM0ttu6jrBtUiwJDNZgNvehlK4c4yDVheum/6kTM8ikQbFwW7hBXinsPlzfYtJSI3IyQNrgcvpNuZXoEop/jA+YelhAwaaW/OeZomYIyMsaIdgtJhvV5bW3EmtDaMce9Dirmq6uVyBfpyucxCQCxPuZioIP1zYwwS87J7gBHtEp2qYhuw1eVyKUlODh8IJ2iM2e12OM+YaJ2BIEoI8JkhBimlsRY2ME5j27ZAwxhji8Xi+vr6w4cPRcUdsv4FhZNSns/nH3744fnpebfb/fDDDx29Sh9FkxI1wjxSjePxqGjAs+973GdHgy91Xf/zP//z9fU1sDtOW/+KSSN5MsQ2hHSYUqptWwiqM8LcYoz39/cQd+FEEkVN72nFuvc+5VTSvkJJwJFH1VWcDNhKzrk//vGP6PoYY1D7VlWFNTgoSvD+xHK9aA6nU2ZMKBlSGt2Uci7nOpMGz+FwQEGcUrK0hKock2EYQwhCcMay96OQjDO2uV5lsAFjYJwpq5VRtraMsaqu8Uw/fPiwWq0555xxdKSnaXp9fX1+fk4pnc9no/Q0jDwza0wKoTJGMLT6oyRpFjx6qHeUxOh8PpdnjeNZYh6ySXBYGA0bDLRYF94fluBpgA6t8pr2kYzjCDnqcRwhJfsLXn/ae8CJmqZJsHk9nJQyEtW3oBaMKgx0XxHD4UMDTS0Vn86oCVHsJtPqMWQoSLWAMOLQKq1jSlprP80d6VJLFl8fidiHnQ0gAgESKc6rVB7sYvcnwi/OkjEGVgUxWtRG8CwcSbQ1uCFIanCxII0kIlbj1E0kFe5DFEK0bYucy3u/3W5xeABZpJRAD2eMjdMUszdGxZiurtbWIvPikKFG3sfYbE/oCroYUEwIIfb7PRpl8AsIcpoGL3POi7YNKWLuvXhbGFzXdWg/1LZCv5QxhhWP6HHBx5V8B2NZCAaCiPkIQjiT8L/e+6vVGiGwZAaoJCrbqItZM0Y1WV030Lyrqnqa3Pncte0yBM/4zNktSBoj6oEklcYClzHGkKLCaDGVjewYHrbs8AE55HX7djgeHh4eNA3YMyoISm4OX1YqVEaFSCSOg9U2xuidr5ua5Tz0/ThNgHoYzdYJknnH6S10Yfy2ruuGw8GHsGgXmAEEvodWcCYlcxyrQMvMUXBgiz08Mi6Wc77b7ZbLJcpWAJiIJfjwqqqurq72b1v8FcQw7OPSWmMICQV3IQcDH0sp3d3dFagT7uynXoIPkceEJExIKaUWuhQxFS01QhEJMOSygkcNgSkQ6AYm2gDoQ1iuVqfTqRv6qqoyw20XJcwH0t3b7XbFMPBdQgg34U+YEML7SWsVUzqdTueua9pFVVtt9ak7QWrFGsOFyInlkHPOXZo1fh4e7qMPxpjVZo1cXmv99PSEbJVV9fl0UlI2VX06ntq2FVwEHyLL2c9pu5RyoG24WLUNgRBPapuIrMCCkNzA7FF7gWqRUtrtdhDqn6YZWXIk89z3/TRNq9XKkPZiIhXLojvwv/r6ObgUQ3ROcIw/xBBizCxmVtHy2KIBh4MdaPMzHC5uZYkNqJRLRMEiC2yZx/lcr9fYHwCFVPg4BCGrjTWmMZXgQgmJFABEzKKImWiMSNMoaSkvCtYJ+mVKs77NOEws85znxcUFuePEYh69O3VntHrAuO9O5+Nujzh8PB7f3t6KC/bBD34SWtXLhdSKCc4F11pPbuQi1021aOuqtikHxvNytdDWHo6Hru9X67WxlglhrOWCK6VTYlqbdrGytoJU+PnUKclra6xRq1VrtEzBj30XQ6zb2kcfUmCSNW3DJReK++BTCNF5kaE3zATjwflxGMau5yn7cRKMu2GMPkTna2NzjP3pnFKqmloZfeq7RdvmnDHFegnow9Ecj0dPW8bO53Pf95yxaRjdOI394MYJX5p8QDXGhQgxOuf6aQw5uRAG50LKMTMf0+SDUFrbqqqb+4dba3WIvqoMF1lpwUVul4viFmHfAKkNydzC9Etpyy9WGyHsIdFLPojMKmN4RvwPSPSgmYGUpeS8xa7gRvu+n/uNtNUK4LugIUqhhK3NqT+du1NV2eh99M4qNQzTODrvAkvZGqs45yler9e3t7dSGaZsFDYJvbq6Xq2Wy7biOZ0P+9enp9eX52E4CR6DH07nPWpHnDtkLcilSnqErAhYBAaJlFI3Nze/+tWv/umf/ul4PP74449VU++OB66kMnpzdaWNefzwXggR48wHQ6b/9vaGIjWEyLlcLlfT5Op64VxYr69iDJmFYexSDsPYKc1TDiE6pWQIgXHmvF+uVjElpmTIqRToBSFEtvvy8vL8/IyUDoe0bdvj8WiqmkmpbXVzf9+0y9v7B20rWzeVqUQWdzd30UeeeA5ZCeUmFwJ4ek4InnNUSlS12r1tkw9aazRzpZRcipgTY9FYbYzhXHCmjofx9//2aezjfnsc+zHHPPSjnxzLzE9OZJFjklggJiXn3FZVvWhMZdfr9RdffIFxoqenp++++2673X7/w/f745EJLrU6911kWWqdWE4x5ZhyiDzlsR/6c+cn58cJDE9H2gcF4s6JuckP/RhDrqvGmConfjrO1F4Ur6+vr99+++0wDM/Pz7/73e+22y1UcEo31FObtmTMBU74Ba+fVQ/W6CllzlhOyBpmmYTNZoOdPMhYy/chTy8ZdyREKNNksqeRelBZJL3iBVnIkjQegpCh9dmaMSllDMHljKwHPUNo3pam9ECLrwVpu3JqkjPGvJ+cQ5LSoHhwzseYvIuTGwtjMpCuS0opUycToNY4DDlnCLmAHYQ6jjFmrLVN5VMUmQsh5sUMKVtr0Q88nY5SSs4V52yapkWT6sUCOIO2BgkD4zylXNnGWiuECj5Mo6+rBWPcGs1yNkoPQxdVUJWNIYTorKiFEjHHEEPTNowxoaTgPIYZFuOMSSEwN4LCE+DM9fX1OI5p3i82KanGNKJi3Ww2+/1eGc0gxnkhGnMJ4ADE0CS5KITou47TLvu2bbm1k3NIaQHBx5wWlXXO+Rxx8xlnXPBxmtWllFJaKefdOA5eOqVkzunt7fXLL79E4lwaS6V7gf6hoFXksDQ8IJwT4CFAb1DlzDOfIXLOwzxXOF8sPny9XkPBJpGeFfYDQo8LZl9KUnxs3/evr69VVfV9l1JsmsZ7F7zPOQUflZYpMyFEU9csBcn40B1vbu7ettvB+ZAil/rh3ePt1frf/uWfX193XAohU9VU3vmJeRdiFpI3M5sAmWZKCR0LZEjI3JGHJdJxatv25ubmr//6r//mb/4G5/pwPDLGRjctqloJGWMUnGtj6kXztt2WfG6aJkqNDY7MYjGLbimllDZ9P2DMQggeYxJiXhTRdX0PIF5weMkYo9WmdDLwC2GHqKchVwxwz1orlZp8GMepaRZam3Gcqrrm1GlHNVxXtdYa2vXjMOactcGgO4/JZxZD8DIrDHJbY+AiGOfDOAbam8sYc86No/+7v/t7IczhdBy60U8BpVVwKWeeZcbWT05Lf6WUylohRN/3V1dXQLPRNIJxoqmulNIkMKy1lhdNXKM1Bg/3+z3wkmLYqCaVUs555zyEY7XWwQfUkUgLkM4CTco59/2AQoRzDqGq7XZbGnKwT2TtGDX7ZeHhT3sPM6IaQgEH0FeE34HzLfYkiKdUvAkq9IIDgNmC/l5ZY1CaE0opUGIYQeTiggR52Z1DFyvR7ofSDS7BCS0aQb30glYXPJRfsFlyzlIKFBmFOsVIO0TQJpNMcxLoRgI3L4UL7kbJWwO16dBZ0Vovl8v7+3tEGtBpkKkVMApOhzEmhays1UppqWpbLerm/bvHFGKxuXKLpJDTNA+X4NBeX1/PD0XKxFnMmUkxeRdyGqYpsXw6nSBeBtInbiznnGXGGeeZ1VWVQsTOMi1VJiGQTAo5uHYAa4yqRrRbS8LOiaZc+rGc82ItqEWEEDmnEOal6pomsIQUMUY4ceBd2GCaSGEJ1oisAmUcTgsY95c9mPKYFGmQDMMQWR6m0QWvrQk5cSlCTj4EKQVmEYB3VbQwTmuNA1zRS5H0NLuQra5omSA4l0BLyhHAk0XLB6NquKiP7+5+89WHrz7cLaysFPduur17+C//+1++/+LLerGUyhrTdP0UEuudP5463PZMW5rhxNHpdbQprJxcgHtKqb7v//zP/xxs4JTSy8uLpMkA6HqllJCXgI9U3EfXdQ8PD5H4XaDqNU2z2+2mcQrEBJMSbQ8upcy0HQG+iRPNPV3sHsALvS5kit77h4eHm5ubu7s7PLjz+QxqACafYF2SFKMDbSTENKgmIdtSyXVdB8Ih6hLMEMCMMccOd+FIEP5wOPzf/+N//PDjj3ApWLReOs+M8NKCVeIUF2XJ5XKJ1g6jTZqKlEDZBY5taVFS6WUyxo7HY7khBeUG7HY+nxnLBUiPMY7TgGwA6AUsHLwyGL+UEn8OIOenuvZChRoLYH7B62fVAy7P0Tr70tbzRFU6HA6lRV4wnEDawkirAaXhz6dpgv2VwXpcQ/kWnLF4MSKH7y1DBsUyAHEAf4dNFNYK+s+lBEmkkh1CYCx1Xa+UTrRMQmvNGC/IplJqs9m0bYvJKRzpAhbDVaWUYPf4/Xhs0zTd3N0Ofuy6TpJ8oxCS0dnINLCG1UmYmNMkSW0oy2CMCcarqvLOBx8MLRhwxNYHVs452+12Sqnz0Blbow8paeaOUdcks8wZH8bRWOuDh44xoOdpmp6entALhe4sUDLOeKE/BczQMsYI02QkZi5oGg7zOzDlsjkK0p6BVi3OiRtjUsrgHYDv7XYbU9BaT5MzVhurvWchZCkFFzNJCS/O+TAM33zzDb6oBFR+IeYIAyjsHXx7Ke84rdyCS9Jap2GAaCPCrfd+mhzqD9AHAAeX2hx31ZI2apn0xMGD30cyIWnrJK4ah3bZbpyfcNs367WUPPhwOh4fbzc6h7Y2P5jX0zCGaRKCP7x73y5X//qv/+Jjenx8/OMP38fEOef7w4mnnyjLWFx6c3NjaRU5bj601gExoev7+fPn3/72tw8PDxgq3O/3IN0jumDmPKYktWKcMcYQ5JqmeXt7u7+/L0kMDgIUwuu6YplrZZVUla0FV0Jwa+pluwS+X4hk89S6mhsDieQfdrsd3gMXgX6yMQbuuO96tI5Knw9+A6urMkYraGmSkkoqxeL86LVSIQYhxLE7o0CBaDYabGiZlCOPnnCixVZgORpjsH0Ba2YKYslIT0jNyozj4XCAN99sNhArhBbZ1dVVuW+K9saXvgL+Liob2I+1FhRYsBUeHh7GcRzHPrNU1zbEsLlajeOotHh7e0VmvN1ut9stwpL3s6hJWTJREBoEKoRSPy/B/E8itgbamwb/WPrJjjYEZCKGJuKt4t9L8l5OckoJ7UEcobZtSyqEiFLmrWBVpftXABzY1o8//ghfJmhiGYbiSVwXIae05vC/kGxO0yQER9MGcQg/chx/Gh7BFuiSAiA9kTSIy6hWgGMqNU1VVctli+49LkRKuVouUXPA4gUxiEDPQKaQcwYxEe+s61oKsWxbwXlOyU3Tfrcb+l7QdCEegabdW85NnhZsoU5HZNLWZMakUkjmtTH9MKScccBokiCh6xNjrKytrG3qujufp3F007RoGtw0RQt2cMCKth0SHPjNS8eK7CaTyjFSJ6T5nFhqRUtSkyRtSXvVxe4QxhjuEt5W3g90gpEmZeGwYmgfiSo6DcWS8TmwTGy8yYwJKbUxjHMfQkwz/62gozBCMECAU5VqAGEVocsYUyQ6EJ+AJRZ8EocZ9Q3oj0iQD/v99br96sO793dX/9vXH/78wyPzw3H/lqK7vln/2ddfPT7cv3//brVajcMYQ3x+ftlut2W3IO5qYSLhPiDJwLF9enr6wx/+UI7kb3/726Zp7u7uIKFaMqrSCYgkLLher51zq9UK4NK7d+9Op1Nd12Uo5KuvvlouV09PzzhBeEGmAlULShCcIBQH+KhS3qHRCv+7Xq+BFcM5oghDN8taezgckBzs93sId5/PZwQqJCJCCCE4OG9zx4jPPKuqqm5ubnD/NSl4o2TBG/DP0s9PMSF/laQ+EomeG4nBiNjJGINidqadd0KI+/t70L1wHwqVoxRkgtQJ4RLh4hXNXaG8rmiJhZRSG306HbmYpdELcgPgSJKA1XK5fHi4x0ItBH5Ni7uBzMPsJxoHFr+09/Czv4bjXfL34kz5BYcMj1yQkC9a0MAECpjAScuQF3ZaSuhACJKmYIzB7j3J6CMkIK7A8eE3SKJmJ+J4lUY0Km6cZ2RwcGeZ9qhwzo2Zld/DT8Nr8ebmBqkKwkMxkfJdl/1PmB36UeVuDMPQ90NBw6w12LoDIBKln70QkoLHKf8syEyMsaxVgDUoGjYE+kT4FRb5hhDmrAGZMiJHCGHoh5H2P5cX6v1E8qs4aWWyCQ8UvNvj8Wit9RS0/IXu0EgLmhjV0aXISyT8wC6U8SEBpLWWpAAPTKOqqqapF4u6aSrOc1WZpqmapqoqg/vPOYdaDowkhACgmdFUnSAmdKbRCsyXsQuR7URyh+XCpVKIbYg6OPalwsi0VAea2wDuEHUCyb0gHsNJFc4l/pMxBubYzc3NOI7ozZR0RxT6MufDMGSWd/tj3ZhfffXhy/vbTWsWhteKbRYVS2G9bN6/v7NG/frLL282682yNVz2fY+kG5dTBjPRMgFmiAopxrhcLosf+Yd/+Iec84cPH0AKgiEtFgsQInPORuu+6yO9UIU0TfP999+DSoS4iMpyuVy2bTuObpr8arUxphJCpZS9j855YDWMFAEKj6PwL/f7PWZ3rLWbzQZq4dPPF4KmnDAGcX9/j+1m3377LQCD5+fn/X4P6805Q2gZDxowYLto67q2tsIPwM3XGr1o/vz8jILGOQfGFybvlFJQ2SulM/5d8Bk5L2zs4vEKsAHmQkXr/ACLRZLQgGmV3C7RrhRNHH2YBN4/0RqYYeicm6Zp/Pz5R++n5+fPT0+fPn/+8e3tDR+LiAguz+3t3cePHzebDX4D/Ke+2GOhaRMzwLpfFh7+HUFvQVNRCAnOueMx4AaBsZeIMsRJCFPNesgzSSnTdExhuBc4DB4NbRZHkmeM1urCIWYagADAjZiBaXLkKSg88WkwEWQH5akImmQ2xoQQgfEBaWGMIYTB+yNy4DDAXhUNXZdKLZDqp7UW5xOLMbx30s40kkyTIlJKIaWhxU/gk0AEou+6giOj7IAjYTnzlGtjh3PHOY/OT/1wfX2dUgIjQIjZIYbAQeStSGZqpE0jYIVzKZjgx9NJGY2NTHVd39/fw98Bi4dyTnR+6HvJeFM3m+XqdDqtlkvO+eCmUpChgYa07pKVgNQP5wTHBvet5CwoQeqmkbQdGsH1uNsvFk1BSEvOAYgMnlTQQg4MD5cMt0D/jDEUZzjwVVVhtQAIV5E2kcGHCiG0NZN3XIpx6NdXGyR3McYwzWsQBSnmg/z64cOHGY8igUjgTngEuMyqqoqYElJ4Rnvcip2Dmgz8JCXhjKmr+h//5z/+n//Hf+MprdtqWjf/5av3UejluhlSyKtGCPX6tru+Wn/18QPjvDud+7FDkCu9DTxK/FTYFWSXhmFA7vz6+gplbxhwXddZzH2LlBKogzlnCBPwxFHkIZtGJwD+DllXuNh7jKl13DRjjNaGMYZS+HA4YKeho4nXRVXjB5dnulqtjsdj0zQ3NzdSSkxiIzgtl0tI8H7+/DmEAJxktVrB1F9eXpChw0FjMEXMUvzd6XSqKqOU4pxtNhsMFqAoAVXyeDzudrvNZgNfPE0TTrH3PgfvSaMCHSY8R0uNaPgoePOiKcJI/eH5+Xm73b57905rvd1u0f0GpRj9IfhlSXKK6CgAgUCwx6zVMAwvLy+rVWutPZ9PmfZeIJPztGEByWXZmodc53A4oLUQade0JNoxJw4RWPu/4PWz6iHSnDAyOE8bxD59+oQ4fDgckPLjABS0BKV9oLWj+BycE9wUTswTQQwQQTPYmXYoBlLJRuTgF23Pvu+cm5Dnwv4kLVHCMZ5oXhSFSCbmO+esqizSTE4DcUIIpXQJWpp0FzjnGFRhtFELFQYAmbZtK1pbhGdf103pxBD0NE/2AzxVFKjmMEDrWpGMSwGSruJsrnsE9TxQNsE4YNxlorugcMVJgVXlvPPePz09lX0pmRRp4C6vrq6K2lrTNIUECedraW0Ao1KPk7g6jFXSLsYS6cuJLVk2I7F72ABuy+l0ent7g5yqpR2/QKj+xDYCrRJB1oYsHnMYjMYdkEQzmrpCnaFpRJlz7rx3zsGZzit0lEa3FjcZP/jt7Q0dQjjHEMLT09Pr6yvgYMQV9Gbg+5BeeO+xvRUINXIIZJcYzywkRZRQjCBQxtjV1VVV1//2+z/ud4fD4cBiqIz6+P7u6y8e/uzjw8d3d+/vb2427bqp/dAta7Osq1//6s/u7u5wjYzK1re3t8PhgIwSGBd+gHPu+++/h6M/n88vLy9IHRJNKkAACk8Ef1KqTIR5JPiAFmH2INHjpZS+2lz33fD89JJi5kxwJk7Hc4xzyYtyAYyyvusd0Ulyzm9vb+h/oBYHqIsSPOcMerFUElwMzvl+v9fUA0PfFRcI4KHv+/1+fzyBTDjjwHBQly0iLLLEgDpAbMz9CHq//rlYpyPVH/bzJXGwTEZiU3jK+MzHx0chBLbLHY/Hl5cXQZo9CGwA3HDV+Byc4oqUE7332+0W69deX5/7/uz9NAydEOx43KcUcp6nSlEA4XLatoVUOGMMQi+UPgb0QjabDXAwdGJKK/d/9fUzQe9//t0fBBecc2srpdQ0uZxzXdfOTZfgO4Ktdy4637btMI6TcymnhOXuXCSaC8WNKNYGVhZ+KzJ6hHFGE3mMMeRxfd9rLWLyOaeco9LSVqbvz1LKGJkxljE+DKNSerFoqwrDUOF0OqYUOWekuowaE45PtW3bdT36NsYaqHAD/8m04/pwOBQwGpED0ajMnYFcCxhUaSW1HMfRGlNXVQhhGichhNHGOxdcsNqcjyejdHB+6HuxUP4AACAASURBVPoUw831lZ8mN43W6JySNdpoJZUA11MoMUzDMA0hhn7sc2aEQTEMQ1hbGWNSjotFMwzDYtGeTx1nQkkdQhRCfP70WQnZd900TuvlMkyuaVuUR1fX1xpEQyGXbRtyquqaCb5oWy5E1dSmspmxkFNKuW5qHyMTfLVev7y+Xl1fV3V96s790CfGjLXL9UpImVnuh54LkRnTRkupYkq7/Y5zEVLMjHHBhRBGa85Ypp02hQhUGvX74yFz5oKHdH5i+dPTZzAIEZC22y2WFZfTlUjOAck7Ov8dwSyZsaZddH13PndV3V5f355P3X6/f/fuEft/Yow8M5YyS7lt23EYWc6VrQTjwzjCMAJt4oStgsmqSXACPQBE+t1uBwrA3d0dkjthpLHGR8/nJW7c+TCM0+lwaBbtol2O0zSMXYqeZ7esleVM8eSHfrvdDUMvtPEp2apul20pE8FEAqg1jiNG3NH4UTR4iNTk6uoKVAsEYKkN59KamnHZjxPLrB8nbSzUA1C9oWA6n8/v378fxwHS9MZoJFjQMDfa7Pf78/mMWUIEwvP5fDweog85ZSXk7m1bmyrFWH6q1hr0MHC3FqumXbWM5xBTVddV3XCp6nqRchqG4d27d5czHNCsLo2HgXY0+eDP52MI88zTbrdvmsVhf3QhhJBiZlXdpMx8gH5frGprrBVSnruz855LwaWQSkGPAB5WXIg4CMYRPpHlIMwIKVfrlXNutV5bY+q6ZilVxn7/w3fH04FzdjweoKfpvXNuen19g1QlAsNisUDasd/vQalHCEcBgd3DSukQ0jhOSpkYk/fRmCqE2LbLGFNdN5vNVVXV8M+KBBQQQdu2bdsWUecSXUDE+q9/+ef/0fDwP//1W86At0QhBJZsNE2DFSjee7QBgf31w8ByFkL0Q++mKaYopGDQmRDzRCvyQXah4C1I/ICT5mXJ31HyT7QXSAhhjFVK9/0QQqxslTPjnE+TB9ABr13enHPe7w9wGpyLGJP3IcagtZFSKqWNMTEma21KOQQf4qwSA1dVpoVRAJYWFrgHYObAlyFMaq2lUjGnYRgqa+fwMI04bFLMY7GwP6B2q9UKjy0Qb7hkamCq4SmgcqqqSut5mShdppBSCSligtL1gBpCSoXcSmv99vrKUL6EkGPima2vrrQ169VaCOGmyVobseOMzRpNKALm/DelaXI+eEui80AIrbXgzwAoQEtWkTIEF4JxVtU18AqI76MrXkqrSLrZpSrF9V4ieGj5RlLdYZzXVYU+/9vbG9ZacFpLWR4EmsZd1wkppVZd33PBnXdCyhDCol0YbVerVdd12902xljXFUxu7HuGfYg5T+N4Op2M1raqcGLLYCAq47mTSb1Wzvl2u12tVsgHC9rZNM1mszkej0JKCL5abayZh7fPp5OS4tPnT5zzYRxyjjG49XLRLmql7TROMefMpQupG6flauOcF1JgcRYE5PF1pQGOvBsTHkD5t9vt119/vd/vC4zJOTdVFWNsmnoYeikFE7wfB6mkFFyTUGPhEVwqtwPBR4l8PB6klNaa7fbt/v7OVjbnBIWdGKMUAumgUkoLKbno+v7x8dEY8/z8DDmZtm2VVv3Q1U2jlBqG0VprbOW9//jxCyE4Kgzn3HfffQe1VMh3YtsagGh4W6xSQITWpHY3TVMIqes7rbVU0lY2hOC8n9zULlulFRczNxJLTEvv93KHDyxKK1Wee2lzjuNoq0pKVSDot5dXa+36an0+n7///vuyEYgILAzeEqknqLElWqAURoMaf6KUAtrSNAs8BfzncrkEXQ02CT9saVErQGBk1aDUF1AkXywT/GXh4WfgEic9JU69aEds5UIDx0mOMbI8L0WpqsqQ8vtI288D6WhGWt2HogH+AvfLXiy64iStjE/ouo5T/zkQlR5VRWm34l4wgr9H2u2FSNZ1Hdb7xYt5CCJKzfQb+KwCZwVSjgN8Ab7TZQYB2F2TGgT6JfOwd8CgwDw/wWnMojw5fMt8fqhw5tROx/WCBoP7zEnEgtMCdyCqx+OhEDBCCN4HZB8zdCY4Y0xgVwENELALEWYkX0DbcGMlrUik+M3xg0sLCosPUdcXARl8CLiJ4B2WjhF6g3i/Iol/vB+J4X6/x2EDloUHiqY0J0YNkl9HGgmXvb5iEn+CYmla3QNPV0Dtvh+Ap8Fxl4gliVPPOV+tVphumabJ+flLUZ5PJF2FkFZqzZubG6DYsMaC84DFgNyF09wGOKl105hFsz93f/8v//R//ff//v0Pn7bbQ/RsHOI4hZBZZiLkxKU8Ho9CyZBm6Y62bd+/f397ewsfUYihuOHn87ngQt99910I4fX1tfyqGOPUD1M/ROeNVNF5LWSlDU+zyHFd10hcUJ1AmgLHMBAHYblcbjZXQgiQoHa7XaQZWJyL1WoFxBKzTVCpatv2iy++wFYVxKqmaVKKkHwolJtEK+TwfGtamitJROf5+RnQriKpV1RIFekMKhJtLVguLj/RNoHSThNC4B56oqcjnUchCyDIOdcPwzCOWDQUYgTVbZzGvu/rusK51loH0pctoxUlWuNAIXHBNZ5OJ0Hy+EV6q+CrbdteXV09Pj7e3d3d3t7e3d19/fXXHz9+xPxg27bocKC/DU+CvwtPCFpBIWvhyM8JCvnJX/D6WWv6cDhs1lfOuZQY7ibceiLeKg4ezEIqaVTDhdBaTW6G/nPO1lgkyDhXpbsApwlvCKbKarUqVwIiV4klnPMYM+fJudC2K+/9NHmljJS66wYERgRq5CwhhBASYyJnPo6OcxljHkcnpWDz8j8phOAK/YbsvTdi3iTMSBgcFoaMCbwFXBF+LaOklZPaBGcMNa+kBetdd9ZKl3sFP6tJl6a0SdjFHgJOFH60s1AowI53uwNqKTydeRrTyna14HMvEfmaxqfFGHOeR0xzSvDj3CttDXyod44T8Rf+C+EHzeTz+ey850I0C0yYz12ihgivePS4+WB/I8dpmgbDKAjVKLkg4lRiLdwcrhdscdwE3CXcPUD8uGQ4CJzYRBOIjPTkMZ0E3K/QCqZpOp/PkPe5u7vDJVdV1XezbiV+JyfiGSd6qCRZU+ccD6E7d+/evQOeXpF6Ek4jWoIlTUOKg7bK4XC4vb3Fc1wsFj98/nR7eyuEYLTipkupXS6nyNrrK++GxXrz+fnt8Wbz6fnt7vbWc77bH173px+fX/fn4fPLdnl1ky+WMimlABnB2SEVu7q6wrcDhIRKEmS0Md+LdBsq5bvdfNuPx71SqmmqlDI6bWWIDHM2JfdEBuBnhd2olPDeKyWk5EIwa7Uxyrk57Wua5ng4jOP4eP/w9PTEOV8ulwjJxapjjMvlKuV8Op2ur27RSIBrlkouFguIXH38+PH5+fnDhw8IwAjkKLhxKoVgNzc3nmZNSr58PndoX5V5ZhSXz8/siy++MKQZiooZDSr09suo6cw16AeYIpwSzixQ9LqqvPNSSimEMWYkdUjG2Nvb29dffx1pFu987vGDL10EqgGgTImI5ui3rddrLFWD08/UoMVx+/jxIz6NkQ4QDiPYX4yx5XK5Wq1QQUri5c9Y3C/dJ/qz6qHve9BaEJZBmyuMIOccbquk8QJGxYGkCRr8X+Ch4OmXfg46JCC9IRcmbEQymi/lnEPmEx0/QD0lV7LWwhcI4iMmWveKZAdBFbNm4HulNIOJhZeJf2pjLhNeHCqkEiX3RwcikkJWCHOejk9IKXEaD3bOleZEgS8RbAytS9JaY3yvsG6APpUEaqbZ0Pgl3okDAIsHuIGUoaqqnBPoKJK2rscYlZJaa045HZ6p976kDyVn59SoR7GFGjGnxNi8fqOAP5w2xcKaL6ufTLKpJeMbSXzYGIMrKm9LREAo4VbQqB3AX1TN5bZYWsWBhBRsMUDSCDOMOvDARgBf4OkjBdlsNuM4Ns3CXQgaoyjBfQA/GEnZPIUjREwRUQTzIiUtABcDfCp8CJrY8FClUPjhhx9w89/e3sBjQbNkmqbufM5CNIuWc7O+usvayGp5GsKnl/3/+4//8nd//0//9vs/Pj2/dv1orBVcMcastWhLIsktaA+UMzabDf55Pp+HYXh9fb2/vz8cDpn23DHGttvt5x8/uXHqTueh60+H4/l4moZR8lk4APk4hN40baBEEAJHINIcD6MhWbRSPW39YowdTyfwqbz3cFLe+2+++WaxWGAvAh5QCB4UWHBeQXbC2CaGzxljwCewbAfAnaJBHFyvtXa5XFlrsZMOyHuhWhScA8AAzGO/33/+/Pnz58+lT3sJkPZ9X34zJC58DCHFxLLUikuBSk4qGWM8HI5d172+vHz69CmQmNDV1RUjjAT0y6Zpbm9v0YhaLBaYhuv7/rvvvnt5eXHOgYGGqZT7+/vr62tOI3WC6GHFJ6C0Xa1WkHgB5FsusFws+tL4K2DxIQH6z5HkU7SbF+kYI55vIlI5tjDOjVxaBoIzL4SwtpJSZpaLby3df7gwpHgltCAkSBqUY1RqANGCH5ZSN01bVY1ShjER4yy5w2huhTGmtWaMS6G1ssGnnLibQorMmhr5e6LJgzizHlWKKeeMVgoQJECBBRODC0afp1BQGCmZw4IR5wqUUVXV9dU1vCqsGbKLgtbmgMZQAgz0XMdxRC8HrESEB0vbYzjnNzc32KWFjgtiNi4ZgCzqdIQQaywOeUWKh4WXhd9QbjLuPO4PglzXdeeu8wS1OdrMUQwRVTkqtjIFAoPBTwXKhCOx2WwEEUYFSY3CgRbm6ziOiF74EMARSCwQ0mA2oF3BHUeaQUHZAdgNN00phXgAZBYUu67rhOBISuBEIJ+AcIinWVCgnPPpeOy6DmhDKVkAvyh6YdIC2gZ4TEXYFbfx5eVFaw3HetmA8d67wQmmbFVxZQeXkjKyab3UNw/vHz58sb65++KrX0OmcHfYw9uWpZiImsiZ8NyFEI+Pjw8PDx8+fMDgAvhp8CCJBMpi8m/bl/1huz9snR+HsVNapBwETUuVKU4QrgLJB4QQ3t7eYoxvb2+2strIceonNzCeYvKn88FWumnqvu8lJbaRpsngkZ1z7969w5nqexyXAQUivh0kwJTSdrtFLo8ptsfHR0BVhQmGFGS5XILIj6wfLJ2c8/Pz893d3d3dHZIPZP2IScvlEm1hTy+lVF3Xm82mlDXn8xm5L2IGXDPc2iWaL4Q4Hg+73e5I2SSS18I2Bgtc0CYuvDD8z4gaWiR7V6vV9fU1AthyuYwkRY42G1KfRKsTEMvxcDebDY4VUqimaZAYnc9nfF1pE+L5/mJw6U+npmETsJgCCsN6UNgiX3DOgVxJqTEQBW2t1YCeaAtNaa+Xew1IEd+Im8WoJ1GgZPvTUqSIjD4RCV3SSkJOzFTGWE4/BZjS/cZhFiQrjSQXUIm1Bu/XpK0U6FWcO44Kskg4NaTAyDiQKUwklIZ4A3MBfVPQmJsQYh41iDHO0vODuJCrijTGCbAC9oEMKJL2EaO54vP5DADdmNmfGmMgkoHeALIVuLMySQAuNk4ablqZ6AHSgiTaO4fTiwCTSEVKkVp4oRWUvKz4PngHpRRWu4BHwYiBmkmgpcBrKJg4iabglmYSzIAHQVKGTHaz2WB+AqkxjgryFQBBkpYCwULwq0DvQTwoTUigt4gfuApA25vN5ub2NqeMJL3A3IzqYxgVnC+oTciZUOvgo15eXoC8wxOx/5+192qSLEmyM41d5tyDZmZldVb3bKN7B4AsF8HjiuzvXxGM7GKAQQMzxZIF83B6mZF9+FytvEb2qbr9oSQq0sP93mtmSo4ePaoUT5LHWNry8euDNWVRVOVk9k/f//jjw9Pj/tD7WM/mRdW4qppO56vl1TiO+/0xXLSmgmjxv+f1Gkc6P/7u7/7uw4cPf/jDH25ubgAZbm5uFovFcrmMMaqkulN73B8evnzdbl6P+8Pnj5+6tuu7nl4zuNFRRkiR/URhPFOEe91sggy0eX19JRfsuq7remvt/rDXWuMX/+Vf/gWifd/3nz9/VkrRyVwUxWIxx4YW5xm6R2DYVsbGKZGaV0pR54CMi31hgzFJiSvESlRVxS1Yma+QTScTrvg61oWAGoPGeIlxHDebDXuVTWgKF7Xq/XhoTz7FqFUy50lcxtgkLTidgEV8L1EXPNfMbWGvUpzgLGSzyc83Nzc4VOdc1kHIEQ8nK4d0UQbTYhyIj4/HY5Li/PPzMz5JCY8cG/U3cA/Hw5EyEZW04/FIrIrTB+UIF+VccnzutiiKyWRaVxWKRvzG/VqQI1w0bmTMSkuxmg9sZKqUEr5ThrP0hWxfkjGKtM7FGKuqpj2k67px9IQMZVFm1jMuhyspyiLX9FkGKy/CvQxfZL5KXmzezO5UKRl6A8Gp9LmNPojifCOjhtmj8BHhAhlj2NPjON7f30fRBwS2ImfUWmfFpPGsuuVQUkMqKqlUloW1xjk7DEOIZ1i8LMu26whSWAXAmShjp6ixO/HlPOek1NAPhKX5l9mxKWnOzGCrlQY3PgqrmkRdgzyjE8loNnpd1bPplDcPw9DUTSujwbL/sNbOZjPK9T6EUYLZfhho/kopGWu7rguYXZny1Pf9+/fvCQIwK3Vd400x1oC/HFc2Vdd17J9aWjKbSUPgmetnhfAL9MXE3IwmdV1HG7BSio/NbQpkGGcH41zf95O6HgcfQzDOlnXjqmp7PD5vdy+v+8Oxi0ppa9dXV7gfAPTckMwdEeSWImyMGXLOoZV0f3/PzsfnzWazm5ubqixDGI/H/el0eHz82vftdrv5+PGntj1vie12+9NPP202G76Rci4o9mI+/+//7b8dDofdfnc6HqrSrdfL3Xaz3W7a0+Hl5UmlNA5jiufpT+QNNDnHGH/44Yf9fs+R32w2r5vt3d0dwBFAcVEUi8Xi+flpOp1eX19zWvNIJeIbtlwpfd0STabT6XQ4HLuu6/vh6ur6eDy3GqSUKJUTZWeosL9Qc2JjkxBXMkDXe8/YLqtNiqksyrqsnLHz6exqvV4sFk1dpxjbtmUw+3a7fXx8fHl5yagjStp4PlaffMIKH5LU56effoKKFqVjIyA/LhaVXccKZrgla7S0bZt1XIg4yXXyntQXBD8tJKC/1j3oaJJPJhmTVPLBpGhSLIwGvYKRwjrh7rpx0NaQAoXR65hM0sThoA05js7hGAWMjLZrrZUyz8+blLQxriiqu7s3zpVNc3bF7A9QERYAaMtJnzaL6gqTVK/0eDhukhqM9UWpprPSWGuMq6qGWD4mn1SwTltrsnNSSpUiKkcpEkOGgSPiSClcXa2MUcPQjWPv/WCtTsGHfkij992gY1rM5k3dTCaT6bRpmkrrxEx4a7VSUalYFDZGX5budDoYo+q6TCmUpbM63V6v5/PpMHTKpJD8qTuMcXTOrNdLa3UIo9bJWj2Z1LPZ1Frz8vJsrVEqGJNiGoxNk2ldlmWK5/wvpRTgVlmrY7LGLBcL731ZllGroJJSsSxdjB5p9hDGyaQex95YY62FbZIxXzI/JQhbuhDFy6m3kfZpKpxRFCBAdcFwCuv2221T1SapyhXO2LosS1fwZoIjPM05TXE2qlRUZe/HY9ceT6d26H2KUal60ihjJtOJKwtXFGVdRXXun88IHviDNknpOJtPisKmdNb1Sim5qtTOGme7YWj7zhZufzxady7kkDdjibgYApcoo77YM8R3TgaR4nicc7Nm0pSVCjHGeGxPtnCzxTxpXU2Lauq0SyH5elJra4uq3h9P9XzZhVRO5vVk3kxnPgZjTNOcPQTbHjbB1dWVk+5ZshZiETYwvyGgwQ7e3t5OpxN2Mm4Sk/H6+jqZTpM0kG42GwZkbrfb0+FIfWLz/BJ92G5eN0/Pu5fXw25/2O6borparj/99NFp+/mnj6fDoW/bcRgGwQxny0U9Patkgr8RHlEd6dphPlv6MWKXX19fqqoIcTzstqWzVqWx7+rC7V83yY9XqyXKK0rF0+lQlm4yqauqGIbRGmdNqZVVyY5D2L7ug0/j4Kuynk5mddXMZ4uhH1XSXdsB39PVRD5EjjJtJtGHuijDMHbH03wypVfpuN0pH9r9wSZllbZJ+bbfvWz6tqPKCiy83e92x8Pz8+br18eyrK0tvI8YnJeXV6RE2rb9+eefn56esJ/W2nfv3uW6OjIhtEGAEUVhbGe7xL/mLc1fGek4ZrcXwqbhhQXIwdxvzh5+1ffwH//hn/Azo2jzntuXRFBhtVoBfT4+PnK5mAMlUhxKYCLuE9+Yq/b4auqQAL5VVYUQk7DTcrZhRA2NzyEXKcsScgvl/vwea21ZwmUK+/2hKMrJZDKbzbXWk8kUh8wTd+4XphYOZhgG1BHwwyCPvI08lMZ3rVUhtG6s3rl8F8+EOYyCOUu3Ft6H7XYXQpzPF8bYvh9CiIvFsm27YRi1Nvf3bx4eHr0Pi8VyNpkVrjx13WQyDTGdTm1VN8ZYZyzYMeQEtoIPAZ6YtbYoyul02rbdfD5v6sl2uwshFK5omsZobbWZz+d103jvb25uyF6T8P9SOk/txhyjWfvp86fpbE54niPlICRXVqEWjRclEkwsZRAib0448kYHzrbW+mHgT0gmhr6fTqd+HI05o2f8UyZSZxYHzJxm0igBUp1zRVmA/GpzxpeIRjNoSxrU931VnXFO1ogDScXYWEvDOeY1psRBVcKSyiUrAkBio7wcVgQyc45FhMi+hb+oREnwdDotV8u+72NMzhV13Xgf6OxpmmldN4zM47LVRWe4kT7tEAKQSwaC+FjvPVRgkkLYMvwhlvH29s5a5304Hk/DMDbN5HQ6LZYrdPeI/9AmsdZ+/vjJGEOzCKvQdd0w9OMwpqTatuv7IYZYVXVKarv9pYeAMyLiox0ukyLcdrvlGDaT6e3d7ffff0/6+8MPP2w2m91213fd89PT6XR6enwE+v/hhx9ubm6SUgTRlKAfHh5YwaqqjDVKpdGPReGGoS+rcjqdtF2rtDJGr9erp+ensiyUVkolkksei1IKRnJVlCAQAAyLxaIW5wF6oLWez2bH43G73X7//fdd15HNt207mUzIOylskA9hqZGZCCHQyfH169eMoFJl5BDxhPu+n8/n4ziiOOJkOjp9QhlFILbLpV/sD3mJkfFr1MApBmemKObuf/uf//wb3MOviK3dOLRDb60NKo0UAlQK6SyHmVK6HMvMfsUH5kjcy7DDSoZqYZJIlIx0P5JdQmyNcdRa5/Eg7GboBGwCgkpOHcYomxutNfksvU4c6VL0XJVSxuiUVLZcVjRPcmHHOceN5GIG5lJUlc5jnL0w97OqIAeJNcteSgny4KRf34gwlBGq63Q6pY2eOhi1691uV5bl6+srVPezrRlHsul4MWOgG/omNoS0OS4ry7Kq6qqqDlpba5u6Hscxjl5LrylwBEUz6VSPGTRje81ms8VioZLKfjQKUUrJtNsMzhYy9NvKtE6yFsSNcz+REimYw+GQO2BZ0LqukQTHAc/ms2EYmEzAhQ2ilqFExpn7BU3OoJ8TOSxgViMchIznPj8/UznkX82FEL+SkeAw3yaTCU87Oyo8DU87509kvZcdMJgGL8JQnNviPCrnrNFCcSJPheT2M3mBVxQ6I3YWj6WEbdU0Ddy/DIwo0YPLQSinrxYdwKqq9vt9M2mMMW+/eVdPmpDiy8vLZDZVRh+Px5ubm69fv2IQD4fD8/Mzf3s8HhngiFrRp0+frq6u8ENK5o+ms/RyxOzyiM4UCTn+SinQVNZuuVwOfd8eT4V1P/zL9+3xdNjtSVlenp6stW/evIG1aK19eHj4x3/8x+V61TQN6Csn6PX19bA/vn3zVimVQiysM8bUZVVYt1gsmJW93+/fvXs3m836vp9Op2175u+yoDAvIEFke8IBub29xaRgyrIIOQvEE0jCeHTOsRwY9w8fPnz69OnPf/5z1ihLIvjIY0Gsk2fFQ5vP5/QMYtbpm2FrwUAjhL0U0VKCBmfJppzC8jPVRFAmTgHhzm94/evSNLeK8+Q48XvGGbJySqmc0l5yewaZEKCUykeR/yUwh0sjIfZ5WAf/itvkaRKPx4thA0q0/oPIdeBI+Ku+71NSISStbVU1dT1RymhtlTLISeZqJwYlyUxmQElKuKDGToZDZcI718nF4F2ckOeMzIQhfGZ5+r4fhlEpU1XNdDqfzRYxKmNcCMn7aG0xmy1WqyulTNNMtbbkpM6VzpXD4MchaGVVMnU1KYRwOV6MQTbGFEWllOn7sSgqOjyOx3a/2+uznJQDCalEXhjvTlEXDdQgo6GzC8TzrVar9XqVEz6cYiWS7CwlD4fF5XGxH8g2CPTwxGxirpkK1iCD7/OjDiFUdZ3Smc3lpfUss6I5AMVF70guDOSkECY0YqVapLryaawvlMydKBQpmRrNhrcXDXf5ABMlaGHd8K/cQi4qBJnMBdAEHoWxyANtyLAxECoZZ8vgk9FOK3s6dkPvtbLUyQk5aecE0GMt8KxJyNwsBOZ4MpnQ1KaUohoBHIqfw9aXRd1349X65rsPf/i7P/zx9ub+9ub+2/cfaFv73e9+V9c1brKX2R44uXghP1dVVUp6vz+yn9fr62HwxjilzoCVUoq9ul6vF4vF/f2t1mkyqfu+Hcd+v9++vDx13WkymXz58oXU/NOnTx8/fnx6esK5oqYH/4L2ETQ8cjX069evPMaUwpcvn3a7VxBardNsNun7NqXQNFVKIaXw9PQwm03GsbdWU5XMMdw56ZSGU+4dxgE8QMrsCGI+PDxwPQS4TsbGZA3ExWKRZ3TTunh/f8/2wzrd3NzQoDqOI99ClLnb7bIzZgqLtXa/3zPFWixbAkoioOn7nlpakPk3uBbeUBQFb+a78uH925Smy7o6tqeQoo+hrKtu6OtJYwuHEUwpPT09sXGVdOUpKSfiWnCtl/F47p7lE9iyvQiRY5vY+lZ6BbAd8/mcrArSgr1ojUF1juwpWyiSfZiRyvW0wwAAIABJREFU9oIcpWRUjhbFLsqJbIvMluHacubBVWEC+D3QMwwEjgGlV8xxkNn0SSg6WEnCTCwOXzGbzZCJZ5fkSDZJvTfjCXDdkLfEdhTSKJ9SapoGWzwMw263s2hoWxeFbqSEdWpEYh3zge3OtoCHo86y5IyBqLDCePFedC8g5mJVrRDYMKBGyLv8bWYrAvjwMPlAqnbI4WEUMM3wg3mSLNPr62tZlgweiDGyHwqZJU4mlC6K7ZhO4Ec2UkqJh5avPFOMSMkHGQjhnMOg8xvqJYgEJ2l6yAXbTPUx0iibSyZKuvdfXl4I33h6gABsG75xFL16drUXCZYkUtLYZdjr7BZqD9nGpYsuzhACF++EGEPuhRXux8E4q4x2ZVFU5f3bN23fvf3mXVVVjJchaIUzCheOicr39/dXV1c3NzfffPPN9fX1u3fvMjxLn7lz7ve//z0tXZnGhhu20qIMNOqk4XnSNOMwUDn146hS6trWaN113WQyYeAYuSMsvi9fvjw+Pn758oW52T/99NNPP/306dOn3XY7DsNuu/XjaLQ+HY9lUTw+PFIxJtLn0bEEuYWFg5zZFvSmwEKm9tuLVO1msyEDy+jIYrF4+/Yt7Rc3Nzfr9Xq9XtPw/OHDh7u7O5LCN2/ekBTy6P7Nv/k37969e/PmDWcE3kFKiXrPZrMhvoHwin2D746lYsR9FJGx/X6/2WzAnbgjNpu6mIIcY3x9fSXvYXv8NvfwK3CJlgoKg0TT3GSO9DOxJ0qHlDGGEnyO95MQMfM1kVeSAheijjvIlOBh8Hn3J5kJkx2Ml2lT+Xsz3GGk7XAcR62N0UYlVZXNGfgOaoxhsTyXQDC4o6iiZpfGLsF1kfXzyTgeIMXj8TidnjsGsiuyIlWdr4Q3K6WsLWNQWtn5bNmeemuK46Gt67prh5RSU5cxKGfLutJGD36MqjbGuBSHSTPbHw+U6I3R3g+0OBTSZ8B2aSazqmy89zGo5WK5XLQpaj+iD3HRyO1DURQ+RbZjzgkEUNL5qCeZcGKtCSladU4UWFYMTc6UlQAySloosGi5gz9Kabeua+QTcrk4yQS9bHZHaVzIXHhsOjA6BzsP4837SstwQH3W5iqxPpqxCilhi8l0vfdt22L1EJGGC0RQ72XSDpHN1dUVrM1RtMRBDzJHGSCUgQQ5puFiWKzZbAZLcrvdcnwopYDbKGVo7z+duqqqQkinU7dYLMqy0Frv9/u7uzvwBCXyNrhhCkVKhqzguXP3AP41F+Ry3YirjZGPolsiKGViVCnpm5sbUoS3b99S1QDXGtozsw7fH2NcLpchxKaebl627MYQwpv7d4+Pj1dX12gNwULGbO12O2NMUdhx7O/ublIK2+12Pp/ClUJ2m9vJEDk4Kpxsnt5yuRy97zcvX758gcZ2Op1+/vlnpZSO8fX1JYRxMplsNhuufxjsfDl7fnk8nva73W6x/LchhsVylqOuh4eHyWTyu9/9jvU9Ho+L6cxLI87nz5/fvHnzl7/8hZ3vpGNutVoBPNAVT50AlB+PwgyJ3W7HZqADgxIRyRxBfd5a0NZpKOn7HmfGEaCHQ2v9/fff//73vwfMr+v66enp6upKSTPK6+vrzc2NvpjoZ6TDtOu61Wq1XC7R41IS7P4N3IMtnNbax5C0GoMvqjJpFVXKBT0rHWGYAyAmTinGSws1Isg0vixxjFnhACvBtYGnwGfyYcMlRKE8ZVOVg1AcAIcfYlxVVjGeYb4cs3SihAVIUtd1bjjQ0maBqTLS+56PFv7Mew8yqLXOy5+djRVmJNYBHTRr7dAPSmm2xSh62vFCDdvLGJ9sdqMIABDjyMkZjscjD01J7WSxWCp95pKCbtERE4Qgz3c55+qmbupmfzpakTG/RISK4qxBRnsKZno6nW2257iYDWcvyHZaGiDyy17wO+n9ydXaVoTsMQFOOkjAW19eXpL0Qm5fX2eL+f6AHG+w9izsTBhOpp/p6k6GhZiL4SIZpcxG3BhzdXXF8IBLv0haUIlAKZV/4oaUEqcu56n8F1yuFrlWsGwtemLk7+miVZ6OPIBsog3UL7hm8IGiKPAl/P54PBpz7quKMbIn7UW76G63Y2Akm58RmKSesF3pKjdCR85YKE63bOrT6XTs2tv5zFXlrCwW61U79FV1bpl0zt3f35PlzGYzkxSJGkqr+Jj9fq+0ub6+dhc0rcVi8fL8fHNz8+nTp2EY1us1Z3O1WsU4hDCGEGCy/qf/9J+urq4eHx9P+/3d3f20blJKi+ns5fFpMp8spjPtLDjJfD5/8+ZNCOHq6qqZTJiXTus4e2k2m1mlfN9/2e3J5r98/LRarXb94GN4ed0gG/X999/f3NxgoDM0ejqdqG/BeQ2i4oMa+Z/+9KfHx8fn52eqFETAZVleXV3lYuHt7S0ryxG4xE4zcQCbgOuldxpLmGSUZC4dEdxQnrm/v8+03Y8fP2bNXb6L8R4Y3ufnZ0SCF4sFZ5/iBP4mYx6FiAD+ZvfwK+bS//tf/tkVRd/3Wil45UVRlK6wxoJTs2VpPOFUAOInESYqRSuCwxBCYHBgdhjE1zxKc27Q3xmjJ5NpErokRhZjzfrhdTBDSqkkaqC8v6qqsqpU0sbosqyqqkKxtes6bRRCNFgWhiMmGQlLfUVfKEBg7rOJeXl5EfBEo+KZMzWWH0gkiRRXPOtGoLDvZrO5Umq320+n0/3+oLXxPkwmkxiTUhopWe99GP04+hLMvT11XVfVtXNOqRiCN8Yaa0qBp9qu1easBlGW5Xq9xrRRcH552aCeXRSFVtoaE4TiAoTFh2ulnLMO3Mnovh/Ks8SWj+k8mqmWgbJKpswGebH0l1A44I+ROWu53M0eyH4atQ9YPUpm/kBNwYUEaWCELJ9LXCwWCQTOlU0COFsUxdAP+Wp5593dHeXcp6cnFg73iV02xnCG+V6gZGL2+Xz+8PBArIO3DjLjCPM0n8+pXuJdCAIAjnILNzWeDEqAj5VlmaJSWk+ayfF46rputVxVZaW1yaNslFL4+5ynZriDjcrkK4wXv2G9ttstz4SlwW4SASSlETAuyzJ4f9gf1uurcRjK4tzgqbX+/e9/T9/rZDJRSR1Pp7ppEEeAOxeTivE8/vb65ma33zd144Ov6nPwxz+RWpVl6ZwBM6mq6rvvPuz2++l0OpvP9vtDWdfNpElazRbzoiyqupnOz8Peb29vrUxYcc5VdZ1UAscnHmUz7HfbqnJJxZRiWRZtdyrL4nQ67o97Y+zz41PXdYVzfdfF4Oez+WQyJQbHiL2+vvJ1MQTiv67rdvv9zc3NqW33hwMTvbQxpIPL1aqZTJRST4+PzWRCII9ntc7lcafULbCHwHSUsq+vrzkIuY5FbRxeE56J3Zs7okgvcDasNYi6kXZFbCAODJvGD5kUgP5YxqD+1//pT3+te/jH//wvfhhVTCnE0hXdqa3LKviQ2z6pkhG2YHAzEShPywnS3wEygBXDuINUcDLDWci6HcbeWjOZ1CEErCG2DwcrAhKow1IpLbRWztmicEolrZXWqijc6IeqLkP0ZVV4P1hnhqFLKYGYsZaX5ApMPLZMKQW7oK7Lw2GfUlQqWWtOp2PXtXV9NlKr1Qr0n1s+Hk/G2BiT96Gum7btIHNOp5OyKqzRp/Y0+t77QRutTYopuMJaZ7ROSsWmroaxH/2gVer7LqqojTq1p5SiK1yMwVg3jH70vqzqyXSmjY1JJXXO1TgnuTiPCHHXdbRbNk2lVXJOpxDrshz9WJaFj370o0++ntQx6RiVNkYbq7QpynL0XhuLgdBaX19fQ5eE1cPoLpp6ctaopOeRjCE3HuZ6AJE+G3QymdDiQDQaRIVQKRVidIVTSuUBMiTaiN1b64yx4+jruoFY6VMw1gYfog9Wm+BDYW1KabFccni++eYbpoNB8iuKInctkOiAXGG1468FYLLXV0I9JIjGn1HFza25mZ7HaaRfaT6fQ9Rmd/FODmo/dPf3d11/itEXpTNWa6NSilrr1WpFd1Elwla4W0CeS2fAbsTrGFGvckK/yeELPw/DUDpXOBeDj8GrFE/Hw6Sum7rkNrkLwHSO2HQ22+13SilltAIATUkbrXQqqsIVtp7U1hpl0v64b5paX3Ry4au898bY/f44DP5wODXN9HTqjXGTyWzw3lijrY0p7o9HbWw1qeeL+el42u1219fXSqmmaeAvKKWOpyP1DKS3zhmS94WzBDG4Q45nSq4q68P+OPRj3/Z+DIfdsSgqBqXQpJY7y0IIx8NhGIaY0jAM17c3s9Xy2LVRpa5tXVEUZVFW1fQsBm5GP7qi2O520+l0GEdlzHa3CzF67z9/+kQAdDweb29vM0kSAMpJyydhFvouHBbOFzska2opUbzOfGiMp5Mh1U3TfP36lUgCB59JTZkgk9n/WIl/++fv/lr38B//n/9K1JxjCr6JrnGlFN9EljSI/KeRDvUMFlnRhFEXHaeZ++hE89YY4/3InRRF0feIbEd2NqYcGBfXKtDQmYmfL4l4zZhzwmGkPYQMnQsbZXAud5ohWr6a7oeyLEPwnz59ohyarRueTCmFZAXQ6jAMh8Nea0OBF+BVy3BKpRRuPPt53H4uYBJTQChUSCsbQ5VMa90PfRJCLbaVNkMjNBu2Bek2m9Jo65yLPmitjVZF4VRS1hpnnda6Bzc3OsY40rSozv+LNcnWh880xgAWweBCs4yInvXNZIFc6EaeM5cHcMbjOD49PeEbDofDdDLBACG8EWUcd1EWy9WqF4USgmVjTFGUVVVBQKqy4POk6frz8BY/DEabGKNWKqbE8BytNboLfBQtfkpUY2Gse9HUU6JEyWJBtmbDgCmxPUD/idd4PoPMDQ0hrNdrJ/pjFAlYa1CFq6urS3CsaZrn52cthNS7uzu+mo2KM06iFlNe9PDn6osSaqOXKafskEIGk5GyFEUBk2o2mx2PB+cIp7QWAnfuH2YhPnz4MMh8EXq44MuCd2UmFf/VWiOwOg5jL/qPh8MhvzOI5HtKyfszhrNYLNqud0UR03nWFnmYtXbSNGTqq9UKLNcYM18sPn/5XFUVWfJyuby9vXXOOWNKV1ZV7WxhrVst13Xd/O53H05tt91ugbkwNcfjsRcCAvDLcrkkvfPeD32vZSLy9c21MvrPf/5zjHE6mSyXS2KvIC29GduMKVV1bYyJKe4P++v1upN57PwXiiZOvZAWYC8v8lRgopydf/jwoW3bzWbDasJ9SNL0gFx8EmYpQdXHjx9jjPf39zg8qAFOlIzJ47XWFCH+/d//4a91D//lLz9yNWwaHIAVImPOWAlhMoCupVJtpMuB3UPyTlAZpAV3lKm5BAIh+N1+hyOFIde2Hft7s9lk3ksWZogxWmuI4DjthUicp3Tm5vP7/DPMn3ShMa5FgBZkgI0LeujcuYMvB3254kJVk/OJ1th+fwghWpkN4C60pjNqT71UyTg8EvBMp4MubLVhzwFwG2NC/FVPGVUKqiP4A7R0+C5gDeeK9tQarZWK1pjZbFoVpTZKa6OUiioBTHnvi7KYz+cxJgITzD07mEIRrpFZN2xH4msrM1nzM8mOk/guo6scA4xy7vA0xswmU8iLNzc3YHHsmZgi0T2oLst0OBzo+8uHhONxd3/Xdm3f94UrTseT1rplMsc4VnVFNgM0n+sTeD4umL2tpDMGT4AFBODOIOrZdRVFWZYMkMlZETEaYT7ehRyO0IqOsGwvCHdwPGTSg8zK5UxRReP3oLg5qSVQu7wdLA7O++XlRUntEUSOkskgQorZTPNLDiYPJA8vs9K1StFVKUXkjqtA1KGqqkxBBA/Iux0CEohQDhd6GQuYoyXGLBZFUTd1TOnq6oo3AyT0fW+1YUwFoqSvr69nFlNK1AgzXFlVlTG6rEokCkKKZV3NF4u337ybz8+bdrVaAU+llBh6wdVqrf/4xz+yzZqmaU8n51zdNGVZLlerqBJliV7g+yCCkjmyBCzNM06+fPlSFiXWD61Gtg09K0YqxkY6BNjMJAGgrDDi4Nft93uYPpl5wTyVEMLnz5/ZUcMw7Pf7+/t7VpOOVx4LJALYU1QmONq/2T38itg6yDQ0AnBWjl9m2D1e8CAxxKRIShQfC1EHZBvl5xJkNBs2vTq3Z3sSya9fv+73+9fXV0iNTdO8e/eOT9AyDC6HbKNMS89DQoLopkUR92YlaEoiEKtEc/sSzIXKljnsIQRI1l5mVnN+yPU4JJiMKFKvUSYrcIxLmTqihDmapG6JZVkulxDVQZB5YZdzCkmV0hgD37yX8UewpDmBKAZntGGxWNRNPZ3NpAmjyFknGxQqEcbOi7YXvAsgHS3l1lL0hTIKByRKYVmJKicPJ6f2zF8CHuGR8gDv7++By9F1yEzHnJ5qra1zbO6yLAlByJCIKzmKTjrbiea0aPmd97Ex2AV8c7qo9mthLVuZXvD4+EgQg9WgdMkpQldZSdtqEGYEvCwt3XAoHvIV/YVktxNSco712A+DTMnm+tfrNS4h9waSS+U9jJngmjlQeTlwZjnA4vIoL9MRlsdmkHnDB6UVg1vmcwhIs7Ns2/b7779/eno6HA6UMfiiHAJyp4B11IqlNpMwWByZ6XRKIo5xIPjgwIIqI3KWk2l2PoaFh4/sNgjeZrOhGRACDzFcVVXrq5urqxvnyuVyfX//tizrlPSXLw+oEKI09f79+6urq/v7+7qqjNafPn2CfuqcWywWIQQUW9kDSZSGafAmmsTDZUgcEgEBAXbDOffNN9/gY1gRGiBIjjNPMtdiWQ6oSiZPoQ+ByKOu6/fv3wOugDiRjyqlvv32WxLxr1+/8sQQFe5kdChmkCAvY0qY3ErG4v6G16+yh//7P/7n7AN4ZHnnQeMBedfSoWaEU5XxEyXFkyi6j4SoGTZh7+Jdvff7/U79UleIwzDEmECEWDmSMk4ObgkKILFtzv1JoDDTuAQOtvceUUYlM3CMMUivWGsZnIKhF/t4niGcRPGJyL2qqqenp+vrazpZKHF7P1JezjYX24TT4jY7GQLO+vGBKFMaGVJotTHGeLapUimluqmzQyKsSNJMC4OeuigR/RkdKuu+66uyTCkprYoCq6f7fvDea2vKsmQOmkFvSluK8Dks8tIZjkvAzYwyV1VrvV6v2dmE56ATGAKlFPuSD8nZA3+rpC/SD2e8kQjACoXMB79YLPAo6/Wah397ezsMozGmaRqqo3xa0zT9AHJojvu9VucOD2ttlG7nqqpAdZVS+/2ew1zK1EbOFRsMo5ZbwNgk3nugJCtt9kzUaS6m5bDfMsGJMb8QBFig3MFDtqEvemuMMewBACUr6vG5WAJNLtdFcgafY/+cjisRRkSNJwN0mGkCoLIsSTguPXoh3RKIQ4BEwaDv+54GNG6NOzVSYsHWsBMWi8XQD33XMQkc0wzJB9dyOBwIB7nm4/FonTPWFEUBj3axWGw2GyTBs+MnnjgXvZYLwPrcXt73fVmUzrmiLE9tG0SMeb8/TCYNvVlN06CTeDZlWuUFWq1WqP9qrVVMk8mkbprpdFo39TCOXHzfdYR6GSZBd480i4CYtSBuwF7lxoVSJiizB2h7rqoKhVMvfWNsHi6S7ij0Vl9fX8k1AcTYKq+vr7gNYwyZrpaZnvWFshF7G6tohBXZtu3//r/8j7/BPfz/dE1nFIj7hJGSpP1HCdv9l4+QiyBWzW2lGVehB02JzFG40G3mDssLTXyCFI4xezEnNMTULBJvyKkJAaOTSRpa2uU43qTY4l1KoGeCXKxYFB3Z3K8HBA9klFPpvu8Z5aHOc1qGRoa+Uu0goMg0VmNMLZLaBM78U260ybeGIeB4zGYzoIncxXYppuZkfCnPzQmHZxgG6BZVVTlrh2E4Hg/eh1wMYFNyPZh7HFuSsZpRJJoxUjR/cV+FDG0lesWYGuGDBiGhRpEiYNgLIULum8upGLsiU2uOpxNTeTHHLBAriBIGBovnXIswqmL2bTrvQIzvKHIpwzA8Pz+TtithoHLZ8EYIAHF+XHPejcRiPFvcpz6Lg/0yFyzImOK8ScjAeEQ8NO6djJngEUkM6aSZZuPOpxFj4Tjfvn3LweGf6L0o88wirXnafDXkZu7x+fmZ5xBjnM1mX79+ze2ot7e3PEBWn7CPWIoNT4JCvJ8RY84jURF1RzaA1vrz58/DMMQQ8KBKcrjlckkEhg8gBo/n5omw2215MqCORVG8f/8e/kiSlikCFCwvuk+VyKfzUdZYRkPOF8v5fPnx4+fV6mr0HmSGbKO/UNtOSf3ud7+DY/r4+LhYLP7+7/9+uVx+++239/f3iJ/jDtkSlC68cORYF9wnOw1g4+zaBU0qZWhjlPZ7Pk24y2a9Xl9dXZHiMHcEu0R1qpB+0j//+c9JZkgQxUJYAJg1xiCIy5Zjs3kZwIzcEZuHnnMlU6J/w+tXfQ9a+ixghSsRBRO8wvmLmQo5lilEILOX4WWsMTaOvMFcMJrUxTyflFRVn9uJi7PQt+dj+SInxXqcZFEUNKVFUTjIAg+YdSuUKpB9mAM5asOZLxYLBMKc9FKVZYljZ6uzKfOAjuwRgZ4wl1rronB5DFP2ZPyVlllvvDOllBud8O3cLFyXszMrHDBLRif0Bc03I1RsSlwgz4EQNaZYnMW8CM9jiskWlj/p/bjf70OKdKix9fEWhEiZsHRp/rgRDFy+kSR8cCMCHl4mP+eHE0I4HA6cMSrA3KMKv+jDBNGwIibIkm0UHkjPs4OnFVYpNQzDZrOZLWbjOPokulEp4Ym1PRdIlajgYU1wDABcQbizUWRicfz5ynknW8UIhYFViNKFri40OdjnfAveYpSxd1jMpml2ux1f3Z9nYZpcD8BeGxkGcDwen5+f//jHP5Jl5nacKKUmYwwgZyb1RRnPx8LBB7MyQ5DQNZ+RUeYs4bomkwmZd7Zrm81muVzmOsp2u3379i3aghkMDDKFBabW58+fAUjxOovFAvhOi2JNSokooarrtuuIUWDvUP9rmsakc1EQE4+/mU6no/eYGhaI6aSDH5tmWjV13/emMO/ef3M4HZfL1cPDQ/blFBju7u5G7+vDQWtNiHZ1dfWXv/xFoMWzYFdKqawq0l/vPUESBZVKxEMnkwnV4/fv3y+Xy5eXF47Scrkchb1Jnkq6f1mtoRyN9c9WCC8LHsCGX6/Xz8/Pd3d3eb8R0Hz+/DnGeHNzU8gURZDP0+m03+8p0pA3gCJeX1+zZ/jw9Fv7Hn6VPQTRIAvCnGOL9F03yGBkpVVKKaqUfUmOr5P0fGV4wYkWUzaOOcbnDr33MZxRV9bGOTufz6y1s9n0ItwuswOM0tFqRZaDnwkPvfdcDNFW7jXFc2CXnQhaIKKy3W4ReCnLsq6qqiqvrtar1Sql2HfdOA4h0KxwZpJAz6iqCiGB4SzWlrqufX5+HmQMSP5e4lN2mJUml0LkMRaLRTJaO1vVtaF7K0VrLe0aWqZEWWmPvAxgeQLkm0XhjDUhhphSTCkl5VPySZvC2bJUWlvrHh+fOKVFURaiTpEzAy2D/zLUFkUXryxLlZLRpqKyok0MIfqglc7pFzWSzWbjRXeeJYZHwX7omSKVUkrJWBtTMtbWTWPsWe43Y/3e+6ZpqGmllJbLxefPn6214zi8vm7+5Z//+XQ8rlbLZjKxzhnpqU4h+mEsXXHcH9rD0Q+jH8amqgvrdFLRB5USkkQ8fzKYtm2pfOCSEc4k5VcXMy1yhpGdKM6slaGHVVX1wwCMRpCRpImSt202m1x7DDIlyUjlnFXOppwXazQMA7PAGBeTLgYx8bQhFBhjqIhyhVQpvbyOxyP3bmRS4WazmU6ntIlhp3Ae2V2lszMYtI7G/JKmE3MURUG/nrUWUB4LhdgUBxOmgFLqdDptNpuqLAvnVFJ+HId+QMpit92tFksAMXRigijZkFNbY5CvmM6mSqm2bV9eXl5fN6dTW1V14QpICu/fvyfaMMaAIGUvPp/PALK6rnt4eHDCARu93+33wzAU5G11XRaFjinGWDjX9/3peNRKWWOmk0nTNLe3tzFGZkLM5/MzrVEqZ957qNUxxmEcldZVXYcYmVlyatso1KNhGGBnsFI5P0C+6XQ6UYR48+bNYrG4ubm6ubl6fd2M45BSMEbf3FzHC6UfcKcY448//vj58+cff/zxH/7hHz5+/Ei3Xc7tfsPrV7WH/+cf/3smG+RgJKW03bxqpYyzxhgfQ1mVriyMtYVEghwVEnYcST5amZlHyJCNI4HSMAzT6Yw2sfKsvYH2SzOdToehH8eh77uqqpyzxujMzCMqJJLq+17rX4DLTK7gxJIfBOnuzkBKrsvxoM9Jxjh0bRtD0CkG76uiUErNpxNXFjGGqqoeHx/quioKlxJTa6bGaAQquq7dbF6ur6+0NtxdJUrXQOEZ1eGF+UgMihkH65zSOsSgJFbl6RFgQtnmwPAAMVsEpJIXD9roEMLofVRq8L6qm9fdbrFaHU+nEOM4+qaZKKVxyU5E8bBf2DUl1SaAgrx8JinOCcfVaAPQBxuq73siVkpqSgQEe5lUxfkx1iitrbPK6KIsjbVlVfrgWcpsIvH08/l8Op0cDvvt9nUcx+PxcHNzvd/v2rZt6lorPZ/Nj6djURYhxjH4m+ub6WQyiFhL27aL+TylFPpBxVhJ+W4cRlcUFCHBBoMQLrIoW5T+6uwkyGkytsM9eu9DioP3RVkmpXwMwzi4oji1bWFdzkUyl0kpRdWKvms+vJCufvotyrLc7/c3NzescuYmLpdLfx7ZW7EiJEns3pxeYJrfvHlD4MKHf/fdd5l/kf1NjJEJmuM4vry8sNC4wDh6P4zTpplNJ/PZtK7Krmv7vjfWZQ6lUgrt4cViMQ4DAWxedNC2fO4mk8nLywu2pSzL2XTqrGtPp8NuXxbF6XhEuHQ6nbLJ+eRhGKwx89ls0jQpppTUZDpR1iSlvn79sn19PR73ZVksV4vpbNK2p6ptnnXkAAAgAElEQVQqF7M5gGSWhizLsqrrsiqLsiQ6JJ7DGw3jmFSKkpobbYIPzrmmrOqqVjENfT/0/Ww6HYdhHMZKsNkcN8A51NYcT6eklSuKkGLbdYvl4tSeirJMKk2mE2Pt/rCfTCej98f9AZoAiBP0WZAMAFUrA26HYZhOJ87Z9XpVFE6pOJ/Pjsf91dUqxgQq5ZybTCZYGwBG3M/z8zMbxnv/+vr6f/2f/+E3uId/nT2QEZNSESDw8/ndxkyEEUxiYWTMA/aaUBG7x1WyHdlVmTCK6wbNpL3wzZs3V1dXk8lkPp9zREH5MfSFjNZDn5K6EDEpqE5GS6NMayGWIW0kwsKwYhajEHX42GEYOLejD0rp2Wy+Wl9rbUcfm2baD15rU9cTrY21zlonqjXnBl0rVCXvPaEN3A+yVCdMZCXSIBnmZjBDrsRwRHN9MqPbVG4AWDJ1KrN3+HCebU7gcBh0p2vhUyLhQlyvZXhDVnPK6Pnl5wShVzpRyWYLntr2dDqVogeZP9PKHEA6zjC1FHK8aNgZaUyJF6QAK1KauUx1bgpRSknZPMtaEMBmUhYLPWmaJDxgDCgqWDFGHgIJ4kQGtfK0waBgpuLIi6JggEyKCbB+OpkuFosC6T3nyqKYTqfz2ayuqqaqYwjjMPjRV2V5Op5STElGpscYid8LaSSCG52DA7DTfG1wftAR4phAF84eNwlpW10AgDxnKtgwnoEv8GcM0crlkCwSgy3GMVih1gwyaqZt28Ph2Pc98SzPNm8PUNZzOX0cgSW5Ed7AvYP+a6H2J+GP0ZKWoV1jDJlN1sf9/vvvOT6w0jkvjw8PSmnv/Xw+f/v2HVOaaXNDWpVbAAXCaJwhO2ursjLajKLKbkiUx7GSOaPAj7n4oYSQfXd3R8IE+4DyAPnNly9fNpsNxyS7inEcmZza930MZ+1FldR8PldJDcOg1flo7Pd7Wk0JATMPgtNHDQmzWVX1YrF68+bten393Xd/uLt7czp1f/zjH7/77rv7+/vb29v379+/ffu2KAqaTthFoOVa681m8/j4+Bt8w792D/7X46A5ezi0nFafqzF1PZ/PMtuPHZxNdrbLVCzzzmAHN01DrWm9XgPSaSlQY6qsTHUHVWdJCFg4+W3bAiwWIvSWwRCg3vyH4EK8TQtdSl+M7uParKj9aGuNK4wrBu+rpklaa2tCiikpa23X9cvl6v37b5fLpXOFc0XmpeRyizEGlKkQgi8hagZb+C4eRWZbeZEGUpKxYUmBs6kS57XI+5WdTYTOL8HxCEMyNTuJMgzHng2N0UyioF7KIN+cLkQpunAX2hgfwjCOjHxq2/bp+TmX4pVwfkibbEZ7pJbITggXE0k3mw2ZL5COEl41mTKhU+5xJWSRzvZzcz4AlBLZ/V5iEaUU8Rd4Djl4ZolkJeAshspOoCHcOTebTIP3VpvpZBJG74ytq2o5XzR144fRj74syhSiSqqu6qqq6rJSMR33e2fscj5PIVh9zsa0gPWVvIz0kVhhS2d7hNPKyLWV5sdMVQAQA93Of5h9OSab80URKIggDWALZp2cA6eCxlyQtt7M5oAzzelARG8ynRbFWVaZsI+K62Qy+ennn5VSt7e3XEbf9/BrwUzYaUgiQo7njOOSUSWhy8GK9iLBBPGQUqpt29ft9vr6OqY4jENKiYnW1JMzVlwUBVXZm5sb6ueZUky6MJ1MjDY6qavVOsUUfTgejl46e6i2cka8dG5qmWxGTTh7QQxg1t/FHMHFYKIquFNVlilEZ+zQ94V1TV1HHyoRdttut16mEgAtuAvxaSNTdZ+enne7fd8P6/XVen01DGMI0VqnlPruu+/GceQx3t7eXl9ff/vtt9988w3naLVagVlFaST6a90DVBklrcjUFZTMLM0I+HAmrZscw+a3EYxTveATIMPkwKGSGSasRM4P0q9lJrEsGcG/vr7m2pKULrNHycVVDk8vEg5BXsAa+RjkjnYufpTmBsyoMda5oqrqoiiruklJaeMm07m1brlcz+eLt2/frddXzpXj6I2xGYbOrIOqqhhARAkkH2ziAvYTbF28RVmWTVMHma1mRatZDuc5zuJxXUJ5OXIMF6pE+X95IKXI7XF+rq6umJTClWgZ98gPveh7K+GJJqloaa2VVlVdKa1CDP3QK6Viit6POWIw0hJMzQmDy8rinjN2hGnGGuKquV8t3UCDzNweRJPDSCd/RmMwMYV0oefgoJT5i9keORnwx87JO7kQ2jheB96nUopmIq31er02Qu1VwsgYx9GIXkgMwQ9j4YqqLHfb3el4qutaK1W4AqtXMuLJ+8y6oTaA5zAi3G2FhDafz5+enjBAQYQG8BBR5JDJiigsJ6GeE17Qm5IxHAJqtgFHiR2FHhTR8TkXPJ24I7x1SolugxEtZGPms5nWBrRqvV7DHLPWvn//vigKsrQMfNGj8OHDB8RFeHqLxYJFh7EDsIzWFkvJ2rFhKLRqGR3Gui/mi+3rK/Ykxtg0DS1TOJXpdNq27cPDA4/LyRQydlR7OjmZFwA1az6f+3GEkUE43/1azTOjdsYYFAH0xfgAkpjsh5RoqTrRDcMzdV2nlRqHoa7rSdO0p5ORmbs5VyBD9TK/pJHp9ISSxpiXl83PP388HI43N7fr9VVZVlT0eFz06BH49n3/4cOHP/3pT4vF4ptvvvnw4cN0Ol0ul9mQ/lXuoRYBMv43h9Uc10Ga5oDdOc9KiBN4VPbiKBKybNAMnpAaZxTICJUi5wrsiWyntIz5ZeuDaKWL6XoZ2Ln0TF56vgAWSGUuza6+mP6WS+h8V0zKFaV1hbFOax1ico4hCkkrnZLq+6Ftu6qsVDrbOI6u9x6+mjHGmHPzFM/kcotbmeugpdLeyw0SDl/mB5fRvRFpUiVTTXIEzWbFmHLaeYxGWDdEc9g7yuk52CQa5VETzlvpHWONcsqotKY6Aqm5njQxJS7e/qJrco49c8qCX8zcD+ID3ik8NAXukW1cLbr8xGW5VgQFrpd5kFVVoWWWa7M84csDXJYl6IQVihf9XMSMGZ7KLgpzXFfVYj7fbbctiX/fxxA2Ly/jMMQQ8A1GayczA1KM08lUpUTOoZXWSsHYIVrHRhMJYV5xjXgmivkQTpxzq9WKLl8sdQgBhgzFbRa3kPHCUeQuwAyVDLoJIcCVwrtgu3MfL5+jlKqqCiINajE8EziX1K6JcB8eHoyx8HbihRzpmXMhdKkofLC6rn/88ccYI/hPXhfnHN9IXE9vB6ga2laUrwj/lVKw+0mUP336VJQlEgBIn02n0w8fPrx7987KcDSOw5cvXzBTuauX0CeEsFwsYggpxq5ttVLL5RLqM/MbNpsNqDVAGEVNL33gRhgEVoavcIXurP5S0Nr2i0NqW62UH0cSiy+fP5PNt6cTvylkSBfryC2w89F8zPMqCldut7vTse3a3rni/v5NU0+GYdRa39zcGGPwx5ym4/H45s2b5XL55s2bP/zhD1VVrVaru7u7v4F7yByk/Jsk46OLC54rjUjpQk4V+0hMx+4nIrjcGUmqZznpzn8LCJ7fpn49SKAQVR+OHG9DGdBIV6oWBF+LggVxWf6BdNuL5LiReY2ENlaEkYdh7IdxGEdtrNLWh5CU8j6kSEm1KIsq+GiMS+n8cHIZvJFhOIWMO8bK4zKTUKcypnSOH7suG32sZCkthFEYlmzW+mIAyCXyczZSUuzhzUGGImAUMk3z0nNw9tjfuT8r43KFMFlBupNW1rlmMoEZpYy2zkUZZJ/bOyppTddC+sxFrCDNFixB3lq5DpHppBDDyDbYSACPZypLjDRg8zD5Tc4saRvGxxCe87FaONbQZFkXACuMGvF+d2pTiJO6OR2PVVEyuCb64MexLiuINynEsR8O+72KyYhm8Hq1Ct475yohxbEHnHPgyNknlWUJcM9KkQfQVPjy8rJerymJe++32y35x8PDA0F3KYp7+Anu3XuPpBUULFIH2rhqac7HzM3nc1AjNgC9flqaQFlKlO+qqrq+vu667uXl+ePHj4+PD13XcgHEZLmqN5vPqXJprZ+enggpDofDp0+fMuAcpV8niH4RQA2PiKeRT673nmkKb9++nc/n2+2WwqMfxyDzOYrCEXTjUYxIBbPN2GngBMjGpBifHh+bqo4+OGuHvh+Hoaoq5rg55+CVstsbkeKHHgoIyS1HUVIBPcsgYU4Wc3XNWYeHCN53p/bzx0/73S6Mvj2eUHEH0cLTgGjRZqRlTjvY2unUnk5t3w1PT89FUTpbpKjoHUZzqaoqOgGxQtCa8VU5RmnO84P/OvfAepNjRqlMYnoySKeUIusM0miWkXe2eyXCEiwbJ3YcR9LJjJInkU/JljHnE6VM3Obbo4g/ZwAqipCOkkZBjgopqpI6GDdF4HmZ4HPlhELWWhqwJ5OJ0lpxVEJ0zlV1fXN7V5aVOZejVVlWi8XCWlfXTVVVWv0ir481x7CSVfA0sLZOOH/YJoqE3LXWGk8TQsjBF/41J0b43aurK3yYFW5PfyF8krGmvIK57pcD/Jy8R6kxZmhSy0SRy9W0ko/jJtHvtIXDN9STpqzO8+75inRBUvAi8oOaMY8dJZwoo5NyZqkuYJMcWLCySub2aGEBWKl2ZlwrZ/2kSlCAsPtA4VxMjj8yn4r4GqTl3HhfVYVzs+m0sK5r2xRid2pjCGM/nI5HZ+zYD1qpFON28/ry/Hw6HuuqOuz39H+ByvtxXM4XpGIgqLvdDrWGHBLl8MgYQxE+r6ZzDlkXwh0MMYuVLubp9tLC6USp3orkV1EUGHcKKrmUrWQWOh4Xa0s1m4czDMPpdKLOye85icfjCcyS/moeOxlz0zSI3KGIc319PY7j3d0dKQulXdBU6uScxExayXsMoEZrvdvtuGDMHFc1n8+TUqRZnET6wipRxcBbxAsZY1wRcU9dVkPXD31fV5Uz1ih93B9SPDdpjdKXHmOEypXd6iCigRkEwxBfXV1h0ApR+AcyVTLCQWnVta1RurBu0jRD3+9et34c/+s//VN/Hn/djuP4/Py83W5ZstfX16enJ3zG3d0dvW+LxaKq6tlstt3udrt9CGE6nTlXEAbtdjuSTvwEGQ81GBZ3tVoxZ+lv4B7sudv2yDYdZd7h6L2WagRx09D1Y9f3MpVFyxw3PkfLrCXi6Jy/B9HnQnYigyGHw4E5StnHcDCi9FFz8wBt+8M2xFHpqHQMcVQqJhWcO18et5DBXKS5lTq7imEYi6JU6lyAyuE2kIWztu9aZ61Rqiic0aoqC2v1tKms09Yaa7UrbFk6rVPT1Nr8ItTBBRNo8H3b3U4bQ0lhHEdHB1NKMaVT2+72+34YjqfT6L02v9wsGpl8oJcx0dkNkybzlKC6Up7NWRevy7KEkYkL+Q34zkxuwWTP53MCKJIGrBXO1Xs/9H0YvR9GnZRRyijdVNVyvnDWZWxQS/efFTU3QAwr/VnZIJJM5FCLvBjPFGMEoCtFz0DJADKsJw+H6mWuonMS2CS5deOXHMto5WzX90prbfTo/VnpxGhbuGEcq6qyxhSuKKyLIbR9ZwtXN82xbaez2eD9+vrKlaUunHZ2dzwErYJW5aRRzvoQtHQ2GWMQngK8Bhxg/JmT0WNYc/IYWNfluT+reHl5LYrq9XX79u034+iV0sMwKqVV0k0zTVHNZou+G5Qy1jrOP8aLgjZxbiGcXU4WNI1cAx+kA5z1ravqdDiWRRF8GPrealO4Io6+KArG/M1mC6bLWeuaZtI0DVoa3vurq6sQwna7Pban1XpVN/XLZuNDMNbe3NyQo7y8vNR1DfwForLf79u+T1q5sri6vmYgnyuLMYSo0na/CymOwXfDsDvsQ0qGqX99r1KaTSbX6/WkrMdT10irHV5faw0hgn3C/uGfcHXGWePc/ni8ub1dLJdv3r09tqdhHOqq6kTuc7lc/vM///N+vyd1w/FkRA6phb7tdq/b9nhqqtoZ8/qyaY+nsetNUmEYh7Ybul7HNJ9M66IsylIZ3Q1D0vrUtT99/HhsW1sUHz9+zJKLIQRUWq21THUlTby7u/v22293u13fd01TGWuWq0UIiKS9WmtSSta5nz9+BBR5fn7u264qyvV6jUHm1NCcm6mnf5V7IEbj6ObA53Q6jcFbd4ZEYozjMB73B7rJjHSQEp9eoiscQiukIKwS1oGfiX/BWJ0w/LCSVHS5HgZ34BtzRJxiDN73Xe/9mOK5FewypUjnDpQhBK/UuaJ7PJzGwVvjrHEZ6YPTQpJUFm7SVFrF5XxWl4WKwWpdFM4YfWqPXd9aazavL58+fxz9uY6Xw2e+/SxE6qwy2sdgCxdSTEY300lkQI+zgx+jSlElhvHxoDAuMUZwSRJDngOuwoiGOfEXEQ1lVZ454Q//iniZlmI1eAXvyfAocXclwuOop+RIPIi4KTGjSmkchvZ0MtoYrVVShXOMCM6ZItFrJ02UeVPlS6LJnKvNLFuSyCAjTS5Z1Dkzo1GAK6FejR8KQts1xnTdubsN8QNYVVprbXRUqZ40/Tj4EHDG4zhWda20jirFGFVSfdcBSQ/j2HadMtoVzji7Wq+U1mVdaWuiVsqaMXjjbD1pmumkEU1mojYr+rXX19fz+Zw0LtfnibdYQcSIlJCGc4H96ur69vYW4kNKKiWltDbGTKczP/qu6/e7/ShCWLkgQWAbpYU+R8R8O85pvV7nEiWLNZtMU4xVUTpjulM7DsN8OqUOwUlfLBbG2JS0UrZwVRaI5NiS4Hrv275vptOiKpXRz5uXxXq1ur76wx/+8PT09PLygrh6EFGT/X7/stm0XVdPmmbSnLr2+eUlpmids84VZdkPw/6w3+33D48PSaWmaaghV1WVQrRaO2vHfjCilENU4UTJTinFdkLgjxWxzrnC7Q/7583LGPxsPrfOPT09MSIJWI+Y9fn5mXj8cjeO47jdbo3WZVFUZdl3nR/HqqzKoui7LsdhoFgYa+rSxlpt9NPz0+h913fDOFzfnEcDvX37Fqj8dDr9+OOPdV1fXV2tVqvT6fT4+LjZbDD0fd+PfvB+CGE8nvaH4+7UHpMKL5uXh6fHpmlObfvDDz/0fc8JRRYBDAMFh6zy9Btevx4H9E/fe+mKHKWreRTxJSOMeOqKlxlDxvEzHpItNZAxlivGyK6COMwnD9JmjLszIkIJMjCOI92PURQCxvE8BxtSRErK+2CMU0oHGYSSj4q1Z6WmzebVOTeOfhzHoii8H4/HA971+voai4zHghAJeEc3ExaKpI+scLvdVlU1jgP9sYNMIdZaD8PQiRA6n4mhB1Fh25FrVzJqQkm1GffJFWa4MEgnnRFOJC4tBxr5YVJmNNJcraTMZYzZbrccey/6RUnYwNk0s16sgpYXoTroOWFjHhKgRFVQrHPHA+S0VBcD3DOi2nUdagQIxmXPmk1kIS0L8dck9LxPSBEIjYEdtEiB9l23Wq6Q0iLNohjY9f3+cMiDWXjCTdNMZzO+wihN+7eTZnUeO5YXKxNitM4qKZZgjs9O0ZzFNobz1vJd1yWjoRjgM4zw8XK6QNaV92pV1Tc3N9vt9ttvv729vWUghLoYl6K0Ap8sy7KqSpJymsgKaT/KFD6+DohSC3GOp52x4uPxWAm+xG+YTtH3/Wq1apoGTjCWpWmasqpc4fKfs0lIUHKUjbgA2hL/w9/93dPTUyWTVpVSbdtOZzNrDSBb3/fwo8gvMcS96IpzDGOMMEFZUzbtZDLBkeQ1Ilrqum4xn2e+LJt5vV6D3eWQaBgG1vTl5aWsKpa1KIqnpyct0kFTYd8qpbI+UhStOS+ym2CkOept27ZtW9rCoFRRsia4KeSFP8OIJWF+Hw4HSu4ssff+/v4+h1aUKMiWML+TyeTh8XE6mV5fX3/5/DmlVLqibdu275xzuASKcPBi/sP/8e9/g3v4VfYQpY9pkM5nrixDvVq6t7jufIyTaBfnY6OUAtYvy5KqEXanEJWY/IEZMKlkXqZSCrgW5CdnG6+vr+M4xqjG0ceY0OQiuoHSkA1KlNlVZLWZt5e39fF0Ajdj+x6Px4eHB557JUoeVlhVFCqGYeC/TvhRSvjjGERcaZQByBlwz76Tdw4ynAR1towIhxByOw9iA/kcEt1DHmf2Mv2xGcHPXiTnfElKr5y0JGMYcvE5Myydc1TVFosFVO6MbueCDQJqg0wRUBdEUmzEKH12OeTn64zIQuQhH3waviEXS3DqAnMfrbUg0bmsGmTwgFIK7uZms9lut1nQwhgzjmfdCJj1KSUUrcui5PLYAFFoLXhKvteJNBAJ0HjRjof/4/C/vr7CtPEXw7zAD30IuFJXFJ8+f0a7jTFzHIEc/WQkahTNypQSmMZ6vUYG5rvvvmNxt9tNSnE2m1prnDPr9bLrTlor4Hgt4qYZy6Wom8HoUfolKXGz+bM2VNd1bAM4/hhfggwObI7VspHK3JAgXC+m6aHBRxzw+vrKPNT3799nYiun+PX1ValzRfr5+RlfAiTIF4G20xurtX58ePjxxx+1jDZKUijOAxu4cvZtLwouYHp8Qgjh9vYWxhRrjVtdLBaz+TwED/aV001if0qARBJZl4IYjuv8/Pkz7hOXzPcWMrMBf0D6CCmGm1osFkoGN+H88IjU6jsRQyTC2O12b9++vbu7I7IJ0qoN/k/ANwzDYrF4++4dlYyiKBgBggHnxtkVv8E3/Gv3YERQk5QHx8gjyHXLfJKddEhw8HiFi74nCkQZoGC35UGkmeuCtbXSNMuV5GfEo0fRJdvirutR8THGeP/L5Bx9oeWHjS6k9yK3BxtjnLN1VZcy9oBrpiBGXxsZA5Y0d5TwNtZbn+lP5xyL8CSImptKqRahQDacl/5nL/Q76L+MxCKys9LlFGPc7XZoAg4y9TrGSGzCZ7YynBbnQaDK+ckjqLK/L0Vmh/Qu5wf5G1lQLELOJ/hlIR2O2HGAo+w4MwBlhKOceatsJ4AjIkGu9vHx0cl0e3fRhRdESxyBMy2UNiudTfBlg5D0+74HH8/lq7Ztf/75Z2MMsSESyhg7zEcSkiUxmpGRakYkFsiujFA/g0yvyveYozlgPaVUPwzH9qS0HvxIB2VZV5PZlIRmvV6T552LW1KBCDK/lj3GZkNVAsWnm5sbKMjOudPpmFLU+jyBjnoD1irXbzLGgoUyQkVL0oHFAj09PVlRaiEPVkrBi8+Ljm/m90ibXN44+4p1oUiZbXcuhGb/fXNzs1qtMm1vsVhMJxOlzhQ77z0jbpRSlFLQA2ceJ9sD0a1cdi6K4vn52Ui9My8f7yRmDyG8vLxwzbPZbBiGH374gR2F8cGGhhDubm+HcSTGN9ICgsci7OBp5zSUKMdae319zTAlUsNSJLyM1PkI5gBsYc2SQ/BOHiCOkwcIRVspRSVpu932ff/4+MgzZy5F27Y00GEWmPXC0I5v379HBh/2CpINWINCWLl/A/eglMo+zUgtMZu5oiiw+JiM3OuvJA/wwplRkgpkMIofMkKthbyV3QxWMpcWnHTSYp2N+aUFb7lc1/UkJZWSVsrsdoeUVFXV+ZxHEe7nS5XQOrGM3nvh9ZxVYJFISyKqmgNMYITNZoP/hwlOtO7PTW1ngYpchBCs4MzNZcNhUnO/npE2Op6hFpYhhw2aPBsRf5CfAGU33GQQqSUlMzIJE8hVV6sVpyUf+3hBKeaduHYrvSbsaQLYQkTZMksVOkeQWfNgVqMI7eHXseM5peDPscjAiRnEyP6jqipCKgyWlItG4DKAEcLhUpSuC5n1zS1wwNSZe+3GcUR4I6eA2hgnnfZOxrhyzbvdrpCWLhyMl/o2ATLOlVwnixzgpXIQVxTFZrM5tScvJGYyBudc3lfZkVPexH/jpIEplFKUneq6ZpBtdidfHz77MGqTitIqHQ+HXVUXh8OeeAhOAbAMp4nPwb4T21kZPkGVMklvM/YIdtl0Or27u8MMAViBwFDZNtIjgjHSIgyDk+svphzj75VS2+2W2TUZVCFOijEWRQm7aT6f58y+LEseJpEyISnxPt9YVVX2Z+SXlQywQ0l7uVxCc6K8kYeWz2YzSFysLw+EFDYplV0mv0S0xopAIb/JGBHhEY6Qw9vIzBKWA3qhtfbnn3/+d//u32UiwJs3b7799tu7uzsjsDwvVL+en5+dcw8PD13XQQpgT76+vu73ewpsSBlhbbKs+tXVldLq8+fPUMK6rjsej2/fvgXmdULRhA34N3APlJrZJXzBcrmkyJZrpBlichdqrBkwiRfUl2wpMAcZgMZIZbybw+yELO9EuymXYUHtWRVURMqyNMaS52qti+KXEYnZk2E4wv/H2pv+SHZtV35nuGPMGRmZWQNZIiVZkgVBhqcWGnb/9QYMGHDD7jbQjfbTE1+TVTnHHHHHc8/xh1/s8+IR/kS9AkEki5mR955hD2uvvfYwpOlFMBIrdklC+y7W92IHIzCFEuYVngazzsrgsXD7wzC0bcfHctlispWkSXyd6KgwQGx8fzVur5X5LeTL1tqPHz9aawn2cXiTyYRTAh6FXQjSN0CEy7XHyQUh7aRpSoLC1aJePUiDHnXFVkaAUdnDA1lpe+QYfPjwAVxIC4BuZRZVuEqV+GY6luMz4NoJoGKQEXvffoVSQk4j14mjEK2MLic64ZeyUzGQjA3/9JThRYZhaJumLAqOqBYVJtw5YAsWgQ1iO3AtRVHQTDQMw3q97oVXzsCcqqo2m03btlXTVG2zPexb1zd9Z9O0db3NUqC/vu/pmuZ0YexAHsj9Y+JIEADHEc0upVTft8MwtG3DJhpjggqr1SrNLgYXv0vK66463hNhkWBq+QJLgfPIsmy9XuP1YcrTC911Hb2T5/N5NputVivSKSC7GLLEtn8WjSSmqqrFYkHWS9qBvix+Gp6FUkqpMJ/PSUpQ9YjI6iDcXM7MZDJJs2w6naIuF/eFCwikRkxDMoqn5PRCGAUyKstyuVymopnBMwagwwIAACAASURBVHddZ7QmRWMsWKTqUYRYLpcsL7qnxpjZbAYParfbUThZLBbOOYR4Y3A5nU5fX1/X6/WXL1/W63VMStgLd+nbSPFnfd+D/JRlScM8+RNOgjM2yBjKu7s7WKq8yG63m8/mQWbhoE1CmzRL/fb2NgzD4+Pj09PTn8E9MLg8iEo2sSTU5tiRpP9UuUFJlxY1NCW9dbEZysvEG2stLxZ9gxU9anvVX66kz+tCmur7GLwTp8eeu+gJMtHkCCI+E4F+rk2QQdMXRqO6cG3jjxjhXwUZNKQksHIiasRZ7GR6jLU2y1JMakRdLh8l9at4EHFFqVBvwU+pHfHKmLy6rolxYh5K9RXIlZUBx7TWAjF/+vQJcC+5GvGGccQgjkajb9++KaHYe5Ek4pjGQBLLTtyKL6epNfoADMRsNpvNZhg4LHWMSfH6HNPYlaaF4094oaThMXqmQaZVk9hh2kgi+ZsYCinJUbBH8M1iSuScy/OCiBJHYq3lzqcyPwvTFkFq4BFObBChQOx4eTXque97ZEi898ShQSY4HQ4HpmtgsPq+b9qGO8Kpw7J0XYehpP+LYoaWhgaSDJ7TOXd7e/vTTz+9vb3RvTGdjvu+O5+P1mrnujS1IVy0YbwMUEpkegrB9Xa7Jf+LhwceB3+JwSIkwpZBx5jNZp8+fQK4eH5+/vTp0/F4xNBEFDSCnOv1msAomguoHKfT6bp0oUXXnWueZZlNEgrVRJyZtE+rqxmC8IhIiE/Ho1Lq7u5uu91ymBmncz6fv3371nUdYk3clNgkkUgTDFvA8WbBY+YaQjhXVdu2o9Ho9fUVVCDKhfG04Ps0aUbJHIjXBFJlWcIH415jLdu2vbm5+f3vf//09HQ+n7HREcslfMGMgGfOZjOU9YzMy7u5ucH307g+Ho9vbm7wu6XoudZ1fXt7i8mN01vZnYeHB8DVeLOINv617oG1iyYyRmrGJGU5VsooZfK8DEE75/t+cG3XVDW3oulak1gXBi/6rJieQYQfsD69qDVgCAaZCcPhI44zxkCCPBwOQLfe++12i42uRSQ9WvYgXXvxY3uZw5PnZVW1yMqFEHxwWZ7kRepcNxqN4EQFGWFE+VRftZJNp1NQEfwQD5lKI0yW5sEFNSjXukk5CS5kNsuTvChKzGUnmp0EicqHtqrrc5UlaaJNapOuadu6IVNhp40xT09Pg8gDxOA6lznvdMNC9AI4UgLd8oc4DqeFM2DSDgI4Mf8L0sKG41dKYcjwfDGc7/ueYgAIDxeM34s99d6T0nkZNofDi/RH+oFJ/6PkIlb+cDh4oazk0oXLioHCB2knjjSPQkY6A4J57+/u7i6gRPBZniutu763STKeTKaz2fL2NslSuNFBdJwA04am6861dv58OLZNM4i2StN3aZblRTGZTheLxX6/HxelUkFpn2ZWm+CGrqpPXd8Yqx6/fUuNdW2nfVCDP2x3i+nMKl2W5adPn56fn0nv0HTruq46H3XwTX3umrbrutF0dqqbth/67sIzAatwQ7PbvxurvFfD4Ktz8/z8qoJJ08KaLM8ufDkvVbdLhOF9nqSH7e58OKrBD12f2UQNflyUVunMJuOiNEHpoExQq5tlkiSz2Yzw6+HhAXtHjv7161d254/u37nUJn3bzaczNfgw+DB43zvfu9P+UKTZbDw5bHd5kg5dnyepsTZoVZRl3TafvvtcN005Hu32+8EPHAAYKwQrbdsOKrSuz0flbrdr6+Z8PA29q85V74esyPth0NY0XedVmN/cmMRaEe97eXmJBI1yWmZFOpmPi1Hu9eD10Puu7qpBuWKc6yQE44txZlKVlUnnLul+WZaHwwH0pm3b+XyelUWSZ0PwaZGnRT6oUIxG5XikrV3e3s4XiwPqLF1vgprP5rFWAehEv7qxdjydVk3z+Py8Oxy0tVXTaGMiJmGMiV0+Dw8P4DSHw4FebrpB6YWEpvjy8hKduh+G426/mM9HZcn33H/84MIFlCZKo3pEkvpncA9KqgiEgTEk5A9O0gu7qWlqpRTYujE2TVMGs9g0wUBjVUkCIvQRhNYZpDEiSjIMw0DuTyrH83jhccbi2K96YqnWkhcPUhg3MueS/+yumhuVUvjtVNTcaDRlNBAoFqtJmE9sCGqJuSyKgrCU0xkp5zEoiHYZc4ml1lr3UDjSFIHD4H3bNH4YuvYCd+JN4yLHMxTL40GokJyAyBTAHRLpWJF+jGXzuLO0BXQyniiySiiN8M2AYNjuTKYaRKwsSJciB4PlxY7H7yF1qOsa7hmeHstuZS4jcBkbRC2H3WG7GUgO+F7X9Xa7JXzje5DDNMYsl0t4L1G6YzqfLZY32pq279I809b0g8M7EjoEIcJZa13fa6WYWRIT9nNVpUnadK1SarfbLZdL9FNVCKRTWnpQ6CDJhJJrZERKEIpE13U8Nv6773vn+ixJjdYcV8Yap2kKx+H9/Z3dJIQchuF43Gtlp5P5MITTsWrbvq6avne5yMaBbNCLUFWV8iERwR+SHsh+5EDOOQKg8+mkRc3Me88wBs48JpK7SbiA2eKcTMbj4D2y2Kfj0RpD6h8vHc2tMD6Px+NoPG66Vhuz3e3SPGu77uHDA/eOUUVjabOIdQvS1p9++mm73W43m+++++5wOGx3u/1h33bd4IckTX3wcY4Tz9aLClme516FtutW93fM1+IfbUzvnDa2KMu7+4f7hw9ZXnz4+AELRusihfHlckl0WBTFw4cPUNjP53Pd1HXT9K5XWs9ms9l0GrPP4+nY9/3Ly8sf/vAHOp9Pp5MPoRyVk+kkSZOu775++7rZbow1SZb2fX93d/f58+f9fn8+n29ubjhUXIq7u7uXl5eyLO/u7uI4rPv7eyYRNTIfnpDr69evQGeRIxpNHGhNXddPT083Nze/zT38Sd/D//F//j/4QMwTtjvP8xBUeiXwQvneWjs4dzqdBu8X8zn04cEP1piu7RKRcYaB60U+mmNKOS6Czlp62ciIowmr6zrLMtYIYbJhGIqijOpgg7C2tQy17qWJl4JMTGUifhJphVrr8XiMb7j+EMCZ4YpxPMhoTHIdZJap00ahuulsdjyfsiLPy0IpNXgPJ4GjhmFSokhopD5/PB7zorDWzBeLyDoniwyigZhl2WKxGI1GcVIYXoFAI4TABcOF8wUOmJ+NF96JsAQFGBaHpTZXHVWslfceQBZNYCc8NGx6rBthhrz3EDmaywyfMRztQdQOlFLZlZYXQAevj2mIJ1AL0RN9wNj5GRfHi8IKsmVg1ngdBjgraTQDw5xOp1+/foUUyyOBvC2Xy+A920E5FOhss9kkWeq9z9KMGZkqhMRar4JNLFAMfJVLTO0v9TNqpHgaUj24sIvFAovZdV1eFOu3V2PMZDp5eX4N6vLjWpvxaHQ6n4qiuL1ddl13PB2cc9aaLMlXt7f0IRqtXe+CvxR1AH+AQBme0zUtfASca8QSGe1A+QRVJR4+Cq9WVYW1whTifSkDVFUVTz7xClu/2+0wx2x9BOvO5zOJSJplSXrhAlCdZsglVf0I6AEpD8OQSHNG33Vt0z4+PuKuuEqwSLWwkzmHz8/PXGEqEF3X2cSANwKL8W2x6kMezytjZ4qiJMKDSpCKWhqF0sVisdlsgsh94oeMNsxyVyFU55NW+tu3b0FKwSDe2+22bZo8zxHpAZ//+vXrpfkuy7HvGMOiKECH8jxfr9ck2ev1moNEKEB8DFPGiqxnBE65ccMwoOfIQnGbEJo1xvzbf/OPv8E9/LpruixLoB6cD8iDMcp7p3Vo29q5zhh1Ph+Px0PXdajg+Mtkq2I0Gk0nF3E3wpbIvbMiTxS9BXZTizxfKj29EFfiI0UYOggNiSAoFj/4Ap5ZVVXolhAax1jeyAgaQn6WHuJK5B1ijPzVNF0tvUWDcH64eNPpFKQYCJ6aWC88fX6WgiHvRR5tZHTSBWsSNSHWBICIdJscJRUte64QjBdyLO4JhkCJADuHKUbHSmrmXE4C3kQklUB+tKhnxzugtcYfJKK8xg96kf9zIogbpEGS5IOAC2cTREsxyMQbHH8rM4vQIMJF4UfZbsrsMdhnBlYIgaoYhng2m5F6E352XYeXOp1OseMB+CjGEFQFsReQr7xMhSIL8VI8K0ejpq55C5CloFQIgU3sRRqd5H00Gn14eMAxlDJDmNwChaUvX77giUEJYAT1fc8r815ATxgRnDF6FWmaMEAm4v5OxGl2uy3vtdvtchH16/s+zTJWT0srohUVyOVyyd/EDeq6P7b9N03z+PhIBwNz1tj6i821Nk1TjBFbzwrUovzqnNvtdufzuSgKoPZYqPPCT+H00hkKnAjmjK3gC6JSbtbnz5/P5/Pvf//72Av9/v5OzhRh27u7u7e3N3IOXFRdtSqY6WTe1F2WFioYo5PEZonNtLJGJyoYa1Kjk651aZIrpSaTCRKKIP6fP3+GlXA6neiZ5QZ1QjHvu24YhlFZFkVR183b29t4MokVeM4e0UySptT5SXCZP/j+/s5FIPza7/cwUF9fXznhX79+hRn89va22Wyqqlqv11+/foX04b1/fHyE5AKqAQXWirYbzgbr50WIiQj7X+sebm5uuDmDkFuguzgZSsfJY712223btn7w2I5hGLy/KAdgbiIMFZuqONyw97RowEVqEHcvF1kO7hJmglvNkSVy52RjOjkxhMx4CHAV2C8xVCf5VUrB3yAhHY1Gu92OX0fHLChBEEG9VNpWY8Gc0PXi2MuyKMtyNOqdW9zcMDzZqwuBB3uHG+ivuPx4IMpiKoQkSYn6E5GkZfoVFpAAnxkj1G/ILp3IbHhp5uTKcQOxyJ0MKoiG2wubwIn0EP47kTESES7gNIMxOmnXsEJwpGgfC31RKPt4PCLbySW/HDJjdrvdIEq9mUxmL2TeA8Y6wtxO9JcwZ7BceF9I6Fg0DD0Xj9Ion0mwTF5IsRTrCTswVvUjR9Y5xwj4vu+nk4kRkTh4qACPWZqRAJEQEyU0TZOkF5YXMR0XIU3T1WrF/mJV8Z0TYV7ybL2oBgCRUZyE5qi1Ho3G5aicTiZ91+VZVuR52zRFnqdJkqVZ0zRoZC2XS8Kd6XTa1HVcQDYRCqlSCl4mZyC6N4p8eD6mRkcjCPwVXS/OgO2G40sdmFyTDSXG8t4/PT2RdpCycNRDCCSyvB1HkWiaTs/D4UCaTj5XVRXgJL8dvibZyXq9Vko1TQMFjt/Lc/JvArWqqiK15JrKBQ7B2aOCRUx2c3MDV43C7+l0+vbtG0V7K+qiSmTG266r6/rDw8PpdBpJSA0xhM6VLM/9MJD6x7iKJyQa4NRBEGdBfv7559VqRTJHZKOlpSyC3h8/fiTJoHePDgmt9el0ms/nVVVRt+fSMQ1QKcWK/WvdQ4SVuLGs0fl8Pp0O3rskMdZqa3WWJdbqvMiMMW4YMCvg+1xybHovDYdYSXwaCX6QQXL+Su47k6Zr4II8z29vb/HDdV2z9HV9YYYEqUtjXyi6KhE61sJAHWTuCo4twkqd9GNHCkdMgXkeJYg85jUVHSqip1yUWGJJBqiKMzQ4p4VqSW3WiVhsK6MsooFOpAEV5i5hONkJD8niHI9HosggzQE4SGCTiEQpqTDDb0tEySoW27EIw5VYW3QSWuSwemmBZkdwySwLOR8pVyqDNHjrSgaXXpORIpMKoqcRshnZTCrSUvikGGeQHgFtYZioQHCiiNaJsLSM8uYAxGpNkAFWgzA7SYYgGjiZiggFq23bQoZsF2VZFmWsUmitq/MZS4r7IZPgxlpr66oGJaNmBq8pYpKdNNzQW/fzzz9Pp1Mq+YB1WjpOiKuWyyVEeyhkZVEQi4DGoA/BQnFhwQDxDSwmOAxf8Ebj8ZiIFQyNv4cgRPwX8TraxzqRDXeiv7LZbDCgdV1TtNtut3SuMU1zsVgQERpjVqvV7373u1QUILjpkGKtTLLDKVpr1+v14+MjDpJTwXFlREFEcamZr1YrKGTDMGy3W8xi1GLQF8FUo5Rxzo9Gk6pqYGJ7r9brLbrDs9mi69zT00vXucPhFNkT+HgSU0AC1vDl5QXf46ST+XQ+N01jjamqKi+K+XzuhmGxWKxWK6Z7/vDDD4vF4i++fMFzExLhw7TMQIQW1coAV4IG6EnAJ7kooSlJXmuZ80i2xIIDLHsZc/L6+rrf7+fzOdxcjkesFPwZ3IO+YqBzJTjr9Gh4qbkPw5ClqHabLMvSNCHYx1EbmWKfXU2lj7aS9L8X+SAOcbzAoMDQxSIUaGSuQ5omxGIRSQ8i02REQDCGhIMI7SohL15D5/ELirfgj7GfIBcFGOKRRCbYwOZGrUEbg6weuntt33kV0jxDLz5GSYM0o3L08cHUS4ui8JJpee8pj0f6Jl6KYDNabUyJuRJA1X+qZhHxBHyGkyGj3Ct+O8Vn/kTQL4JFaZpCY2UHQcYG4Y/yPZno/gfhLwyiokF+ya+jarJcLufz+XClL1LK9NkQAqmVtZZvjuIZQQg5uChqrf6qn4OKd9d1m82GmcmNDK/FNPOfrAZV4lQUb+KNzbIMXWhrLfMVgrooXxF+4hWS9JInKREXwDwF6QAlbH99ff327RvEWbYM+4gb6EQ6YpB58ewaxm4QQcyI/9wuV+OibM6V9uGw3ZVZnmgT3JBImEIVAZirKIrFYkGPGCY7wvSQ94mUlVIge7FSFaTTvhTpXycywKwPWPxoNIIyMB6Pf/7559fX1w8fPjARoW3bz58/E6iCHb08P5NYRG4b3oXzHGNequLPz8+Ej6m0QM1ms8Tav/mbv+E8EHzwauDstcz+JFcGgCLuIV4hNjdCrWZNkBwnuoodS8OVJAwrxjWMuTswQCRxoAcThElMzBeJefP5/OHhASnyNEv5WJCfsiy///57PhzLQzc7ySjFxZubmy9fvhBr3t/f/+Vf/iX5RDxyb29vq9UqttREQ0cIGw0mbYnGmMPhgGDUn8E9xBhTS38ZYaCxpndd7zqlw+CdsTqoi15Fll4C8K7rXN8HH5S+eBoKdNHMmQtJ4wL1aGmXo9atxM3GWNXKZDHqyby2Uhdbn0pbULjqDqtFnwvHQDpWluV0OvVCtvFSJGezMcfYOwxr+qdNtoAbi8WCqG21WmGqRqNRnmWYpLa5kNZVCMh1cHuNqLZF7mYUXudka60n43FR5JvNhsoqtT4lHcLRvRFHYNciABUdZCMzhayImOKTsFCUuzGIsawn63npzovfn8ufWDHinsRFo7LCpoegrjeUX9f3/S+//KJkvGJU9eGWssK8HRGZFlJsdHW0UBljPnz4QBJNt0esS/FI3HACqyRJyOKdjM8jnOf+J6IaQpGwkbGRTlg6b29v8B2cc2mSJkLtTZLEsoxCue5ktAbFcMapEm6zfSEEJgHEpI0Jl9gINwxlWUbB7RDCYrGYzabO9bQQR7QwTVOt1GQ6SbMMoeymbbXRx9MJ6LIoivF4xP3CueLkSImoBJBUsWhYZELg+ACsHkWgIAo64MngFSxUVVUkKGBf0+mUIXcfP37kTCLYxa+7v78n1btU793g+h7Yc3BDTH+NMUDzOHJGF8Sa0+ru7sOHD+Px+NOnT58+fTIy5hOhrYeHh1ya/ynPTiYTdAy7ttfKBK9cP9RVk2eFVmZw3mgbvNpt90Zb1w9tc8HNBunLY/fBxJDVIUIahuH29raViVWbzTpWGvb7vTYXCSInXThsdNd11liIWBwJY8xsNru5uVmtVphHrTU2gRobnz8ajT59+nR/f0/F4rvvvvv48SMtDh8/fmzb9vHx8e///u9ns4saaxRCfn5+/vHHH9u2nU6ny+WSUk2Mxf8M7gFzAGRxvfpZmqtgBheKvBxcKItRluZd73RivVbW2sEN9bnqmtZ1vet6LY1vQQZ/pjLxLbISY8ivJATmoBNIWmuGoQ9hqOszeItz3jnf95dprviGVloKY6DNycMGYaMBmlFNCTLilCAUvgcmtRC9HStKNRGjCDItGTrNMAyc4yzLiOaGvs9sEtxgtTFKDW7wMvGG234pzLp+PJ0qY0yS9IMbQuDfSkbV4yYRC4rpVMzHgQWCdBpHlBnXSPyFielFwo/Mj7EhIOnOOdSluJ9k60a0xBsRqLAik0AEjcth/YxJ0jRXyhqdvr9trU2Nubh/bZQbuiTVzvWDyAqxralMmY2ORykVpWxizyp/SZjPfcA/RbKctZbiJ7DyINOiOmkxoeOUFCFSZjkknBN+6TAM2DuODXIRwzC8vb56F9qqc90wGY+bptE6VG3lw9B3g+t9muQUPLWyaZKfTxXkExYc5TXehQ5HzB/pWtt2VdP1g+/6QRldFEVzPqnBjct8OhtX1anrGnbzdnmX2Ox9sw/WjGYTm6XldDzoYNK0HfpTfVbaGRsm05HSw+G48aGfzcdeq/nNomqa6XwWtCrH48fnp/V2Q1gt9+gyjItdJuoklelEBCKi3qifOlErwAH3fU9dCqN2f38P2HV7e0vUwm3SSuVJapVW3g+969tODf58OtVVFSM2OOKU02L1C2MNQptY63t3M5vroKwxWZEHo3s/aGuXy5VzPs/L3e4wnc67zi0Wy67rJtNx7zpj9eCdG/qub7M8JZXpum6/3zM6qRdJGPzBdrsdpBU3HicQWsCDDx8+8JdVXa+3m6qpR5PJZr8zWRZbr/lwpdTDw0Pf923XwjsgZOHGoXaD9iI8C6EAtD/++KVpKue65XI5Hk8+PHzKs3I2XXx4+DwZz+bzxel0eni42+02m81bUaT390vv+65rttt109Tr9Xqz2Wx3u91hf3d/P3j/yy+/zOfz0/6Q2T+SA3+7e1CiFESMmYvWRZblIajRaByCTtMsSTKtDSEMHowQNUkSL8yWaJv6S4PxpYM6AtMEOIMwgpSITxiRrc6Ly9Qwto2oDdvXyTiBiOwbkfaLHCQCHxKIWHtQ0p+My8UaEvVgXhE5McbAI1JKxXpJrKCCmAHjcMcoxHGp0jT10iyNV484j1JKaZWkyWQ6yfK8HJWQ/7gbxC+xBotXIxEBuKNwB2qhhNM1CEtKKZVl2Xa7BbeJgQPln/V6zRe4gRACikOYBi/t5bGYrK8YhDx5uMh18Nam73om6GmFlQn8uhDC4bBHqj7LMnQC8BNKegJwCaQIILwxdIpJYSpjqIEI6OcoZNDNer2GZsPpiiSuINxlksLJZIJyKg9gZMhSLm19mYiGAS8wy6yuah7gfD5jueq6PhwPYKcsVPTlEZyE7ROrKWmaQhaKcWIMZYy1BINU1GPcivOL2tHcPnqJyvFo8L7tutliPr7MNvdZlmdZPp8vDodT07RN0zrX+xAm08ngfVGW5agsR6NWRCSVtLDgMvEWOEvuKTtIMsFF2Gw2MZ3l8e7u7lAd5whFVSUrU6MBpoCkopJmmqZB5kppbVqRnzGi2NiLGBS7n+f5ZrslR/fejycTOCbn0xlc8Xg4ss7OOSYq73a71WqFMFq0xYm0l2uj+Kdp6+1u07S1D38c7c63ccKrqiIxioxSrTV93bwgtXRFYpdnm+3mdD57GTixXq8JC2ySMI0OzCrmand3d+RMi8UCXY2iKIwxLy/Pp9NRKXU8HrMsG41KUtXT6cxUAi51kqTj8Xi/3+3227ZtlbpIw/2H//AfGaqW5/npfA4hgGLhy91vnffwa0HvGGKnokGIHEIIgdyZC8bxpR2JNcUagjZiWaIRt6LqioXtRUaCb0uloyKIFgUUwCz7Y508gvh4Fz4HcCaipYPMm0xESpZDQ9QQg1MtqkGgB9xMsmknxMHuSpTUCaE2k9Z5UI5OZKZApVgxIw0EWmtqjLilVlT8Ir4cLWAs1V7XD5SMaMUlWBFcomKmJOvCa6IEQMEWGxqr0PEJMQe4YUB/nHrEKzMRJulFBSjWzIergRMhXBA5XFdRFMNVlgOFbr/fR1IveQmLH2ElgLWo4qWuim+xtszfw1VlN3uR4WPQ5vl8JrKLhSuc3yDsFzCfw+GAh8D68F44G2w3SDeQepIkyJeORqOqqpbLJaA87kdJp84geoVsFthdRK5xciGE5XLJMcP8UWJhEay1hJwgQuA5hGVMwmBHuDLsCF6ElovBX+ikIYSohdU0zWg0ZrvZtd1uR7xP6AZxCAVs4K9YJU5kTt/PP/9MaBKLDZ3IcvAjrNLLy8v9/b21FooRq+RluHE0zVztuq6p7iRJstlu+UW9sHtxP4BsqBk+PT3FUA+kt6qq+/v7YfDwr7Iswy3BC6KfH7YCaU0nsmPUJCKIFC9FzDsJGsAqI/J8OBze39+Z/MxVen9/xyglIl0DCodOqpKmhyRJttstZdpB+sMQYaUNkLfmrBYinMw9ur+/h6xFtbzvL9yz4/G42Wy11kZfLgW1wL53PAnVX6XC6+sr8AbYBtMjuMIAvH8G98CVJsQwopyay/hfJUriEbzm+BKTwr4KQmnFM7PB8QheY98cguFK4DpK0ziZpuAuKo8pCCOfHLNCLSOp+YLLX4tSLrH2IFJOnWjBK5kH672HpYeBY9v4fCZGMWiTU8K2pdKczOnnzgAsDCK81YsGkROJiNiygAPTQvAlcMili4oci3Tk/f2d+wPLLaJntegyWRFEghWD90pFilVfTc1z0p/B/kIMw4/SyRUJWvwIlQYsMslZzADatmnb1vtLh9pisZCy3mWsjdY6BFXXDXEf9hewGzuF26jrmtg/kXk4QeZ5RNyMIxfjL/IzErWnpycvSnxK2kesyKKEENbrdSlS3pxkaEV4x1i6x5Uir9L3/WQyIQVhna1Ip1XVmeehNst+AejHxkClLpNlI0eFtT2dTrhtwiZ60THTSia5cgzKsqQ2TjxhhLXM7vAwNNlmaQ6Rv6m7xGaT8ex8qve7Y9+7qmqSJFPKLBbLzWY3Gk1os4UDRqSMC8T9ZELazrLs8fFxOp0+Pj5676fTKYB7rOQBx+12u7u7OxY2lQlaVVVx8Pq+h14FzxinGiIRdwAAIABJREFUgsAf+zUejdq24Wocj0eIkWzuZrNB18iLIufLy0td15vNxvW9Ugrg7vPnz3zPTz/9RE8AD0PQwG6S28Eq1lpvt1vnOnq2nOuyLFHKG6Mmk1EnCgJ8phfVeufcZrN5eXlhjgU+43Q65aIETNp9e3u7XC5JmCIsAZ2XnlmyQyPS8RjJzWYDsQInATON/IBi8uFwcH3ftq33QWu9Xr+fz2djLySX29tbpS7jfp1z4/FkuVzO5wvAkqZt1+s18hvUooD7fpt7+JOu6f/rP/4XqgJaGmrYY4ACUjkIhaTY1lpE5TjNuFASgkFYNNHmRvfAp0VsBMM9CEEWtMEY3TTNbrdvmmY2mxdFMZ3OICzTuRrRZMhFsQ0bQw+Hj0JfjJ13ux015/l8fjgc8N5vb29Qjyk+U+430rOWyIjHEAJ5ktY6drobabWDWBkZLE3T0OrZC/cUy4jNwtZzIKJRZi+xwka0pwjioqQHv8WKPjMmD1CFyTAxp4mYPqsNtr5arYjE40a0bYs8JOEeix8h6U4koLXwgKGLOEenulJK5Xne9xdF6PP5ZKw5Ho9lmZfFiIiYUB05ClwmJw0sPp6KWI2IcTrLGEFbdpY95chxP6fTKb6Bfh3k0AnTkM/Lsow5XHwO0QNZlFE6dnV8/PgR0TellNHW+2E+nzNlgWOjjdbGxprqIFrWBBkAnuDOsfCO+4FVoUSyniycf0dqX6yU9NL6w//CvJZlGZW027YF2cCjeO/RpDtfxitdZlgR53rvl8vlZrOx5kLGJ87jduBH4+fQlmytpS+XJYoFpCzLNpvN3d0dwXgmA44icpsKrdYYg2nm7mw2Gz6NtXLOeZlgiJI2JzBJku1uB4I0n84GAEDn+r5/f38nNMzz3CZJmqV5nqugqvMla3l9ff2Hf/gHPDfRACvQ9z2ivI+Pj6vVbS96CtHcRSFkbN1ut4u2hfeKyIExBtk+cl8utZeO9NT+cQQhv/cSAzV1XhRKKX7KCB/dKN0L3Z89enl5SdOkbRvnhvO5cs6V5Wiz3pRlEacPGGvatknTBKbM8bifTqf73Z5IbzKZctGSJFFau94tl8u+67z3qU2MMf/uf/kf/7Xu4d//3/+Z1wiirY2x4A/hObVKLUOGo94hpxxzSS9rCGG1WgUhz3C3g2g3cqmYXNFc6W5iVqhz9r07Ho+LxU1RFGU5wut+99133FgScHU1L9rLpMlM5O0uzKK2NZeRjTkhQC2DiLEmThpBN5sN6CdoiRZpezSOsFagak5mwwEFECM8PDx0Xdf1vRfbZ0R1i2wGAngho1XJH1nYmAkO0omapmnkSxDxUVUDrEhl0Aevr68m+kUtzAiGEO1Ck+AoA4OwUOR8SgY9gUGT58aeo+4y8zVv204ppYI6n6vJZELKu1jMt9utNgpFsKZpkU8gNMby8gezDq8Xz0p0Q5SKAB8BKasd7cswDHwN5gC0CioVY5dBeP3GmOVymV11POCSm6ZZLBbb7Xa5XB73B16WCge7WRRFCJoFb9vGGHs+n4qiOFdnYxIIvhEAJJqDs0tt08vIW+hDnDTKDL2I/kYJZO4zTK1EemKiilS0zhCQeE13UZP1ZTkqinIY/H5/6Lp+MpmeTufxaBJ8UEElSerckKXZ8XhaLpen4xH7jtPyVzKRvNGF0X86xXnUBMKpCI7x/YPQ2+LRjWbBiui6F4b0MAx0b1zaxHjf49HLSJKILymlqqo6HI8PDw9pmt7f3VXnKs/zvuvSNIWqNJ1OQ/A2Sdzg8jxPbEKeR45YVdWXL1+iJyZfJNwhcZnNJkWZ103Vuy7LUpuYPM/mi9nxeIrgZIxTaY7hXmut39/fUT4ncb+9vQV43G63/Ioiy0ELD4cDLgrr75xL5RoWMivQGKNCQFgTIJTY5XQ6rla3w+D3+10ISmt92B9CoDJahRBe316N0bMZA+FVhwRhUfa9u729nUymMDhMYpVS1hhAvMPhUGT5MAz/7n/9Le7h/2eKEHBQEHkJL4OZ1FW3LTc/MjUj4B6EChkPIiBML8wqeyWkgUuAORAN0HDRKFXBB+jDkUYVix9eKP+YCczoINql1DlBWuq6ZhKTFwGoTuZAbLfbKKdOgZG6t5IygxJBUydk/yg2EBckVtVwAE60m6LFJ2mFZIlRjqBcI7MD9RUPlXtFhNKJWAWfzAcS7lmpijuZbobdZLWtDPnhi/F4fHd3x82BixKtgxdB70bEOeLCYr6Px2Pknh6Px2Fw1loVLqWpGDeE4K+qFKrtukJ01na7HdQ9RCm4jVqaGfFSMAX4G77GVjqZ9oGNoJCA5+PsRVgsBnROmrkwTOSFWkpfBIa8OzQntg9pHY4o++JEE4JwHtyJBeTkW9HzAZUij8FLgVhGH8zug0go4Vv3MvMDE1YUBT1ZJBZpmkb8MBUFgbIsqflTNiAnZo9iBEYlP7Zbeu+1NnVVY/Qpp+N9i6Loug60Op6Eb9++UU4g+SDTwooxsriVOUWxzg+Ag4nn4kBk4sdJCHBvPD97sdvtsBJPT0/L5ZLmPu/9ly9f4hHlSUga6ro+HU+DH/hxRLA5CWVZPj09rdfrT58+4fuN9F0ppXAtp/PpYu+M6WXcPVkaVpWgrRdSFmPv+r6P2CZjhYwxaCCeTqfPnz9fWiPrmnenW4iUiB8H7eQkZ6LLQNLGnDFQLIIAPpMpTMRA7+9vfd+PRuMkSVzvNptN23a0ufzw4w9t265Wq6LIMXEPDw+LxaLIC6Z0MDHi7u4O2/IbfMOv3UM08STLvXR7c/PJWxOZFx8tPgsRISkSWOr75MjcEyry3FvgOS9tqG3b7vd7Jni8vb0dDofzue76LhaiwRaQsmDFY6pIqMWikyUQmkUjorWOEgtexjHi9rnkADV0EWNKMNNsKh4IUMgJTT6GJxExiAb9ErMHlaVpnuX4PyuN0Jw20LDr5in4GDgGKyMNelG744pGGzSIggiLw6Xl2uPqIqBP1KyltZvkj9JFPMpeJjiFK4H0LMsArDBwBAFOxtoAsmWXdnFljKHYYC2knbwsSgJn7JeSSRXsUSVzamMW38pAVhKLyDym6YG0AO2T6F+ttcixeRG6iZWD6Fo6mRgDmhEjBtJfLX9TlOX+cNjt99Qn3OAG79/X6zTLDofDEPy5Ot/cLMFbojQhTkJJa1Uu4lS48EH+JDJPFynQyWSSJqk1tmmavus4IU6GTsdUDzgi/idIDkcRJ1RXTd85KPxN3Q7OZ2leVU1Zjo/H8/lc73aHYQi0EH/+7jtSsf1+r7QuimK92UR8jMX33tM6AL4KKth1XV4U09lss9nENnXcSaQnHQ4HIO8gYQKV51S47NwyMuayLMfjy9AXrfVqteKQf/r06e7ujhYBfAMJHMvSgKmGcDwejbFVdRHFIpalFgKuuFqtrLHKB9f1Td0EH0blqO86a5RWIUns4HqlgnNd1zWxMIljILsyxiC+NJ/PyfkQYhmPx6DKFNKYCDSfz8m89/v98/MzaDAnqyiKcjQKV1PNvTR15XmutL67v0+zLM0ymyRd31ubaG2Ipdq2nYzHoC/H4yFNE631fDHnEg3DUNfNqBwVRdE03Wp1Z4z5+eeft9stQrbW2O+++47tZvLS+M8yazqmiuqKxcQ2W6GOsFggGJ1I6eI5kyQhizTSaxarjqnMCGNTiZ1ZslZmsBAkvr+/Yya2m13fD2maa23btt+sd0U+SpO86zr2iYjSiL5YLOV10rikrrrzIj9KSc6ey+iCWDLxQr3IRFiYG87ixvo5dA74QgTaxJvYVmttoo3xIdEmtdZq07ddLDDE8DyWCqCjgJNCdYhGgaoJYDHMCkqLFwNR1yxjzACMMWTE2OUI0McchbyK1Y7q3zHnoIAZu+1QzgD5IQ7NRNUOn5qmSVWd+r5t29r7wXuvgk6TPM/K5XLJZ+IbCMeIOWqZNRZEGxyAjjoW9QNMBlf3/f0dJVErGpGkdHhW4lAcM4FCxDMpxhKRsHRcfiea8EmWKmu8VtoYZbQPwWZp1TQm1YNyLvR12xzOp6CMV2bw6v7+Hqd4e3urlMLWU5GKjFviJwJhIgDC53jy8zQrskx5nxp7Pp2D913dJDYB107TlNe/GBGZ7xTzxQinaKNO52NQ3lh9Oh8Px/3gHVODtA7W6uNxb4xqmqquzyY1Jk2C1vvT8eXt9dzUp+r87fmJ9IgwE/xkNBo9PT3xzOPJ5HQ+n+tqcjOv+nZ/PF70wM9njGlknbrg277r/TCokBV5OR4tljdaWuI5lupqqjmvw5X8/Pnzy8vLfD7vm3ZclKf9YTweOz+Mp5NzUxfjkU3T0XRyPJ/Hk2mRlsqFzORN1aSpReNnGPrFYvb09E3r4Pvuhy/fzabjssiM8lli5tPx/eo2T8r61I7yse9VmY0SlXa12633XDFCE9J0GL10ayuZLY+JyLKMGeBt2263W1wpdqyt6uNuz9wL13a+dyaotmkIAckg6flP07R1/XQ+s2myur/LijzNM5smvfd970PQxiSLxSLLs/GkTFJjE22sulnOP3x4WC6Xw+BHo0maFr1T33358Vw3g1eLxfL29vbxl68Dm+D96vZ2Np1yAX/4q79c3C7/DO6BP6wFKUwioj1KSOtO1DejA5hOp6gSkqxp0TgiPMxEQBTblF8paihBqzj6MZ+APgx/mesXjf5oNMJpKWnRiGhjpA9ZEbwtyxJsikOZCKUaU65kaGh2paOphFEKA6EVhddIL4GKENvHBhmJzNIFaRFXIhp8nT8BlyfCWI3JRESHnPSyGRloBa5KUZGUHw8R9f5AErw0DQwyfImFhaUTa2VU6riusUKupOMEGGcQsR3iJpwN/cYs6Ww2swkIoYlGPPLZIuTFZimR9/nV/+37frvdsshALjFDhb8bL1UrPOlY18VZKqVIwCElRwoKsQhQXtu2sFoTabnyIti13W4JKU6nk9IqSZK8yL33aXaZs2atRW8cdIIUimIPLV3YOzAQ5xwETc4nOSXLiEMFIJpOp0pIok3TdG3biy4hWq18M8kujhPcj4Dml19+Qc2NHLqWuVi5TIWq6vNmu17d3fau00Zttuv15r13EsPlWQQJkXCgIu1l7hPdalSkAG2yLEOM+bvvv39/f48YF4EdUQhii0opRMOcc0Gp1d1dISNmaRKGCDs4N58vwpVCzMePH19eXpqmofxDHvDx40cueJ7nNrGJtDjkRcGpC1IwI77B7Gy3WygnURuKQCfPc3ilu+0uTdNROfI+pEna1DWN1lFuhJ0iBy2KInZQDsKNHI/HcJHzPKdOqa4INfSsBJlNubq748zHYVCPj48fPnygct7K8PkkSZbL5dA7jlwIgZ4nJWMbzuczEPqXL1+4sGmadl1flqP7u3ulFBdtv9/D4yfSvb29Xa1WRIEENL/hz5+Upv/Tf/kDBoUCWpCKayIyNfZKYyeIFJKTAdlcUYwp4SpntxHRMQ4EHHMAWSdq29xbkkqxtmEymRpjmqadTqdKae9DklibmNjmE4GUCC7xybEszOI65+Ku8FSR2sg5aJoGM92KFAyjWthF7z2BEmYI7MVfce3TKJngXN/3GnmSJEmSxCFzFDxRBgxdfCFryzpHCD5CXgiw8G2I/RFU8iS5DNGLJCh+MKLq3nvKnkDbvCZ8IdytFgE+pdThcGDxI2qMpyH6wNADr81mcyyv9yFNM2ON1hcFOieq+hx3FjnLMkBFHoz7FpM8aqRseuxlw4gvFgv2MdLD8XC4Z6wzHi7K7wCzUmfqhbtirWX8AGGNlQmj7orIgCPhZVleDMHhcBiPxwgvKxksCjxNrhZLmnyBF4dKtFgsmHVYyUR7DNZht6ffitfhxA7ep1n29PTUikIcNYZWxrNTgoo5pVKqadrj8ciLcyBHMql3Npu9vb0tl0uwmslkcj6fIhxKSzmvOUjHD653sVgQf0SnTi9SUZaTyeTnP/zXQhi9WMamaR4eHvb7/eF4vL+/DyGs12smz8xms/PxRFDIIhRFsVqtHh8fjbXa6FjcBuu31s5ms/f397Is4VABs9CXB2icZTmVvL7vVfBd31L1iUiGc24xn7++vjL2ALdN9QJXEYvhTP5RWg/eb7fb29tbunZ6URHNRN6VxILbYWUmK70p1loqYd4NVV1neW7T5FxVRVkCGcFGxeWwEUC7OqjEJq7vVQiz6aw6n7/7/N1ut0uTBG8XwbdhGGj/ZGAR15M5H8fjcfDeGF3X9RhOmtJVVfXS+Uh7RCHUqR++v/sN7uHXw0SxC0T6SiL0iIEQyBA+Ewkm0qWVSLMoJVaMHf/mTJOy4UjAgrntiSgusEMxoifkVyLdo7U2RseicXQ/1LQXiwVEcqxeKsprPDPuit+OTSGliFV3orbkSkGTqJnXvK408O4MH+9ltAiwRqRjGaONMUoUZAFtCDC9TFpWV8TZRGSxeR1iUvhOkYoaZDgrS2GFs0SGFx0kPq8XQWMnfaE4JMAiONf4lVTaOOj19yKgG+9VlOIAW+eNrLGR4xR9jLtq3XByRo1otnM3gP4xauz1MAzwQXPREeHvAbjd1ZgmmttxVBEoi49ExhnTNdwwr/nDDz/gko0xOKEI0/HMoEOYxTRNkURmAABgTnTeFEt5BUJCXs2LmgtOkUPIp728vLACtKfgZXEMeZ5vt1vadFPpxg8hIIqOU2dnUTblxengxWPt9/tPnz798MMPcFWzLKmqUwjDeFwmiRmPy7/8yx/m82kvLGEuwvPzM/Un5gTAIoO4AXIY25imk4kfBohYHz5+GIYBbVomKmO8mJZBrs+hIrwAeeP43d3dgfUN3q/Xa1gGRIpPT0+FDIv8p3/6J+DNb9++Afsk0nF5Pp/f398R5CiKIqiA1yf7HIbhw4cPoNyRJkcO0UubFKiRE2YNPmk6mZZl2fc9eDXmKBYRMQuEj+i9a/nDPGc85bmqnHNkUZwiMFVtL7h3FBbkkGBYyrKkOyFWtp1zy+UyF21QUluSxcfHRzqiKbTQjeGcS5O0qqqXl5eiKL777rvYYYMl4TSu12s00H7Dnz/JHv7z//tfYyMlxTH8gbkaRcJNYNsodXL67+/vEWaxojLkhaEEe5VYm/XCsffSSQ9cwK0jdDLGWGvKchRCaFsKoTkf23YN4Xy0g9gO59zT01Mv40SoW0IYx1ukIiHFiUFJNBGdZIwsBIaiKCiQBvnDCYZdQ4KPY2BlopUnSQo+XPRhQqD+qbRm3YiVBlEdZ6lpoI0BONkDgerNzU0tksv4M4RMMpEIVVJKAevj1bTWsXaSZRlgRfQfw1XnNpB9zAW5ACEEGCa0H9NX0Ym8qFLaWKuVVkqniR28w2YBXidJQtaMt359fY3pJieqruuiKJ6fn9kIbKUWGhg/SKAXROt0uVxyRYkWnYyRoZ0C/0TsT6geozxcBY0CHAbOAJ6GgkGUa8SpaOnPj7QFgl/qNHiFSAYNMpqC56nr+ubmBuZ0zCxjEf6CvAfFh0fPR0VxsVzu9/vX11eWMcgwPv7Te08tVH52gF5MVYZqWezjo0VgsVhgqjhXnNvrapkxJk+zWJoqiuL9/f3u7g7THGlmp+o8mU7runZdnybJeDzmskR8L8tzY00uMiRYiZubGy+YDKedYm/bdU174bWTo5P3kDMxeyfiKvwUtkhrg7TP7fLWuaGuqzRNJpMJaBLRjDHGaL1YLCifgJIB3FHIIWBdLpcXCtnpVI5GTOxAnDwGMSSyPAk+9f39navHATMiUkCgmWU59p1wrbtMEvQR/6QFfTQavby8TMeT4P3gHCn+bDoty/L29vbt7Q3shJI414riGYNmYanM5/PHx0drbZKmrndKKVq1P9w/7Ha71d0d8TRC7gCe6/X6f/4f/v43uIc/yR5IZmP6H/FxLI6TGSNGBI44QNba6XQKfYjPieA1yRR+1UmLlpOh8OCSQLoEyyy3utJMjrC7lwESBN2FzGkgMFcy8zneJeccXQJBdKFj3E2wDP3cigAfPdj4G055EIY4pKNI2qM8FaNmLifBjpeJC/H3Et6WMioOy+K9hySTiJIoh4AviA5wRTFmJ2bpRQ3bSdcucT2f40W4n33R0t8QRIqK+BTDF8sM7AvrlskADCxRxP0Iz621w+AIIFif6G7j/mIQCat5TeJNCjm4B6QDY7meODoGPmSZWFuIlYxo53m6K50JJyM6MP1WesVhvPQyPgG/AkTjnKOmiju01lKPiUwbah7EFkSU8/mccjFCnsaYt7e3PM95L7wmOiuz2YwaLx3Ff/VXf3UW9RsO2OvrK1vDdvAF2HTfdUgwZVnGYyulvPcUQrl9SsRFqMAbmUmeJMnt7e3t7W2WpeNJaRPdtNW5OhZlVtWn0bjAVXNiy7L89OnThw8fptMp17YsSzROCIBQ3Xh7exPPbZgEF88q8UTktpFl8l5obltr/+Vf/gVY34qWNbHLh4eHtm1pbGILCGLG4zFoGCg528oF5AoweHm32z09PWVZ1rUdFrCSmbX8CNAZT6WUok2MG0EMShx5SVWt7bo2Ewk1LhT4VQRFAP2xObvdDh8QkUnYTZPxBDPFgScLf319TZIE1hC2kRdPkxQyMZRxvp+4djabwTyu63q5XFprP3z4kApz8unpCdkMY8x8PseWHk8nwuuyLB8fHxeLxWKx+OGHH2IKGDniv8E3qF9lD//b//7v2RWtNW3xSvRNY00vBq3kqtwQsi3iXG6dEv8PwsPhY3EJUljl+Xwe68lc7whPG6NDUEmSbLc7Y0yaZLQmJMkliidJ5N/Yvre3N+ccOfhoNGIPvIxAiL1sFJ1iLSFJEmjvjOPAUMZRJFTde+kS6EUwg1AxFZosdYu6rkMIfddrrUeTsbVWoX6qL/pIJGT4TnwPjidWCHDSBD7Ya/4+rnMQ+W6uQbiaahdhaEJg7IhSKhJ5IR1gULi0YAJQs/gcIkcMHOugZE6U1rrr+iRJ+g4VP2uMLkcoewdiNOxsJPiiiYtdw++SJnYyJ2c2m6HiwNdKKepydV0Dfdze3i4WiwyO6TDgJ/DfgBhsK9nhx48fD4fDly9fnp+fvfd3d3cYYiN0YfwEvYEkUkbGT6GzxLJjhcfjMb4EptMwDFhwZg7GshyFukgQuLm5GY1Gv/zyC5PaYAFg+47How4qIsKpjHYYvD+dz8ya53as12sWkECSDix36Ymjd9Jvt9vZbAbWVFz0fM5ucHd3KxYEILeqqtn0wr9smga2Pv4gTzNySqIlY8zXr1//+q//mkpJjcRhmiZpkqbpbrPRSk2nUzw38cqleiSDVIF5lVLv7++3N0uMNUeC8sBms0nSjJuitV6tVuv1moINb/3ly5fNZkMyR2WIOG88nhhjgaTKskyzxLmexIuS/gUkMJdJGFgSYsE8z/lBrrCXrsCqrqu6CjJKgLXi4sTuUdbHGEOlJwiZPhFFqTRN0yzTxqRZVpRF23VFWWZ53rSt954WBBLx+/v73/3ud8aY1Ji2bqwxTVUf9vv71V0Y/Pl4Wt3dweQmErq/v6f0MplMKGWBgn79+vXv/u7vqrp2w1BXFc7seDwWWZ7nubEW47bb7WgoYQjg//Tf/7e/wT38uvZAoEf017ZtEL4Hd8bJvBcvFCN31QjG19ElxCJ+BPStsNk4TNh0YN8oHcGuYKqca51zeZ4OQz/4TpsLao91VgJY84NEQzGw9ZcO2D8+Wy+NGtE0Y5RR3CW2iuFtURSLxQLqBfUDwkyMGmcoEoF4Zi0twaSuGGte/Hw8VaeLaltMCMhDef5CholitblpEHi0VjQKYNo2mw1rxW8fruZbuCsCT6yjahkvDBk/F3FcrHZd16+vr9HD8czRhmIIQJxRDoyGOE1TaxE8V1obgkoo2845+oz2+33X1lr5ujo50TvZbLZPT09pmgy+n05H/81f/7i6XWAvEtHgIkVgrwFqCat5l7ZtldJkM4fDkevadX0I4enpabi0VWqllJcJXFytqMLLxB4O+SAjV3vhv+PqTqcTn1ZVFVC7tfbx8TGadUKTruu22y2nGrYF/4uLEIKfTqd1Uxtzqfz30iFk5I9zzmi92+2YjYjC88PDAwdv/b6G9LFaragn8YR0J3Al397eYFIYk55Op/f3tVKKCSJpmiild7sdMDSib2VeLBc3OqigzHxxM55MmGCDN3XO7fb7frgMsHt7fWXj7u7u9/t97xxyHHXTGGun06k2hkNLPdkonSVpmReH4xFU3YeASUnS1CaJtXZUltZYrVR1Pg/DUJ3PwKHe+3/+539myBoqn1oYKG3bFkUxm03d4Bjk4b0/V2euDLmdI3bpe3wYx54g6ebmhpZm7Nh6vQZctUmy2+2I4jvp9Cpksj3BLkeRRrkgiiDqCkThui2Xyx9//PHz588AKtg0XgGUmHNy2O/ni0XXdWRXVOOOx+PLy8vLywuxkfd+s9k8PT3d3NzQNIcEFoma1vrr169t22qZEUuczUSQYRgOh0OaJre3t58/f2ZJ0/Q3tsX9KXPpP/2L8qEf3Gw2G4L3wZvUtn2X2iS6By/NVkopEjHaKanaUckheexlaMZms+EmkK1vt1sKL/2VYijLzekEjgxhaLtWG1VVp/P5eDodijI7HvfIfDjRXISlQziGYYpUWiMjd6IrGq7k+cglI8xNoWm9ftNaZVlaFLm1xrneGI0FTmXUOGeCKkUmE0ZJC/BDSmulL8IMp+PROWeNydJ0GC5Tq4hSldB2gSyBRzBDh8Ohd62xOOZhcN5au98f0zTrur7rLpSeRuRJUumRJsZX0h1KgIzt49k4jjGaLoricDgQgBPfQTUhJIl1477vs6wI4ULlVFpleVY3tbVJ8KooyuCDGwYjCoDeK+X90LcmOOWHvu/bplfaGG/O5/PgBxecH7pJmX6+X27fX3bHytrk48ePoJGEP1VVoWcAWNz3/fPzc1EUfe+U0lqbNM2apsnzwrmhKIq+76jlRFIc76KupACW78itAAAgAElEQVR7UX3g1HGeuWCpSJiwJt57kHHi3OgSuCkojqRpulgsADzhOAHZN01TN+fxuChHRVC+KLK8yOqm0jqEQVFfwa4RV+Z5Pp5MDocDofd2uyU0KcvSBJUmydvb22g0Vlp5raqmzvOCyJrsln1vmiZLi+B1lhZN02VZ0dSN64emabFEQ+9UCK7rF/N5czqXRTlom+RZmhchhNXqrutaa0xV11mRV1WV5VmC0F7XpTYpysKr0Ds3Go/SLO36Dk2hvMiTBB6OWi5utpvND99/KbLseD6ZNKmaundOaZ2kaVGWh+PxuN/7YcjS1BqrgppPZy/Pz6NyVDe1F7Is15aInj1q22Y2my6XN33f9n3bdA2f3FT1bDI1WodheHt7U8I2jmFcKn8A/SjJEKoWZWmtHfwFqyREY3/BpmATEDJyg+i8SdOUnJIrE/FDiiWcN+9903dBqw8fP769vhZZroIa+v50PCqtZ4t571ySpU3XzheLrMi9CpCwYYvAfwHA4JGAH5qmAXE5n89Zmh0Ph+PhwJMkaepDMEZNJuO6rtq2ubtbNU39/v5mjP63/+a/+w3u4dfT4mJI4qWwrITCH2QumzEmNlh50X2LZSiyQlaKT7vGo7iuMCnxkyASkO2MMLq6rjPGWpOooI221iTBq7bp+t4husBdjZ8AgTIW/YhzeRISRjJNLX3dqXDtY54Y2ywwAQQaFEsx4vSLsfExtWRZglCGsGI0uNG1V8lsbaWUlQ4+jhQrid/id2lh6CultNKDtBQ4adnjmHKAIuJ/XdSJICwAEStApgyAzodjXMBSY7G3Ew0+vGy8VxRXrUhhYxNZgcPxGGWgjKgskGLCCf7bv/1bCHkhhKqqjkfpDE+SpmmtTUJQHz9+TpIUtHC1WoGZFkXx6dMnbiz4DHj3ZDK5u1ux2lQCIoWm7y8k6RDC/f09sYuSkj6hAN1JbSsWU9KpJEmAfdEQZdnjReB8cozf3t7o3zyfz3mez2az29tbQABQTS7RaDTa74+nU+XcMJstRqPJ7e1KaxOzOiOqRKmY4NfX16qq5rM5qUOapMf9AZc2Go1c39dNzZFo2gbWAGgtJzDLMkSeWaUkSZg8jBJJjNVub2+pvVnRfo+YO7kLa0Il5v39fRAd7Fa0DrGSt7e3sWI3DENik2EY3DBcFi0ENj2EsF6voQzwN5TWQgi04/G/urZlVAPdMHgIthIWH9gXxUittesvNOWmvQz+w+5r0UUGmFXC0qaAz68gjRuGwWg9mUzu7+9JLmlmokimBDzHwRRFgUeH5gfynGUZhHgOjL1qjOUi39zc7Ha79Xr9j//4j52MaSmKgnLpcrm8u7vLsmy9Xg/Crt5sNj/99BPVaaUUJZ/1ev3t2zeQMZC3aC4ACYMoBVCHj5jw73//+/v7+yRJqGz/a90DoTFrxyU3oulN/nh7e0vYRT6eXWQVdGRJYk/58XDFqxmuZltifEGQtNBGrQwYIOzlJndddz5XTdMYY70Pp9Mpyy5CkoPoT2BBBhkrjS2IoRkFTyOtAOQlWD3sEQeUYBOiCM6AS4LoE+eMqAS/yCfEaiEWBHCcX2Sl+SvIEAWeuZNJL2SRrNLxePz27Rt+0UpzrFIqSdJWhDdwEq53TdMA6VrpJYyFBw5ZZN3E28LTApfhSjlAfIO1llnHMFsI34B04nI5NwRpxuaI4zWpfXVdN/jLCARWG20SeEG5DHkeREpdKWW0GY3Gq9VDlhUfPnwcy4B7TBv+AJfAwgKdQx1JkosGkfeeKjSphhWFc/4wpq2VnnzsLNgpWwPhmB9srv5wsPEBxI/8GwdA8p5Ld+cgLDJE1sB/bm5uVFBvb2+n0zlJ0sQmSqn9bg/Ef39/n8pM4E7a4oIP8+ls874O3hdZ7rq+a1s8R1XXRCfj8YSjpdVlGEY8pfqP81QCcAQ3NMqIsRdKKezvYrEoRyMlwxApMiulaKQCT6OQVtc1RNg0TW9ubqy1dOSFS0fFpVaHRosfBlhhxmiadair0a5FRA9rsZXBSn3fZ1n27fGR0zWZTECi6rrebDbr9Zq94wDH451mKRnzZDJFDJHGOv4vmT2mKXLruSDoYVwgVuesNavV6v7+ngACwBla1yDtJpH1jkCFEnUJLAwxGXsBSnHd9bVcLne73X6/j2YKB1PLAKibmxsKJ9gZLNXj46MxBnlEnuqXX37hLykzcDgZW3Q4HD58+MCVJIh5e3vz3kPYPR6P33//ff9nGSYau6K8KPVziCMWxLdh6ZLLILmMgD3euhhpFjJGlDXloh6PR0oR3CusXiai85y/yWTSdZ33KgTt3ECDc5pmzl2YORhEJxMpIgIogzL66MxjyT6XSTiJkJp5QewOpicGdBHmjr4tSKf3IMw2mrka0fzh6PurFr9YpImpWLjSsCOa43jF6kgMV/M8L0flaFRyBNM0bduubduu77j2EE6UUtxPzj1fY7m01jc3NziDX8X10Gz4dZQWCBVjbZyLgbeLhZaYVtZ1HecjKintBB8idEtioY3BshO4oRoUQkgSQ7VtNpsVedG1w3Qyz6TrJRKTIj3GWltVFWYiF83zxWKBel0umhNBxCLf3t4iJkCkgnWIIHVVVbQoc+TwhUTodLQB/ZHGcTxwgewFdEysW2zcw5jipXiL5+fX6tz0ndPKnE6VVjZJsurcRLcai2SsmzFmOplopWrRbAdbCCFAISe7jV0dMZDHTUYcPEksd4qLxo7T59y27WKx4Cg654o8vxVnwLLgVGhyBt8jMjXGvL+/M18h3qNMxmaw7FVdExYAzc1nc+99LA4rpWBtpmmKA4YyQFcgY5dGZfnp0ydK8TG4BDSLx5KSGIJ3RVF8/Pgxz/P5Yk5ho5fmXKgEZPOEKQQBqAAcj0eOWQjB2oQe8sViwQXJRIWFKJlnxjhwROONYNcIdzgh5GpaONw4EgZCUMmAl5UkCSd5u91CgyQag03Oh8epc+TlP/744263g0HDvaCer5SCB0HPCo9ayUCU3W73yy+/PD09/Tbf8Gv3EOEX7pIVzV4jo+eJzcPV9GCghhj1Q8sJQnVnxbUwMpVSkKb5EXI9SKgEEVRrYaowrxRp4jTNMKxKusnIjqOGK3cYIjzHKJVWOzYV8DFerWEY4FNdt/IZUZuAbI72Vl3XWNggOsZcxdVqlYnyD24j3igtna6tiNDF+vN1KhNb9tWVGCK/Issy7yVjcC6EUBQ5Nh3PSiakZeQn/2llQDQUA5oGcMaRRhlrJ1qk2iOhlhAe05Cm6W63I+hTSikVoJ/i/GikiHyepml654Zh6PsuAhHz+Rx4Bw4oLxvUkKYpNPa2dcdjVVXt4+NrluWgB5SyYpDIKA7WkLe4ubmx1njvp9MpJt7IPKWqOjvn2IVcpLRAkyqZOOtkrtF1OMnZi8kN64keeClCjdgX8lEt8nlYImstJCWKnNzb2Wx2f//gfWjbbjqdFUU5n89PpzPfQ7bHH8BVrRQw3X63q87nNEmq89kPw2azcc4xdqZpmru7O6RZ6BVo2zYRoXI8qBXhIOIkTAwGhSPKG3lpWiaJhLyntSaEL8vy/f2dIZ3ROqdp+vz8nMjUDaoC2L6iKPquq0RPexiG9/Uaf8nvJSKJguGr1UopRV/629ub1nq5XDLvoSxLmDZ8FOp4sUIJl0wptVqtcAA4M+ccpezo5okSNptNKYrfkFwxU0iqHA6Hqq6wVGw9iQufw1EkAZ3P5zCMOxF5xM9tt9tEGpUwodBVkyRhhKcRdRzQHq01LiqRkYVo1Hth+pB83Nzc4PVZfCKzh4cHoL8gRUT6MVnP77//HoIWdFtgKJ7tN/uGX7uHRAY8ZFcTtYYrCQ2eLL621hq+EMgyxTpCTi8t6V6mhBJbgWDm0ih/DVJFdJuTNwze2tT74H04nyvnfNv2SplU+GQxd+lF3wY0ICbUpEHR58VMjZ2DzR1xqkKmCXLcwTGAmDtRoUEgupfm2KgoaURfgXMwiChTIQOfQwjaGCJZL6LTVNeDdNipK1Egol2iQqZ/sPLMTYzJB0aQ+yMokKPQEjFTJfVYczVNGsPHLkeAAr42gSpZQjwV3v9x370QB621icSPRuskSYyxOA9rbZok8LhZpWhq/WXAVBic/+ff/Ut1bo6HMxUn3B4lE2woMo6Ii7Bxt7e3HB7Y+uT11zQkzsPj4yMnrZHhbuxjhGJQPWL8Cw4YxgSOk93hcBoZd8ORo1qA00plXBoxgZIB4Nba5XKltbm9XXVd33X96XQehpCmWZ7nUMVIYhoRc2yb5ng4Ds7td/v1+9oPQ5okf/jpD5xqY8xoPOr7y/NPphNOPp4vkwbj3W7b95cae5Zl1Lc5k8vlknwdYu7xeCRiZU9pBOF6UoHw3uOKPn78yF2mOEFENchUYKr3SimgPGCZ8Xj8+vpqhKzBbmqtX19f8bVs5TAMf/EXfwHgziLAK+HAMxwU/IqOAZjTNzc37+/vw+CJYKw1eZaXZVmdz1mW0ZuSZRm/a7/f0/jCCLbYPsXodWvty/MLngOMBNyVoIdMAu/CkY6pfJBWKkCe2E6B+4mo9c3NDfcopiDgFpEsDqxCtz/XhAtLehRTXjYRUnUscdENQ3mDAieenoONhcFDx+HYfwb3ANkjvVJ4dsKbjLjzr6LjWJfG3F/XEqK/5QNZdPBH8BwlI3N7Eaxn4aIdIRmPpEAvityJTELmwVijWuY6ABpqrfM8B+r1MhYGO6uu8lYvYhVKuE9gCBHy4qDwDETEPDPfoyTtxayT5bEyYFycJNf3VVVRJHRSzCf0i86Mj7q9vf3y5QtOi5NB16UVjb84YzkVradBBhZS9+YLLxQdfmMhArpONPtYH/5wb2Hv4YHYXynwXEpK1ykUm6sE0/fed13LJU8Sq5QajcdcDAq/i8Xi5mYR3eEw+DTNu87VVVOUJaET/4uGDAyfFaFWkFxrbQgXXhaGnqUryzJJUndJYnpxQh6oh7yQpKQsS/ocoZZ10oQI1hFLaE567rC2sZ0eXJHUDSu53W6fn59xz0Sml3J6VSc2GY1GXdudjqf9bm9FYZ7WXGJkDGhibJYkQ+/GZbldr4/7Qxg8lTF0UHDD6Gihl357e5skCS0ReMqqql9fX7BxdAPQhEGyy+HMZTBwVVVKXXybtZYD472P2Mu3b99++OGH+Xz+8PD/sfamvZJcV7bYmWPM+eYdqlhFUmyKVlttt9sPz3hu+IMB/3cDPeDBz1arRbXEIqvq1p1yiozxjP6w8hyVGv5iqi8EgSzeyiEizt5rr7322jdoRCcID6h3c3Pj4s4f5y/LXBGCy7LAk4B4jcO43+9/85vfJGdD1JQ3NzegwubzOag8yLcwUHZ/f39zc+OjTey7d+9eXl7weF9fXwPAAVKAKIb3Bl7hj3/8Ixq84Fug9C3LcrFYwDTJO4cHFVQBSA4wNokxDlHehocKQTmFoERXgvnwUWiTCGRkQcjHCSE4ayEEJAMUfyxajSFOov0JdS8ULhDdIMfj0cIcHCqDb775hsYxQHgL8Ti9hOoEhM1/QHqgjGVFDkxqjXHWMkIVv3jwIu7jM/koB0REptGflXNOKXPOCyGVyoQQWk+zWT2fz1LfGJUHnsV/R0OROMOV53kmpdVacEaC984yQjilmZQsLgJLvkngqUy06LBRyRNCmMapKkrBePBejxOnTHKRq4wSVlV1pnKU+cZoY7S1BmjaRUtaJDzQwbhEKZTj+ZjP5+CpRbTTwi5GhKcs7hsRQrBo/4AfpCIIkHAaE9uDpy1TOQlcqayq6sjdZ0BvgCHYZZT6YCGqsJDMUvJDIAtxNVO6RDT6KufR8QaSDM4FY1wIWdczSpkQkjGe5wUhBJ8wQSdKKefMWAO9Y/Bh6CejbaHKpjm/7Hba2v3p/P0ffvh4/6nrO0rJcr0uypIxVuYFY0wosW8OzdD2XQdwAM1CaioU0RwU+QAsdlHkXJCub+bzqusaLsg4dd4brS/dTtxEoE5EQwxPIbThloGbglqJUop2BYnrkkQUswEcbLfb5+fnEB3Ip4tDnMJxyLNs6HolpGCcOM8J5ZRlQnDKrNZ3N7dd207jOPQ9C7TtOsrYx/v7Tw8PhNFxmrq+J5RaYi2xx/ZognbUtePZMz9bL0Y9eUKenp93u523fmw7r+3Q913XgQM0xqB3YoypZ6U2um3PWabgOQGDPGusEhJ1GChZH4Kxhjg39R0PwUyTFHwaR9RGzjlsIQUMwkErioJyJpQsqlIqNRktpMyKnAteVeW5PedFkRVF1w/tOM5X6/uHJ+v9qTl776uy6vuuqsqXx8f7+3sAUJRuGGYWnGcqc849PDwgd6I9Hkd/qPf+8fER9iqU0r7ruratypJRttvvhnFYrtdZnpdl+enhgXGWF0XbtfefPh1PJy64yjLr3DhN2hhtzXK1CpQY2M+M42K+sMYG7+uqqspq6HtoY0ggwflpGI02wfl5NaOetKdz17RjNzrtqCfe+OPhiJzqvb+6ukLEz7Ls+fFJcsEIJYQYa30IhDFtTJGr4K3gtCwyEqyz09Cfq7JIfmIQZaC/QikNnozDxLnwPnz55VdFUTrrbbSvfvXqVVVV9E/DXkUMGsV2e11V9X5/oPT/w5n7/3d6CJSwOHdmjQ0+MEIzdWl1IqLRz4zY0phJCp0x9AfOhblMn2VCcC4osi5ID9DKfd+DBCSxHco+GzHjjE7joMfxkouMloKXZZH4X/R2sB8iSQXKsry6ukLfOIRAQ5BCCM69ddM4dm07DgMJIQSaZ0WeFwjH3jvGaN+fCSGMMRRrqHjgAEHIZT8abh5SSOoKIkXhd/AZkJ+apkEcQQ2SCDrkHrAiKo7p4wro6GNaVZVzHr1N/FfkBqX+tADcRyeAJL4GJ4ZLkcpMvCkMKhKphaIbmQnB113c1wnMcZ3zhNAQiFKZlJc+hxAC6A8FB2U0MKKtORwP1lrBOQ1UcPH08On55Zkx3g3TfLmuZgtj7TgOjHOVKWtt33VKyX7sT935+bjDjhrQx6BTkSFIFIbleQ7qklKaF3kIzlq9vd4YO1VV0XVnHyxlFIEeQBLpE7dAxtU02GkD8s05h7VuYEXQy53iKiHguzFuooVCGlcJXBkAgZm0M7Y9nw/7/TSOaMwOfT+OQwj25eWprgoluZ4GKVhRKEop5cwF//j81HbdMI2E0W7oqWDVvB7N9PHhU1bmNngmeVZkssgHM3VjfzyeGKXduW2b5nQ44uykCXOAXMaCEHTSQ9OcOL+s5xqGITh3bpqyKKQQIDyZ4MYa73Tfnl+eH5vTUeuLOyySKBDSx48f0a5QSgVCuBDjNAkptzfXzvvTucmL4tw2Prhh7AklVV01besJabuBCZHlhSekbdtMKc4oC0EIit3DoEnxdmVZZnlOGUWXNclwMS2cosr5fD6dTtipJzingUguMN09aT3paTK6qKvlerU7HOr5TOV5Pw6jnqz3RVXKTE1GG2dVlqk8W282lDNOWXM8Cca+evu2bc5D10vOb69vvHVFlrdNIzh31g5dZ7R+eX6elXWwnrjQn9uuOXPCyizPs6xpGkilQYgBMZdF0XfdYj53znHBffCBBJVlwXslxayqMinLorDGrJerl6cnOHPUdZ1Gbay1xlgplXPeWT+Nuj13gkvnfFXV0P6C3UJQ6rrufG7x+0VRMsarql6t1ta6/4j0EEKISxDxJyjVP3etoXEYAn/u4owrUDBiFs6bMdpayzlyaQ6xYwKtKLIYY+iGubjZBhX0NE0+BE+CJ0HlGeNcW2Ods+4iWwpx15iKayfwgzQGXss5xwTnUgglrXdFVXpCRj15EoQQ2MAH1b+PbkWgVpPZIZrPEFmmjgL+AdStjBvY8V0QhhA7YO/loxIfNSZeDUQ/KlPGGBycWBQIoTDCyQcxgo8Hl2yENhtXXYrPlliAKiWEJM0+8pyIjkx4KdCgKcGgpACFQqMhHRJVYvNUtG2fohdTav5zzkLwWM1wPp/HYRCCg8BFiEdnEsRR6tKjuQ2mG/bLNG6mJISgLiaEwDk1hLBer/HuUFhxLoZhdM5nWb5eb5bLlRCSx3XiSJkgNCDixO3AN4L8ZrPZIDFkcR1s2i6F5yeZgIH9p5Q+Pj7iwU7Xc7FYMEaRw0BD4yLDSgRMy7/9278ppc7nM1yMJq29c0WeC8bPzbk5now2fdefTo33YbO5IoSO4+Scd85778GnMcalksaYc9sej8fHx8eHhwcosl69egWM37bt+dwCpuC+u+i96JwDewO6g0eV2jRN8/kcm9uR/HCCFotFWZYosHS0vKzKUgrBKHXGzmczjEa/ur1F1AC6h9IGXA2mi3GcT6cTVF4InTAchcwX7Zy6quaz+Xw+//bbb4HEEX8SPYgshWRv41Q89G8oMdEY76P1KXjgqqoShsP8oItTEWDbQMwmOSWCGOAvpCtoS6BX0XXd0/MzFwLDX+M0GWsDIavlCswY2plJK4H+AcN4Y+R1tdY+hK4ftLGUcyHUYrlyPgRKjscjCjVoc29vbymlNKouP3z4gGuLyYa+7/Aunz59wjYL4CqAPwAdnDv07f4D0kOiiYCRx2iqfD6fMTli4tZD8edDDFFUd1k5gBjqPeymOSFEyovFo4jLcEzcwYkSCawINlpgluR0boy1k9b9MFjvCKVCScY5j26moHRUdJFTSs3nc7Bv6BCM4zhOk3WOUBrNALx1bohrf4CmkRQR7IDH0UwDDY2XwpVBceA/cxUc48oKQkgaKUgpAQWWiF7NSGbg1vGOSMksbiQFuYS7i48XovAJ4zyMMSjwfBz2SRJMfAYV/bRx1BEB0xecz+e42rjdNq5AIITk0RI1NTzQxrgA9jyHxCudQ3yGelYJwQnx4zhSGsax12Yaxj5pabDgc7Va4aDiquLv6rh2IouDzeM4wsEYudBFMRjmpxCgp2naveyaphVcno7NZn3lXKirGQmUUoaSHBoVHudCQqDeB8aE1rYoSue8MbauZ1BnJsiSfOJQaVFK7+7ukCEwnQC2CuIWNH4ppdOk0f1G0h3i+kIAguVyeX9/H0JIITLL5DgOIXiVSdRAzhkpOQm0bfuqqouiEkLmWeGcX6+vsKJunEZrTF3XUgjcaEjs0IqAPV8CZLiSj4+PLrpBQIgBFhthCxMwIEJxHUBr4MmEpRXQErh49GC8dedTs9/t9DgVeT70PaNsu92O41DFNVloCWCeK5W5KMhQE1xdXfHLtkGJGb26riljKlOEkLu7O2wYxZOAfwBI2m63AKlgOKGdc8598cUX2+0W/QwoncBK4WtiVAV5Mel68Y1U/DkcDoSQ1WqFjtc4jlj4PI4jDAopzFOz7Hg6nduWcZ7lOePceZ/lORcCoATfFF5hCe92XZfebrlcEkoYF86HYZqkygKl88XSE/rm7ZdQ1uAUn89nyM98CEVRwCnrcDjgF7z3p1MDIGiM+fDhA+4dunQ4JqgFN5sNj9vP/tL0APKaxcE/VAboL6X5ABoXfiHltm378PCASiK1eXHm8/wiOU2K1RA11Ol10OXDXU89W2QL4+yoJyEFCoi8yGfzGe6Einuk0+8j8OGZRtC/tCLRAZYSmcZa60MY/9RxzaG8wkMAPTv+FbxEYtXwINpoKkWiOJJE3ycwMzz+QN2BljghhFIaCJn0hJQ2xsX0kAWDZ08wJzXDs2iECUkJQiqsV1zUt6XLhePn4hrqMW4fSn1jFm02EA3xajSuVzPGYJrhUnVFIa+JowyAk6hFXNQCIvFzzosigxrHewdcA2CBz4M3wk3Jok1WUisAvmEzO7AFegbI9yLaO2PcFAsSGGVKZaigp3HiXJTlRdSBUwpZJ+LI6XTq+4EQwhiTUuV5EQKZpotwEIIWQDxCyMePH+/v7z98+ADHTRoloUDccDdCjoQSF/IwFMGoybq4vQ6IFXuNkG6fnp76tg3OtU2zmM0oIVKKse+D99Y6a6x3oSzKcZjevHnLKG9ODfAKY8xEclJrXdf1q1evcOzRGAP7DBWmjSvGHh8fcVsxvY9BCnRZvvrqKwRT9Icx2AwhHyoJPC2A/PjrzamZVXVwXo+T1aauas641X+yBGbRngjgF76kCUsBbsLBhRCCoE8pveyaVIrHbQ2YeEeiBdMCthbj0zaOv0DvgAcmRJ1IwqYmuuGiiEHPuSxLiA/HccQOWnhUKKV+//vfg9vBycLZwQWHVBRHnnLWnM9ciNliLjNFKB3GsaorfDaIVnmcmgauSkIb51ye53VdcyHreuYDkSrL8mKctBDi3Hbz+fz9+/dJiIhGOiWkaZqrqysEW5wXY8w4DA8PD4vFAiD+cDg8Pj4OwzCbzRIjoqMJJoZmfsbPn3kuvXv/RONSz5QPQhSreO+hrMKjAC9yUAqoX8RlqVzg/HKknbOUEueckELri5ozFV/obOOpwoOeZdnF9NSYvCgYY1meQx2xWq2kUrP5jJKLTSZwOoD5fD6HGgopDQ8ElEbaGBEHmhBqjTHBEyF4lmUAXFVVZllGKem6HhWZ9z6pchFWEHHwoAMdIHIl0gm/QKOCC4gyxNkco/WkDcJQAuw2KkQRiBHR0HCGA8+7d+9g1qiUur6+NsZsNpv3798DBKUKAxkrYTTsHgEV46OxBF4W4QyQykZLfTxJx+OxKAqtDQoOyDOSthrXIbW7wQI553xwZVlkWcYomcaBkhCClyoz1mAJF/6itRbNACklsDkyLr5FUo7heAPMQteBw2DicLgxhlJGCDXG5nleVXXX9d57KRVuKE5ylmVwbpimCUYwlBDnPGecMSa4yFQmBH98fMQnTGUxUhoaTrgyqMYgB/CfLU/FNFNzanT8UkmPxKN3FupFcA6AGkJwiK2+/vprEjyaAeM0Klk057YsqxDQwRJKZcbYsi7zPO+7bhonwTliqzG2KEuExUSujuMopQB1hieQcw6P8SF6fCEi43dA+u12O1AoqB4wAKGyDC+eiqDVJnkAACAASURBVFo8Wehb4Nc4Y9658/lcVMX53DLGrHXOuaqqnXPe+f3xUBR5XdfeWhLIMPTTOBVFnuclPmHSuTDGhJRKKcpYav/gLrx+/drGjRok2lziw0/R9hVPOD4/0ABKPcSBEJ2YrbVQNCE9QGFI41CCiy6faD2mCIDzBX7CGEsoVZmC/SrmRSilZV1CK5+KBugJUTeAPKBxTRml1LvA4860sqyMMVKqoihDuGxgAw7DB+OMn88tKo/7+3scQCml8y7LFKIZLlFqHYGvS52zPM+Px+P/8b//l5+RHv6sekCpgsogiX/BmOMfOOdQgKUKIyUPFHcoxBKN0Pf9NOm+75HZbFzGCSyQEkO6kbhDuHnOO8qoVLIoS6FkXhbGWhJXx7hop4y6BPJH5BjGGCIvPvC5ba131rlxGn3wxlobm7cpnyOdAH766M+TUHNScKZxNmR4YPAh7pVEXCPR20PErevgoHn0rcKpA8Y30TcpdU1w9fBr4jOXEYA7zOYkV2ec1VSxAa2wuBIVmQAp00efDxw8fGt8HRctzBAK/Wd7MZE80qfVWkNKX1UVDqdz1nuHOz+Now8e1h3L5QKVaDKMk9ECD+ADyRuXa7lc3t7e4mTCbwBFQJpvuL29xTWPx+biBqa1OZ/buq4xHJPnBb4yNJE4jUVRMMqVVFIqwcXpdGrPLQLZYrHEHb++vtbRTQvXGY1oGU0Ycek2mw0oOJx8bB/ruksTBTcRqxoSyvvxxx913JhNKeWc3V5vMykk5/NZfXt9HZzjjM7KGnUSaHelMmw+SDHlu+++g2pr6PssjjGSaHOAALRcLquqRrZAsZjG4MFs4JFA+XV/f4+KGeMF6MMnyherh3Cs8HZSyqIsg/dKKSXlYb83xpiouEOf5vr6+nw+o+q6urqC3T0iwzRN6/UaTz76BykQg8MkhNDoCgPCh0V9I0Y9AFBAYaG4xH5D3AiI0IZhgISh7/vlcqm1BqNAKb2+vmaMwTycEALKZbfbwf8YPXAsB8PzM0ZLR4R7zG3UsxmXoqhKLkRzbgIJ9axerlfOeaABaE/LuHgY9w6dy8SsSCnzsqScF2UVCEUt4gMhhMKWVUqZerrTNBlr8zx/eXnBdmGU4zg7x+PxzZs3iWZArYB1VfjwNA5yYcDiZ/z8+21xQGo4lizKOm3U6SdL5GmadrsdSlo8iwgcMeRhFC6M4yDlJU9aczGoQHOVEAJ4nuaxURMQ2IuSIFXGONtcXSFE3t7eaqMZ48QTE7XG4BnTXmi4uiO24nhXde2Dp5Rqo8uq8iEEEoZh8NYzhjoJ8Vdba4sid+4SLlOZyS7mppdlmSEqRIHEUzQHoYQnHmQCqBJUY5AqE0IIJTa6hIJcIlEk+jnzg0iBmAJPfywvxC3HkAe+O45Q0pumKgQZHSHexQERpIRERoFiBsFKCAEmYoyHuO4UP3meA2VrraHjxoYc770QzDgN/w9n7TD0lJLFfHZze5cXhTEGTwWuA5wKEfc555vNBtMG2+2WEHI6nYASPn36tN1u0bp8eHjAx57i/nBr7enUlEUdfAjeT1ovFwtn/TRpIS6QJXHoUSwgjbFCwGs9UyrjXHRd5701RgOO0TgCUhTFfr9HPAVVgtIeUZhE7VOI0xuH/QFNCxqFcyDNE6RFigrRPOOvvvmaUnI8HpWUX3755bt3PwjON1cbxjPGuOCCEuqsy7JcCumDf9m9hBDKquKUKSmHYfDOqSzTxqCNSQhBh0Yp9fLyjBCcwCPO46yu8TxorRH7Xl5ebm9vp2nCljQSZSlaay5EP/RItCjQUQZN46ikgu2Hi+uhyrJsmpN11hhTFGWWZe/ff0AvvarrvMhDCIwSwUVeZKvlSkphrUeMRhwER00IoYyC7sfTstvtEojGqAqN3jMubukw0e4JsD1tRUWvSCn18eNHzFQmUynEXBqnWYO7uJtsNhsEMQxqoLHH41wwcBsXwnkfQlCZMtEpB2+UYDtQMjoroENcdIcDHm3bVskszwvKmMoyPRkpFaFUGz2rK0IIhklTYpumaRwnDCrimqNWqOoKVBLuNWArHj+AmEu7KAr9/7f/9e9+Rnr492Nxn1Nmf2K9KXOBiCzrxt4E56nXXhMaUlhHxzyCmiAEw5rqqqo4F8Mwde2g9dh1Z2OmaRratpmmQetxNqtI3Mo0n88xoBRCoIQarXOVeeusNkpIGsg0jMFeLCGTGjrV+3iIkUIh2WaMeec4odSHYJ2ZNPWhygvifNudJj0aO0G0yrnI87KqFlU1cy44FyjlWVaMow6BFkUJAkRFM9pEsgEkJr4lRGsmEhWZ/jOlUBE3bQHb2rguicUtJSTuULqog4zNVVbmxbyeLecLTtnYDzSQzWpd5oUSUnLhtCE+5Cqrq/pwOBhjMEtB4mYxpRRsv3Ak0B4E1ELLF3kIUQx1MRIkqH9AoRCn5fGcpPEoQpkUahx0WdRlUZfFTGtXlnOtzXy+tNZvNtvZbOFcIIRV1Syd51QGcc4hZGTRDAqJBEcd7wUkxeNEyDgO49QzTqy7PHKMMe/Dfnesqpn3BMdVKcEFcw7GRIiYDJKw2WxW17PTsWnb3nuCeGSd7vo2Ly4NJIBo1BPwfSnLEp4QWJ+AFppQOWFSqHyxWmvjjscTpoI5p4T40+nw8vLUda1SoizzxWKWKVZmcjGr9vvnxaIqZ4Wj7uHlcdS9VEwqQRkR8uJtVVU1I2zsR6vtcrlerNZfvHlbzxeEMXQ+8HXauDJsNq/rWakyURSZkKwos4/37ykLTHFHnMrl4XRq2vNknMqLf/vjD5Oxp7ZbrNbdMBLGqZCOUEp5cL45nYauR+EzGq2dFUJ4Eghj9XxGOT81Tdt1bd9zoabRZqrY7Xa/+MUv8ly9vDy13UkQMi/KSqnT4fjq7k5y5Vyo64XMlA+BMHo4nebLpQ3+/uHTy26HAa6rqyvQtpvNRmutjZmvlv04OBKEUrPloqjKQ3M6HA8YS0b4Bnxu2/b9ux+J82PXT/3wV7/4ZhrH9+/fW2sf7j/ZSc+rmgaSCemMrYqyyHJCCDoWu93u5eXlyy+/hEM7uAHARLy4936cBpEJmUsmmSOuG7rj+TjowfkLT4tBGaRqrXVRF4Me8iqfrGn7Xjsv88IT2nat87Yo8jzPAvHntqnrqqqq46HJs1JwVVfz9twzypXMyqJSKjscTiFQY5z3RAh1Op33u8NsNsMGchA8wGFwGYEkJMQlLp+jvZ+fHqZod5yqP++9C54zHggZhnGA0844eO8pv7iBhovmmqVuJEEzNqpu8AMGFm/UDz3OJLpquJq4zS6upUQEdMaaSSMsVkWZdDUhhCS3RdUGVT7KHQRHG30R8GEYpYQQ8CQhBGN0aobD3wmFOU4m/i4QxDhOic/Fw4SboeJML3AW/lxfjDMvTWY0LQFtdPTxDiEAGZFolo5OLL5OmkXo2nZW12VRGK3LoljM54xSzhhnLM8yzhggn7NWcF4WhYlOdiBMi7gWEVgYvE2Iw89Irih6cEjASmPxtYteRni2wB5gfJdSOgwDMoeSqqpmIRDnPOciy3LvYMBQoivYNA2EZCiJNpsNLjisdRInieQE1dl2u0X7FOwwmOLiM/fZPM+1njhnw9D3fbfb7znnlLLD4eKSgr4IF7zr2qLMCA3OmdmsWi7njNFpGo2ZiiLjXMAOIbkBDkPf9+3bt2+NMTc3N+DQ0L8FagEiQwVM4HAsJeE8MIZWWdd3IL4xl2utfXx83O93TdO8efPGOy+5+vLtV8vVej5btP3w9dffWOfHcez6Tms9TUPSwg3D4KyTQnrnx2HknA/DUNV1PZuhqQBoieMDEma1XOJQ5EWG5ULTNB2Ph/P56IPX1jDOrHOTnnwI02QmaJdDCIRoY9BfHfVUlWWR5c655WLJOLPWLhaLyZpAiMoU49wHLzN17lrGL6xs6tXf3d0RQp6fn5Xkwbu6qmZ1/fHjxzzPN5vNenP16tXrUU+YVNgf9kII5/0wDuCdAFAACyilxmghhLG2aZpxGquqWm82s9kshIt4F713Y8x6vX779i3n/Ljbm7jl+80Xb9brddd1zlpCSNu2VVlKKYP3Ssq729svv/wSPQMw4f/4j/94e3sb4k4z1AfAc6nvSONwKBcc7WLvXBIjILxAJZjnGR5FKaUPHh9Ayos0AzFzu92OcVUabqKMm8ClVMMwAhI1zanrWsaolGK1WuZ51vdDwkyXERxjlFLv378HvxdCgAQWdmf/AekhxCUEn6toJBeZUhg89taZSQ9dr8eJEAJoSQjB76fSHlfWREsJctG5ehLBICWX2ISWGjR/cD9PKn722d43KA289xiaB+SHgI9Hg4cxOtChoEl1MeAwRJksDuWpOEsMyM+iXggkD7g//F0wtukbpeYE6GbUDXg1EtWl6Bngvwoh0CNVSkEjgWcacn68I8wEoX0C3USjAQlyFQhQJD//mS1gapyQaOqnoi9FSvMurpOjny1qRebDo4OUmWUZ2GFCSBZtrNI7ws6FEAIeDyyTEIKL6HhBmRCSxMUJ2LoOyUQqqpC3iqJI84C3t7dI9pRSaDNWqxWYHEQczI68vLwMcZkgbHlIpNT2+/04DJxzpSRjzDtPKUVVhMZJFr3N+76/u7szRjt3mUAE+gGlsFgsXr9+jZuIQh4kmHOuLEs8n4fDAVcMkzH4bFhFgIIbtTJMXLDvF8l1NpshIqxWq2E0/Whub19LVQyDGQdTFrU1oTk1h8NhvV5vNpspehsnY2q4JF1dXaW0mtTkUOCATX39+g34z/v7T5AzbLfbvh8+ffq02Wxub29RLKKflCRk3nuoHgDv8EiAuarrepo0mr14tn2ceYLsB4UUIQSDqB8+fPjrv/5rHNjnlxeVZf0w3N7d7fb7YRw9FqspiUsNG1Qg9NS2BdGPgyaEsMb+62/+pS5KTqjTpj+3s7LaLFeS865tT4cjIzSOs4n1cgXUhUP68ePHL774Avaa4FRh/Q2chLWdNzc38EnFX3l5eYFW2MbF6SD6QwhlWXIuMiGdNtSHXKpcqkxIFsjjpwcz6b7tJBeCcch/iQ9Flishnh8fifP42GPf33/4gO4InNVDCMvl8vn5GU27cRzLsoQoHE0a5HgMsQMTwy5FqcueCaRe/Jpz7vr6GrFxu93iwUhk2s/4+bPew//9L3/AOUQxpbU+nU5KKaUye9n0Say1QvA8z0PwVVnjmI3jiPaOjg5CaMeBr4hhi8KeAUDPOU8pm83mwzDgzKOtZKOVJjwDSMT7wI+73Y4xDt4cJDsMliHFI3EtmozmdFCbweKKEII/NNFTD4hYRetTYNgpGqHQP9+BikcEehVIdZMgATUNAhZquhDVolJKbN2Zpmm+XAYSUBxst9sQNXCArgjraAaiarbaoNuWusRQ9JM4+JbHfah4937oYWMFP8thGO7u7hIwcdHEycapIigdr66uLiZ6URfUdR2Mg+BTbeKwHtaXogJAcUYIZZxPk+aMc8abU5OpHH/ovEPmw8RDFEqJtHwGgkXEX2REnF7OOaoHHvcyoQBCyxqxA2L5uq61NiQQHF2tp34YtturpjkxxvI8Q2+MMc6FaLt2e719fHrM8qyuK8boOE7jNGqtQ/Bd1xVlTimt66ooKtxEVOgwgBJxowngqogbA6XKxnGilCwXizzLp2nEcXj9+hUky1mWwSecc/7FmzeB8g8fPv30/r45d0KoP/7hHWOiKKqu78BpXF/fjOMohSSEeu+8dz5a5FJKsSIJTwKNg7UQRGqtx3GQQu33h6EfBZeb9dXVZvvysiuKTAgYFfu2baW80HFQEPR9XxQFnhkcQzNN2PqZ5dmkNReCMbbf7xfzxaUPMU2QhO52OzQwCCHIyrCePhwO0zgtFgu0UnGgEGcDIcMw4CskAjaEgAyUdMz4B+xr+bu/+7vD4bDb7Q6HA9qzxhhG2Rg3gGqtHx8f8zzfrtZa6+PpBDiCtxNCjMOAyg+jZwnCbjYbiGURqRHxYGyHBx4qG7Q5rbWUXXz+Q5S8D8PAGdvtdvDiBTkP+D9OA8oLuIQtl6vgw8vzDsbGKDKAFPEB3n7x5scffwSpC4+mRGYURW6tZYzCxpFSygVFJnMX+6xL0wKCtDTPiMJiGIb/mN4DMC8AJqg0IQQlhGM81IcyL6ZhHPs+OA8vSR538lg4v/c9BM7o5uOBQ/2LkwOKJrExoAtwfiDjxQ0Q0ZUaUmX01jCChMaDtRa6Lh89eE2cO4MGBjfg1atXaE+h24M16xdrgbhTM8mERLTARXxnUc+OugSBFQ5fqVOKCerknAxoidfEmUH5zzk/Hg7WWsRHgAXcPBsH6/CTmCvvPUZtsyxDOMjzHM89RmBMXDqGiJnnOZZJHQ4HJIYp/uDs4QAMw4CMMpvNXl5eMMKNM4yeMMgoVMpogCeiaYwOo2m3e2KxcH3KquKcY608PjONg2+JbUP1kOIsia6oqEhA8R+Px6ZpsPAE1RukUH1cTwalaVEU9WxmjKGUMMYYpWBChmHASjXvHWzB4FSalC1SytVqBS4LXxl2CGAD0OLDA4DKScfdQRBQQa3f9713l5WC4zhut1s8UXVd/8M//MMvf/lLJMI09frDDz+c236+WjEhnA+n5iyz7Lg/KZVBln08HmVc1V7XFWqsN2/eoIGso3XoarVChxkza+BFcWUYY0ghiCNSyvVqvVgs7+/v8VwBOzPGTqdTatHhIcdANZpARuuqqtpzu1otOedd193dvbq/vwdiw7bLEMJ6vX737t3j4yPK2bqu/+Vf/uXXv/71r3/963PXfvx0P1vMm/ZMGD13bTf0bd8h5uLewVwSsAxnBIEVRTY+p5m05OKrt1/qcTqfmp/e/SgYB6VRFkVVlkWeX202jNL3P/309PR0d3enlMKz3TTN8XhEDxX96mmaYAlOCMEEro1Dpi4OKn38+BFSdTwAcO0GAV6XlTPWajN0vWDcGZtJ1ZxOmVJG6+enJ2uMdy7PsqY56XG02nDKHu4/tc35fGqUkILzxEwMcUcpBEu/+93vIMv+13/9VzCHqZMBmIhwCnkSDLVAUONcQKkFTXy6kkhvQGB/aXqAbDylLDwHKG9BkSfCh1IGeAgu28URwQSlAf0S9BZCMCYo5c4FIVQIVMqMUs65xCRnCh8ARxhAxWuiBEbiQaxJFp7JgXK5XOLaAW6EEDDHiJ/lcnl9fY1vl5TpKAhQWyABwMdCRAdd8FSQAfi4shQtWWB22BahWgKkQuWBXYPoFm632+VyCfUtclgW/aWTLgLUk4+7hlCbow8BpRCiwE8//QQ+Cu8i4npLlCAkOt0iBeImsmjjig+JBIbEDMyYRav6EALUgWBRAZmRDBCXy7JEUMDpQmVGKWOUCy6lVM66sqyUUsMwPj8/CyGenp4w6GSi4zohBCAA33S322FoBkcFJCEKVsBzNISzuDsFogCUSiGuPEIGQkCp6iqBD0KotTYEUhQFIZ4Q33VnpcTpdDidDrvdC/AEZD9lXAYFZvnNmzfYbZCm2zBYw6JgBprd6+trQsl8PquqinGeKukvvviiqip0LNAnxNx4lmeO+G7ohZJc8t1uRwhpmubp8amu65ubGyklevJw3IITRt/3r1+/hqDg/v4esgL/2UZbbKsOIUyTWa+vpMyEUOdz9/j4/PT0wpjw3n/11Vco8fFIK6Xevn3ro6evlBIpE4dXRa9i59x8NsdXPp8b1BCQ1YFbu7u7Q5cFiQduIuM4vnr1arvdgodYLpdo6fuoEjbGAKMMw4AeyfF4fHp6KuM6YR99B7quQ3VbVRW4tYeHhw8fPpzbFrmQEHJ1dTVNE2o1TAJiPA1cTaKR27bd7XZwQsXzrLV+//49ggnibzLuPR6P9/f34GfQNyWELBcL7z2OJATB0FYg6xBCEhEaQiiLAkYNCEpY5sM5B/eIT+6cgxsF6AGsqbi+vkYTLspVAhdUSDbpoazy7fWmKDPnzY8//rher8HE4qpiLbwQYrVabbdbnG5ECawh+hk/f0Yu/f6PH5MUB5Q3IKGgPHhPQgje1WU1jUN+8UkWgBvIB4i842dGTAhYuNNQOSXWEoEsz/Nh6N1n2v8k/kkVAGpzcBSc8zwvUD1gBvLVq1eYh8TjgsYOoCjOFexdP7fJQ75BosbZAKQCMkLwRSYAWpSfOQkieUArwiMQQBzBSUOiTg2PLFr6gLufoj0DeA+AQRonRUQcLEcPZuwvGRqSAUyopTaUifpj5GDOeT8MJLLAt7e3v//971GpoPzCe2GcG6+AhAErp+fn56jny3mUycq4WgPZEWVsVVXwwV6tVue25UJQSjOV7Xc751xZlFrrqq4en54SpQtXH8ZY05xCdIRG5sCTg4oE/GkWN4fLz3aHoPOJIuCnn35C3j2dTldXVxismc/nT0+PhJCiKDhnp9OJ0JBlmZS8aweEjOPxeHd3B8CYZYox0XYoEegwDCiPp2nMsxJwDHgtRMUwi6sxUcDFyakRpcNysWjP56qqTqfTdrudz2e4R8/Pz6DjVqtVnhc846e2EZJLydu20XpUSlJG5osFNAKcCxwWtB82mzWo51QiIzui/GJxlzuJli1oiiIaouamlOaFgmRzs75q25ZSZq3drDdgpNMtxlxnwoyLxYIyNukJm+CsMc46NHKQ/pumwT4GFN8QF4CBpJRyJV/2u0DIf/erXwVKXPAwT2Pk4kwMJjnLMqRShNdXr15h6AqjGE3T3Gy3lNKbmxs4fCgpz00jKKuKQk+T0TpTKlNqMZufT42HoVOm1ut1WVV1XQspp2kicUVK6u1BYQEoBlomrZdYrVZYzaS1RuvYRcPNTKngvff+3DRFUTSnEwlhGsa2ab7+8itGKA2EETr0PWMszxWMTb0PRVEc98dxGKRUzjsgZrQtgQiHYfDWffr0CaQCIGNRFHmeEUogc4BQAsUrYCtKpXEc8a+okAA04f4N1GWt/c//83//l6aHf/39TzxqRfq43odSSgmNWhe3XC7hnEwJ5QJ2qDxd8dS50tEPncR+AOcCqETGaSMctq67UA1V3DeJDIEMD1rNex8FZ0rE+UDO+fF4hPVNutA8Os7med73/du3bzGrgeYkyA2UbLgHQKZo0yUIj1SXeCckADzTuFA+mlOllIbrg3yQhOFQ7IAREkJQxnwIQghwrzTuKwXnDiAP9Iq/xaJ32PPzMz4PBLJ5UaTThaxDCGGcwzUEyenq6gr7v/CbSFr4anCOC3HnFC4F1KVKKc6Fjt4MwMvJw0BHP8HNZnM8HmezWdf1hDBjbJ4V0zhxJjB9xjidYk8bLXeE17Ztu65DLxdRFe9Vxt1nIDowwoL3LeL6HYy8InADTHRdb60ry2och+vrG+ds27UI8avlctLTerXa73d1Pf/06ZOKHsPTNGE9r3ehKPLz+bxer5qmCcFDdhU8ubu7wxR0nudd1yNvoZQBIkZ7UGtNGDPGTtNUFkXbtpvNKlyejYv/GlrEfd9fvMUKud+/eO+yXB2Pe2dt156D85Rx/MLxeJrP5+25xfWvqhKHgkSv+91uV8ZdVTitIQQpRZZlnPHZbA4hmRSSUkYJFUIQgq1tnlK4gBRCiOXicpBDXNmSAtbpeBiGgQtRlEXTnAm7eJ+YaaKEoGTHuPXxcKxnNc4Fbhy+7PF4vH11h0K2ruvNZoOAyzmXEUuhPkbreDabwckVU5Dff/89CrVhGKqiRNJCjSKEODVNWRZVWS6WS9CM/TCgaoe3I2VsPp8vlsssy6SSnPOh60EIg84CetPR9QutHRM3quJsoqkLAgenAFAPJ65pGpxrY4zkAt99u90eDoc0vy2UsBZrlKqiKDgT4zj2wyCkqOoKxyG59QkhJL+4VOEj0YtFkLq9vUkrNQGaUbkinSdql0ZfRchD4DGFMx5C+Hnp4c/IJXxboGxYL3Rd13VdYIFw4qkPjHIlKZeBck8ug77ADiQOXY/R8C585jrnnCPEC8Gs1dbqLJNZJqXkITjMEPDooYSyC4EjaUu2220UBfGiKEDVofiAeCvN1pHPFnNiyVoKuChBZNzTS6N9CIlr7/BrKQGkhnNCwUghgE4uLlnDh0/vDpgMRg4xHZwgapHU1EKMAyS8CHnLkhCCKEZgbkrJbLko6ooKnlelzDMq+GTNZDRXsqyrQInKc8qZ9Z5LgWILYuIff/wRnpQIaofDAdpW9GBcnBtPEfNyJEhwfsxyPptXQrBh6ABFgyckrnEFfEbWoSQQb72dQrCUBU+dp66o83PXwiEVB89Fy/HlcsWYOBxO0zS1bVuUGRc0EAuaCIXFn7QAJAxjqzLhPDZA5FJm1vqrq2vc977vPn36eD4fA/HPL4/r1aJQknr39OnhfGqIJSzw4Tzhi7dt+4tf/GIcx6+//ho2A4yx+Wzx6u610S7PKylLpUolKkaVntxquZ3Vq/lsTYkYej2OmgsaiMtyVRRFCGS12pDA1stVmWdVkXdtK6TYHY6EC+tD149MKE+osf7YnF2gw6SzolyXq5mowuQVUcv5pipnnrB+0sM44jkhxA9Dd26PKhPajO/evUNiVtHXRGv98PjRB12UouuPgei8EMPYBmJm81plsqpLY/VytVgs56fmSEiYRte1Q10tQqBVNbvQMjSgSZv4RkweEEJUkQdGJ6M5nnMhMyGXs3mVF5DuBOuYD24ywdpcZRisQTxFEyXP8+Nuf3O1LVT28af315srybjiosoLrmRgdDSaSTEa/e79T/00BkaRY3766Seok8FMFkWxPx1dCPcPD//wz//08PToQri+vdmfTqeu3Z+OVIqPjw/au15PozN5VdrgwXUrKVFVzKv69vYWxxNSFB4d6TExAI0QsiMo+wt25vzHH38ExIFSlke7FIw9AfWPRs9XS+2sDX62XKgirxdzJoX3VKnc+8vWyNu766++fnu1WTFCJBO5VCyQ4+5wPjXTMNpRN+djUWbvP/zYD60PvTyZTAAAIABJREFUlrIwm1faTFrrxeJiQ4DwlWXZZrO5ubm5urqaz+c3Nzez2eyrr76CtBf5rGmah4cH7EPNf65j659VD//tN/+WOEce5S4I8SJO2IOzE9GOhsWtojQuQkisPfts7Q8hJFmUoG1A44ohPPQkrt5MoxLQ2KF7s1gsQPfDCCGp+5PuGz+Jvsc9ds6B90dugLMF+u3o98L+EJ+NEAKUiqlXfPGk/0OiTj+pTZ3HxYp4+EJ0sIAgJDW6cWu99z5capSUTpALkeFQFYFHRvEB3hArc5Fuk6lDXhSBECGFyjLnL3OkkFrVdf309ARcic+DvwX3fIAm/FcI0tJMspKCUEIIFYJTAgs/MQxD05y54EVRQNeBQwLoNww9Wtnhol4lXPB0W2ENfTwekZnyrLDWTNO0WMxDCOAVq6o8HpukaQakIoRQFigl8LuHzfVisXSXDQGXaWS8shDcGC04q6samg3wjReJC2OTnqBE2u1233333ffff1/XdZbnXPDT6SiV7LoWmS+QILi4fABKvXd5np3PZ8qItZMxVillrVNKrdebcZwypVarVZJaAUCM0xTIBTN1fee9xzV5dXdXqPzm+nq331NCF4tF0zRCShPX26F/FnsYw3w+P52a169fY4kFunFCCMbofD6jlIVAZrM5GD+t9XK5RuzDsuJpmlBElmUZwqXkBVIBUbZcLvf7Pc7jnzCvEEJwjI5n0dq6rusiy4nzJARo3K2xXdvWVaWtKauL4SaJc50hBOwZ3Ww2aGh9+eWXu93OORfoRZDDozkxWgI0kCRvWa/XDw8P6/UacJiyy2K1w/GY5RmBdqDrbu9uhRTjNDbnxjkXCOGR6BZxmfOrV69QsCZBJmI9qEsSW56oYKA4ANmATJxsS8BwAjKCGiLR+09rjTb4er1GLkGQJFHsk+Ay6jw9mUwpRhle0FmnL6tuCOaR0X8GIfnu3bumad6+fQvkhzYMbmLi7REz8edff/01hNQQpIBdHMfxf/lPv/5L08P/89s/wknRRIs9IC/EO9xRFNoIiKkyJZ9NtyI3JEI5/QKANkI/i4sHeDSnA2WBMwm5HtJDktCh9KuqktKLpRIYMAwTgBiJTQ7uLpttLgUEFCDJQwKdYXQR8XyQ6MwFOKCj5zu0LowxZAW8PiEE3WmwmXjEafTDwrdDyYykBY0H5zyQYOLhQSqV0REPvWtAmPRIIT9hfBTVEqYcwCyjJ5+IYxdnOwghaWbSxe3WwPssbp5AhkiECYJCnuc+OCH4OI7L5dJZ2HQzrXXX9bBkj0JnBVpjNpuhyEO/NBlUwM/Re79arRhj5/MZvxl8QKBcrZaMMUKwr9F7T3DLEKFwza2zdV065xljsFyt61nf9YQSSANpdDPEhyeBKCmzLMNmNyxkRxyczedoCTjnVqsVYs1sNqeUPj8/p9uH37fGJn8UAMw8z0+ng1L8+fmlruqu6zlnnAtj7DAOq9USrsaUUqz2zPO8aU7AB1B5IOgHQozWi/n8cDziwVNKqaigw/EGAQunUu89vCCFEOjTjBdfZIf8CnISrZqyLKVQuKHz+RzrATDp8vr1a4S80+kEcSBkFIhlCSsopfb7/Xa7PZ8bQsjz8zOMNwBHNuuNsdY4ywRngh+bU1XX8MnngsN4HBuopJRd193e3gJdISgvFgsI1RilGA4o80IwXhaFmXRzPGVZBr+jJE4zxmAdhY/76vu+xy/Ap3I+nwNSvH///qJHJxQTsmiEoEuMqIoiDOgh6VzQMweDhJPivU97uRFSEBZwooG0dDSbAnTGL6MNno4G2JfUip/iphzGWFXV5/MZYyghaibHcTT2st22qqr3798jv7Ztu9/vx3GELwh82q+urnDuELjQS0f8/Pbbb7E/7qeffnLR8nIcx//yn/+HvzQ9/Nf/9jtkeLS/8XAjmiilSKwA0CLGHBPCDZ5vG/cHpBYrriPyTV3XOLG497g9eN/VapVHhz5EVVxK5Pxpmq6vr9ENZozlecHiiAP6yfiLQOgJGiN8w70OQRPG40gbyAEIYRA7g14EoMYOcWQUwH9KKV4EUTvFbtAgYId9nBuCRAq5BxJJVDPGmHGaQDfh0lVxsTOSLgrz1G2OeJymsY+k3QbMRPkFpS+ebP3ZOPfT0xMk1Qnso4GfKEQEstT2kFKe26YosnGcrLXOwhRWhhCGfuj6DonqApDHEZ8/XYeExbz3wDjn83mz2by8vCC3WYu9V7brujzPOOeMw75FeHc5Y6mNH0JgjBRFfjyevA9CyL7vyrIa+mEYh8VibuL0ootyZBafpST3xOuM07S5unr//v2XX37pvX94eEBxo/VUloUx+nxusNQI//MhUEYpo33fzRdzymjXd4tF/fzyTCk1xoZAsGO1yPNPnz5JJTF8g/05eMayTGGEEHENCcxobSbtneNxbcNyuRRSyrjCFncZ0AobkKJtjUhaj4v+TWVYJqgn432QUjHKvQ+Pj09YxIbHGxoQiG1IlK4BxoLDQbRVSoHXxc1i7OIdnXAJpVQI7qybzWa4etBzG2O0MYgM+EMcFsyLqDguiod5t9vJ6G6NqJp9tjIWZAAq4P1+D94cpTYqaXx34FTwLbAXTPoLrbXgHAQAamhCyKdPn6A1R9/IRd0NQjaEmpC/I/i0bQtvPhwTQCV8BdSjOu5XRxBDLldKAapDHYDzBe0QQCTKIOBUYywwDaTSqVg5nY4AxFrr5XL5ww8/cM632y0GJI0x0CPB7A5JBRgCc519319fX+NI4suih4eI9J/+p1/9jPTwZ70HaIRDdIj7nPzxyZlrmiBbvr29xbdCZKTR8SqhacRQCGQ/MzQWIW6lRlhBAZH0GMBEKGzBQeEL49Lj2KRrzRjD/7Po9It8AJkNIimmrO/v7/u+f3p6apoGvR08FiQOXiVGC11HF4eok34Gn59SmuThgCEszgSk6UStNbqazjmAyvT6Iu6hNdHPKgn+CCHIT4ltQ7KB00Df9+v1Opl84Y2QltA1uWhOGMPfXSwWmE9GpmSMwfkd9ALkJcn6yRhzGdLRRqkMH8xam+f5brfTWs8Xlzydenr4kCjC8JgivqN5MMbNQrPZDBMGgG/ASgDmqetDCMEkcBqNBgJFy1drHYLXWl8o4L7DJcXZds7hS4UQrLssocPNwsD28XgcxrFt27u7u6Zp7u7u0sMJIIlZfUIIFi7hkfBxGcbT0xNEDdiLIISgF6M9ked5IERlWeIGgXa//vpr4E0AprIsX79+nQb1EdyRBnDjMJfz1Vdf4ZKeTqcsy+DEeXNzc3d3B8Hi9fV1Yh7wsBFC4FoKjl4pJYT03n/48OEPf/gDBHvI5fv9vo77onFtoUDFWABCJ8zdqqp6enpKT6CJNpF4TsZpOhyPXAghJZbhqCzDaCGOuRBiNptBUYbknWjbsiyxtEePU5kXVVFmUikhM6nm9WyzWmdZBmAOGfo0Teh7LRYLtI6Q56SUGGQDjJBS3t/fw3sq8cxoZ2JODXwszjU09KnrZq39/vvv0W4ZhuHm5gbVycvLC3g/hG84GujovInbiqgFJIpT/Pd///eodcCOAihkWbbdbiEKByzGZCsmmV6/fs3i0AmS9NPTk5QSR/7NmzcvLy9QOSK5gtqVUmIyA58KnzBZFqKpppT61a9+Vdc1Hnj+cz2X/qx6+D//4f/CYyfi/iOwKz4ulEZlABQPfnOMGz/ACAE+IIigqkjuBUh9NNpFgCoBzMHRwlfFg4hiHFHPOQepDJ7a6+sbxBH8AEfjTfE6IJGS9gCwHYEs6X+Q/xI2wX9K1CdOEcBsOpCpAw8AhXAJ8JJYV8R35CdIp1A/IpQzzutZPUWrZORUuJ+CDoKsCMmDxf3PJC4hQHowceEJclLi65EdQ5QYAmWYKBHGL1dxxzI4ImT3JD6uqkrrqSgyax1jjBA4aE7e+yIvu77DLeafTbR8Lk4DmMDv6Dh9mg4zHgZKWQh+HEfOWV3XPjjvfde1hDBgDhZNT2ez2Xw+G8deSnU+n61x8HTSkzbGhHBxiwrRHgekEGeMEIKFcSbuxBVCBEKstRhRbJoG4aCs8t3+xVqjzeSc5YI5b7WZOOeEBErJNI1aT7Bat85KKYZhnCa9Wq0IodM0teeWUhLiMEqqz/q+n89noGfTH87nc+ucM8ZZ54JHhpNSZvll5R9iH64AAPU333zTNGcbfQSQCJFfnfN5XmhtKGXYLQwLKSA80FB4YDabDYSPyHDQjCaRGB4AlKeIa09PT4xdbIVCCCgZhRDOe0ooIiZQPxiq16+/OBwOierk0dLx5eUFQ0jomiReNFNZOiA4GkCBKlMIqZDcoFHv47Q/dIxQiyXNayKZcc2llFVRAp8hTKXXub+/R+sUYyXAfHDgOB6PuCB3d3ePj4+Pj4/4zMvlEjPhKE/7vsfH89FtKMlJ8HW+/fZb+KqB9BdCYG0c2AVoQ8BfSXFpvqJGSYovHxxYKUClL774AugQdDQuY0rYIa7axMNMCFmtVk9PT6kNDGtIEVeR/93/+N3PSA9/Vj0keprG0SoZHdttXE5Ao7kQIhfGmlCFAeAnaijpgkS0nsZRwTLxBHhxO/F30bvGJcCwbhKZoUbBvZmmCU82II+L283AeqF3Ci7VfbaEboybNPCOWZYtl0sQvklEBLoZOSDEHxF3C5MIz1MbHJciPUP4OqkeR1zwUexLCUmZBscDIAgFAZh9zjmoYVxtMGN4C+SVEDWO4CuTkhUlCIo/QggyBHobmFJer9c2GiaihAeMYnG7clEU680mMeaMscPhAM2+iTuK0edPLJCKrtf4vlBhAVugtwYmB0cUp87FOU/gaLwv9CQotowxGLf23gFDBB8A3KqqyouCUoqECpyBF5HR5hbkwHa7ff36NT6wkIJElg/6XZQ78FnBO+LZAB2ENkliSp+fn733WabAdpZVeT6fKSXGmP1hv9vtn56eTqcTqni0YfB3MU+Tng1jjHcOsHG/38OJK7HSNm5pBoOMI9DHvT3AYXgSlFJQfKK+SYgBdwRMI1BqKqQwjp609sgc3377LTgxTDsiEpHLosoTIizOPp7naZoCJVwKT8K5azfbq9li7oIXUgAh4a+g64tRfBftuMGXxvSgBOfeueZ0OjeN0ZrAYf4zjWaaH0JPDt86aY2Q++fzOQbNhBCJdQDpjUn4aZru7++B/7DpD21UlOCYgQXUg/q2aZq/+Zu/sdZiKSHSCUoQBHoIdgHFkMgBMUED/tM//dPf/u3fAkzf39+DNNvv93geyrLsuu76+hrtBCQ5rfXd3R0qFR/3xsO3FOz0mzdvEqIF4YFYRwgBlYLT1DTN4+Mjip6Xl5fn5+f379//9re/hant27dvf7Zyif27f0cUTtVlIrLJZ4Z9KdSGKIzBGeBxFNnHXWaA3iBecf/SG6FKCJ/JnNI7gu4XURoLZgm/TAjBo5DkmFnceqGjsQ9wJZIwLmiI8xBg+lJrxEXnapwcay3uPaISjhb7bNUPojAiNSpxsHsyjisTQtILIkYgDqbGNYiF1Dwnsa8FSIUWDgkE3ST8V3vR+HcmWmgMwzAOl7HqvuuNMbP6EtEIIXD4QOwm0RiAUnp1dYU5hqSNI1HRi4vvnJNCOhcYY0M/WmthcYP0s1qtQKfiRqc2aYqkKEfwnFRVhXGT4/H4448/IlTt97u2w5YIYYw5Ho/Hw1FrEwfBOJ4lrMnr+54zMZ8v0LwlhHgPdj7HPCOCDlI7BrxRX4JLwfOGW5ZnudWm77qh74FmkJmmaQzBDUOn9ViWufe2KLLT6VKk4vPg4lhr87xgjDPGqrKklBJCnXN6mpy7bD9E4MAlApmGHiPAE572up5N08QERxMOoQriC4xYgxDDJs71en04HK+vr7E+lsZV5+M4Ziqbzxaf7h+6trfGtedu6EdnL4EPp08p9fHjR0IIxl/O5zM2La9WKwxh1HX93XffJQYSzU8pJXb8pSyOukdJRQixcQwF9OB2u33z5g0k9sBGiNdA0K9fvwZ8SexlCAGOO8A9PoqILgMNUuJ+gQyBGqWqKhilADAB0xwPR6MNHE32ux1n3BpT5DmjzHtfVdVms7m6ulJKwZSl67q6rh8fH+/v79+8eTPEHd1Iq/AV3u/3iOlv374FK7Df70FWA7m+vLwAJeNhTlQBJjSFELvd7unp6Ve/+hWi4ocPH5D7wbuiJQZ0n5guAJ3U5b26urq+vgbYB8yt6/qXv/xltKCvV6sVmLfUW7LR9lVK+fj4CHh0Op2Ox+MPP/zwxz/+EYADR+YvTQ+IjLijInpjgR4JcS0GnmkAJR7HhsEvIb5Lqaz1xjhr/TBM1vrValNVM2stZh0YI84ZVEiMEWtdCNRaP476fO6McbvdwVpPKceeQcaEc0GpXMqMMZFUrVlcaCzi0nMTfZCApHzck5MYLaSQ1GRGHEcSwiOL2IqylEabjQTugKeQVBB9ZPQhSI8159w5i2W33ts8V4g7nFOEUXz+siyllBBXOGOHriPOB+sUF9hREZxnhAbrqA9mnDIhnx8eGaGCcadNkedTP3BCGSHB+UwpEgKUIagqsI0A7Dw+GIZCAfyRVkXcRYr6rO97KZXgWfBcypwQNpst4BFmnWGMIfPhSQDswjOKyPjy8kJidRU84Zwfj0elxDj2gdhArA/amAthqLXt+/FwOE+jDZ5zzpOnEIL7MAyMcRI4o8JZHyUihnOa5zln8nhsIIUVknNBJ90bM/Y9TMePzhkhWJ4rKXlwumtONPjVYtkcT3U5I4TN6gXYCbRYcWfhhoIYBCrDWouq69y0JDDBlVJ5COF8boyZfLDGTIR4a/VsVnXd+Xe/+60QLASHpSN4/JKtSNd3jhBHgsyz9faqns9O50ZrbawextY6Tah/eHg4nU6n0/l0ap0NkCHg7qChRQgJJCyW88Vy3g/dOA1csEmPjF9EB2jYKKWG6EuPQwr8gcIXEBDaecC43W6HZD+fz3OV5VJJxnOV7Z5fMqmqolBCns/n9XqNz4DqZ7ValWVVlvXV1XUIFEdVCNU0rXMhz0usKMiyAqHAuTAOOnjanFp+mUj3IThrRhaIZDxYxwl1xn795VcskO16wynTwzh2/cPHe05omRfOGGuMZLzMc2dsez5frTc0EG9tPwxt35dVJaREYfrw8d5OGkvcnp+fkdVQqOF2J3qgKIrvv//+zZs3GNIcxxEaE/QGwGFC65FIJ6UUVtT84he/mM1m//zP/zzFLSZw2tBa73Y7XNvvvvsOfQ7KAhd00sOkB+eN86aqi9m8AnWBSQUU01in+vr1a5gDseh3AlkQfoFFIa8Q4tOnT0m1OAzDfr//zW9+k5Tif2l6IISgAgBcRbEJsAxeWEezvCwuwyGEIFIjT8L6OCpDNPSXqCGWy+Vf/dUvKeUhUM5lnpdwYTLmAi5AT+GsAlz4KIFNsjAcYwioQMWmkkXHDaZJn4BvlBoVuKAIBzo6pEopYYuE64sS8nA4pFYH6CkS15fyz+wocFcQMRM5xhgjJED4gbwS4toy8lk15uIKT2stidQZvqN3jlKKGhz3Ao/vNE3npnHwXOJcxR1V1hjOmHcXSS5uEyp9QgiMqdHawtdJFyeVL+iIwjdfqQxMEb4XPhJArvcefR3Uv5TS1KHB/6fOEzIfEokQgpDgvSvjVi/UPeAe+77HAwAXDWBPiDqQJ1AgQ9jmL7OKNI1SJw4Q1wdtOkIInBLwLUjwJATANMYY2mlSqWk0m802ywqt7X5/XC7X1nrGuBBis9k45zA7jcoyJn4H0WRSW4aogkOhcDgcfvvb33Zdt16vz+dz0zSr1QrgdBzHtm0ZZz4EIG688sPDAwnBe4egg7OGYss5t9/vIXCc4lJ3lEez2eybb74RQlxdXZk0pUgp5/zm5iZ1/qBHgIACHDe+ft/3u90OdmTb7fbt27ewyhiGAc8eKL7j4eCsddY653BGcFPu7u5Q1lNKIRUhhMBXCmARvRPQpzb65aAK7/uLigFvB143zy8+mGj7Lebzjx8+wFbyl7/8JaJzCKFtW0qIEP8va++1ZNtynWemmX758lXbHwcCBCSCUkuhNg+gh1B06DF101IExVY32dEgdQgqgGP33uVrVdVy08/M1MW3MrHB6Js+UF3sqFNnmTlzZg7zj3/8I+q7jpBUa93UdQA/oyhab9YkB2/fvlVKIbO4Xq9JBHEPnJ1AK2JMAE2Czrkvv/ySJY2iCKV3+EtxHJ+dnb148cJ5lq1SKs9zTgfzHP/mb/5ms9mEiYHkRlSVJ5MJMJTwA/4ACRhxipWAvhgCcSnljz/+GGBD+CaQp09OTpxzp6eneKwsy6bTKbcTjADH84cffuAB/YSfPypN/19/+18H3zyyXq/BndlbzvN2sICAAJ2fWAlJDtx8NBpzqxjN2WyGDNZsNn3x4gWkw0AWKori6emZyk+I49h2dJbBpuecg5BCoRNegQ4EEPJA57vk5SdagcYYKBAsOuUKBEeFVwLgF0wA+SkmDGcweJU6rkEIEfquQ90MegM31XWdlCJNU0BP4YftSKWRryNu4oANw5DGifEkwpAPcZzkXgBxzz0A0abAZXybBTalaRrr67S4B54jS0cxCTZB7cclsXRsJr43SRIh9uwatqz2QxCD95JerIZvh3xMOoUXtNYmSZrnWVVVeZ7leV7V5Xa7XSxmcZwaYyktaK/Fkud5nmcUS8E9Dg4OQggMoYi9gUxp07Roq6RpMp1O1qtV0zbT6bip9/2u2CZEt8bjcVNXSurBmDhOur6v62YynbRN23Y1FCPgjul0CqBR181isXh+fj4+PkZX2TkXJnfSYwXWh8snkQrlLpzWfD6nkoQyGOsshAjlKH4BHzg7O43j6Pr6Bonyuq7TFIVdkWXpsJ9e1adpCrMFfP/FixdlWfpKteHsHB0dwSdGuZp9TsbsPE+aF4MysX/AxxA9TNM08TxadgWePs0yY81qtfriiy8CjzOKojTdEzo5MlEUwdIBmQlKNhz50WjUNS2XOgw9tLQ4juM4mk7njO+mhMDmXK1WMCSBYsBvueXgMqfT6fv377/44gs0/t68fbtDgB1a4GCaphmMwVIR4XFhxBDOuc1mA5UjjuPlcvnZZ5/d3NwQu+DkcAMUOebz+fX1NeEd8zmQVgw5GdhRsP7ODxGCUgVh5Pj4mF0UOE4Yz/l8TkLQefE6jgmxWuYnd0GyIN4qiuLq6oookA1JnoHTYjG7rvvnv/ziJ7iHf6rYyvHG7bDjORWJH30eqgUYSjIp6+cfhcI6VQosFGXYohhrHY9Gk743eT4SQkmpq6ph+/JegkECCr4XQrTyyqO4K54ZxJvAJSC4QJzE+OZttnv4S3AA6/WapC/A9JxVsjDtpWykb8+Rnr4NukcQFzhCENecc2RO1hohBObgU7/lnCOaCA4DBh4mki8NhR82BPeOVwCpxLjzXSRwwA7aM+24YN8/ZTElnCXseADf+Fho5tAcUQ7AjzpPQsUWYFP4Liqfzrnb29vb21uSdG6Wux6GQSvNmA2gVSrVURSH80lQSdWRMA1AfDQaLZdLsC9YgNwavRQ8iCRJ2SFt20KzQdErhGDsT9gmSLBgpPAHwzAURT4qJqvnTZGPkzibzw7qqk3i7PWrtwCqWuvb21sgmnDyWQoqxuAwlKCAiSl1suwI4hLvQ+oNAT4K4a2XuxdC3D/cJwnj6mrKEvjjgC+j+tA0Df0xeZ5fXV1xXp6fn+kmIwKgFY6kh4wZYgLpDkacA46g9+PjI31/8I95gojihYT75uaG3PHly5dKqcVigYoG0Cgzi7gRDin1j0DiIIdm2Y0x3CAtDrTxB1yX3zkRtRe7fv/+fSglKt/3yoIfHh5mXlD5w4cPASkl3mfBqQ1w2QcHB9QScj/fgoCDOPjx8fHu7u7p6enrr7/maLNuwelut1ueEXr1dV1DcLi7u9tsNohQUWsEINozI6Ko7/vb29sPHz70fT+fz9u2JXXDMB4dHSnfWEqvHyz2UIpDBRaNZ+3nwQRKAjsKu0eYwjAFZi5hFbHVP+Hnn7bFCT8KiuRXCAHRE+I599M0De7ROUfHTSCQ+BrAvumXdmhGUXKQNpsNknC8nUWPfbMMu0oIQTTX+ulAfBqmHIvcdd1isbi9vcXTCg+CCSGUFwENcJD0jONwPHo/NYgDHwoqWGThyzB4bJR7KUORJQRFcTxl4lvthBCTyWS1eja+sR4D5/YjNHrt54aGVEwIIXxrZbAXLBcu7dPbwQEMnrstvNiJc85Yy8gLEI9hGGIvTil8LzqJEXz/tm3D8DiipHbfDB8bY0i0uSrOKmsSemKhKm63W7Jm7EJQHcjzou0arXVZ7mazWV1XSqm6rrJsJOW+0zsER5PJhBASRI7jdHFxQfBIVAW9BNKOMTaJ067vWA3r7Hq9ms2mbdO2TRtommyq1Wp1enLcNt16s4nj5Pj4WOtIR9FkPHXOcrSwj9Z3NSJQwRXO5/PlcslNBYFI2m44MqArZDmIGbBQHA1GBWBTsB3CU4yMMaw/BQOlRNt2dV0fHR2v12uQUedE17U4+LIs37x5QxDN181ms+Pj49///vccVcwK8E4URWQtWNvQx8OFhRiO1wxetZczst1u+7a11pJ/IIV5cnJinYuTBCkteDVcVdf1uEzADbJ/7VUsY9/HF2K4w8UBJ10IV5ZlFOmPHz8WRX58fGqMeXh4wJfAcQ81YbDrg4MDmKlcPEkYBuTh4eHLL7/c7XZpliqpVqtVked1XSdx/O7du9v7u7IsaZvHOU2nU+rAGAc+E2cQqOfYFhhHuR/pDMwAmAmf6vT09Orqipj15OTk7u4OS3Vzc8NoUkhKRGy03NL0AIJH7EUz42q1Aqunrgk4D3SPtaSNMSDbGIfFYoFNDv1GVIYIksDh/wcQWyMv609svn+FUr2fQ0td61PzxOYWnvUPO5MIlL+HBqg4SqXQwilnJb8Ip6Tvr2N9AAAgAElEQVTQPAA+h0iHU8rthQwGryB8ahnsJpk7oEdAS8AiQhUEgIhTbYwhCmB9+QuPWXjldOkbIIhT2DrcER4llEYIfAhbxCf1CUwtCxgwoizPwgg8rFi4ixBTh880nvWk/TRmltcb8Sj1CiVcbSg88oNHCb5hGIaLiwtOnfCZDV6EPYpdWK1WJAfcMjs+uH9CZihJhKUwIgAMx+MxgwzR5lSeGAZszV/ieN+N75wL2bHxg2aJB4Hy4NcPw3B5eYnnTpIExxb6QuDg8lA2mzWPJvZd9FywUgrCHyaSMh0faIydzRZV1RwfnyZJRu0hivZJ82g0IhXAhoJwksMFCCvPc8YVnJ2d0ZklhKDPgIzw6OjIWjsejzkXQghsHPk+G2AYhjRJQlj99PQ0nU7v7u7Y82R7/I4px0OMx+NvvvmmrutXr169f/+e7PPk5OT+/v7g4OD+/j5gd1hAQjcCXh4xAQEcM5pUptMpQyBoMcuyDIcNq4riXJiXALVvsVjEcUwRHoSK/IYFx27wg3uIogg3L6WEP0qZ6v3793TOgtg8Pz8TSlPrenx8hByllAL6D9VQ7Kn1EqJ7kdfRiGdd+Mld5+fnfd/f3NxIKUNjAUTV09NT6wv1zDnv+/76+joEwWVZ3tzc8KxxSABuJFg//vjjwcHBu3fv4jj+9ttvq6oCfGMnsDLDMKDm9PT0VPuhhz/++ONnn30GcBd5iYfYD97A/4lP5gpDuiErpd6A+SXYBe4ehoE0QnrRTzzEw8PDT/AN/9Q9CA8pQGTWnlOMiSQnbf2E5yiKOGyZl8X38ZcjygOzU5/oCIWMgS9Svqur8124Usrei+kLr36KmcPNaN/+h3UgNsEW9H6iHBaz8/KxoYqAbWWTUaeynmSJZcGaBLuJpcbnYx8bP86az7R+/h1XHpJfpuJAHs28yikrExIOvCzXE9yY81zMACxg77D1fI7yY1AjL2yOD+AeQ7tDSMV4GQIDpPBc0m63I2nAm2Ks7SeUZfw9KQtezXm9JuOVCYQQSZLAbSOMovOOeVVE8SDOflXNwcHByckJOSXgDE8ZVATf1rbtcrkEGWDTA+DsJSviOIo09SFKU3meI7UCnx2sI4Cc7BOACF5PeGX80C7lex6J0eioCLoL4O/KK7ejvs7tcCyBs+i5Bb/CBvGWyWQSeJ/OS2ezN4LfNdYWxYgXN01DKR62LmA3ATu2JuCTTdNcXl4eHh7iAp1zSJAtFovlcvndd9/xLHIvhdn5AfJwWpj92fq+0ZubGyJ3epuFEJ9//jk4OGWJvu+xbvgVUpYsy6gqA4AExCPwUEFaAB5YH8wLSQYzhfAx9CjQJYCtv7u7C9Q4ZktAr2AGFA6ViqD140NWqxXtDi9fviSaruv68vIytF4DORJcp2l6dHSEJ+77/uHh4fT0lA4tUJ3j4+PxeHxwcFCW5fX1dRRFb9++ZY7CMAwoIltr//qv/1oIMZ1OMz+E/LPPPlNeuY+NfXBwcHZ29ubNG/RC+EweLutJ5U95XgnwINqaxpjT01OYxxgKTjE7GQAzyzKOMNULj8EmFDgDV/P/788fgUv/73/9XTEq4jiOk1hHUZZlOtJD33PRudd0FEIQ/vR9T6xEnYdjOZ1OtY7YH5wBDphWkZTq6en58XHvUX1lta78pCqiAETiJpMJqwbk4ry4ExtCeKFTwmH9xxRbggjrhVE5FVLKsiwhYNhP+iTI8UNgjnG0vqHPeqwTk8debH3fA/g4UT/nYTQaxXEElQK/BfokhEjTDP54mqRSKtMPTVNHOiJP506Vn7gZcDCuDfdsPBudW279QBUSmijS42IURXEcRWVZCeusMcIJJeXJ8clmvVZSOuuyNANcJgxvfY/hbDYry1LriICX0B54ZDKZOGesNUppKPxgxyxU3/esoVLq4eHBGFM3NQteVeXR0ZFz9vHxUQgpnBiNxo2XW8CaTCYTa521hnnIQggiO0Dt5+dnIQQexduXPSVBCCeE6Ie+rqvxeJzESd/1wjd2IQHZtm2s4zRNhZJ13RhrnRNSqdF4tN1spZR5XhhjRkXRNq1Usq7rJImFEAhRAEjygLhmvCkhQtd16PPgWREiDCHFYjbXSu+2u1cvX9VVrbV21k0mk816LYXM0rTretZQCmmsubu97bouTRMnbJqkcZyUu9I5m3k5MgiUgBtCCJKVx8fH46Oj5cMyzzIdaXb1bDa7vb1lt0spaQIIWSkBGZ+D0ZlMJviefUF7tcIPhcBls9ksDg6MR2WRWiHJm88PlsvlZDLBNAP/sm6hXj3spdtarVVd10VeNE3T9W2SJG1Ljl50Xf/ZZ59prf/+7//+zZs3z8/P1k+f1FEklRoVRZZlQG3CcyuMMVdXV0SoZVl++dVXN9fXHApwGGHdbrebLeb2k74/KeXz8zMeDlIledJqtbq4uOi67he/+MX9/f1isdBaE6pzRs7Pz0mGHh8fDw8PgfLAdvjGzWZzeHhI+EXYgXjU6elp60WACJ2x3RBYQt3l4uKCbA+ThQ3Bzx0fHxM3M9gKOwA29SngAQqKzaGRcDQaNW37v/2bX/+p7uFv/u63xWg0GCOUFFIkWeqE25Y70w9JkpATCY8ekLng//FOJKTGDEK4rmujSEeRLsud1qoo8jiJL68+brfrx8eHNEucsMb0dt+X1MZxFMeRc1Zrjr2bzeZYH0pVgOZa6zjWSRI7Z6UUTVMje2CtEcINQz8MfRTpNGXbKSlFnmfGDFVV8p+Ir3UdlQarlJ5MJlrFUqi6bru6tsY4Op+VklLGaRLFcd922JqQ6LBHg5x6KEjGcSylGgYTRXEUxW3bxXHSdb0QUkk5KgpF3V7Koe/LXam1Hoauaeq2bYoij+NIa5UkcZLEXddjcUg28QfQRfq+n81m5GHEd13XjYrCWiuF0FoL57RSUkppnTN2MhrbYRgXhTNGWCudq5sGEhfnnxwFDlSaJn3fzWbTrmvX61WaJv3QrDdPRVFI5YqiSNOkLKuhHw4PD6NItW1jren7zjnLssNk3e22w76JOuu7IY6Tum6JJTHfIahJ4rwsK6V0Xdej0bjrWgwZiR3gMpmicy6Oo/F4FOymcyKJU6V0lhdt3ysdCam2ZbnZ7SbTmYqiSEdSSKl0vRfE1k7YYehXq2elpFJCa5kksZAujiOq74HOQMyLO6QYFiquys/HRvWTeE35ASGrx6ejw0NnbVWWhwcHztoWDbuu67u+rqpI6ySOszSty6quKtO546OD+9uboW/fvn5lh6HaVUdHR0JK7DiBAuWNs7PT3W4bR7rvuixJnDWm7/q21Vr1Qz+ZjFer548fP/R9N5/PLi4uutbMZ/OqrIoi3263m81qPB5JKeI46ro2SeLRqMjz7Pb2Jk0TpWTs+3to9+PgG2Omsyl+ZbVa/fjjj09PT9bah4clV7hcLoF/Sb/SJHMOd2KkFE1bKi37vrZO9Kbv+66ua2dt1/WjrCjyYjqbCSGSJPn2229DxGOsHU3G/TBIJYWScZrUbRMlcajGkWGDwRpj3r55lSaxFK7vO8xI1VRCis1mO5lMCDs630CKaQ4VR6JsThZBJGcN/9H7GVCMWCCoR4a273vKy1LKy8tLuLOj0SiOUmNspOODxeHtzd3J8WnbdlEUW2tQmglUWs7gxw9XX331M2Ns1/VSqEjH6/Umz3InXGjNwQbClo68ZCrRNhlzvSuLPC/GozRNkzSN49g4O5vP/uLP/2TmEt4sgP44fyX2mhYkQThP4aXMKbgfHR3FcYzDD6hLKBIQCdJIsVwuCYH5Fyw7TdPGDxFipT4lQjgvfx15+jm7kMxOeMX2ULjvveonD7jx46DxcPwljmMQj/iTgUVaqwDvUM8JsAMuAT+vvWoTsRInlqyFi0EoLXwUmHWe54BX5C68N6DYzrfmKT9JNPIC8dKLxSpPGw3Nemws42V7WXww2clkEuALJwQ3W/uG1ThO+n7A0pFDXF1dBXyJfMI747jdd35aophhGBaLxeHhIZXqw8PDQBmy1qKxHPToHVx1KUF4qWYrpUajUWhMZeTA4AfSAdoAbfPcB9+IRPnu6emJ+kTmZwvCp8JGsJ4XFxexHyquo4hbkFJyLHvfogE81Xt1ps1mw5Xj+K3nfYE7kS4TwREqUnKkkJOm6cuXLxldTkj+8PCAUnQo0rJKQOQUV9p2X0vfbrdlWQUaBRt4Op0eHR1RW9ZaT6fTh4cH5gporauqllIaa7Mse/PmzXq9fnx6AmXGXgAivX37hjAZH3NxcYEEKfEW8MX9/X3qZ7gCuDNRnKIRE7e22+319TVzRFiErutWqxUkHNIptsdqtcKV4jJ7r7zZewH89XodwKjU6w1z1rgXqHSUA6lPcDu+zrSBFNv3PQllCLf/w3/4D0xXpDoVe+mLYRgeHh6m0yk5pZQSRJ5PA58hCyQUuL+/Z/Rb4B+Tqz0/P/PhKIpzkWQeQgj8inPu22+/Bc1OkgSIbDqdBjL98/Mzh4umbmqfNMZfXl5aa9++fQuxBRUZrAcpGkDiMAyvXr2ixlCW5Xq9jr3kHeodmIvOa5ti5f5U99C3Xde0wjqGpg5db/q9aim2++joiIHgrDtWte/709NTDir4OPdjjEFyBJgbwQAhRFmWq9UKL8LVSykXiwVIYqjT9r5/BGJMsO8E6avVar1e44FTP6+NjYLXyT5R33RehggjHuwyZl37ecLj8SRNE2SFsNeRF15WXkJ1GP6wIOwk6bsfeEuAhkOVG1YJ5TJMedCvDYV9sR8+Yz2pQwRMDEMJhmOMIejmmsHQsH2UrcDx7CcqIHXXOimcklGaGGeNcEa48XSSpan5RC4Jg3tycoJzBTcAIoPDKqWIokg40batta4oiizPO6/vHcI3TD/2lHNF7E+ABvM4jmOQK6KhrutoK+29MHjsRcK7rpNeLpQnhfluvU5R5EWEQJmNF/zgICHQyxtjP8IWx0k9gNe3vgNWShnkhni4yms1sjlj39tI4AJusNlsPn78eHl5CY8OqA0Lm+f52dnZP/zDP7x79y7yDQp5nqMB17YtABqPe7fbvXr1mpKAECJNUqB/ag/Kz/bAqTN+IE3Trm2FEFVZHRwcdF13dHT0/fffZ1n27t07KeXd3V1d16E1mk694+Nj59UgWLrlcvntt9+CEDqvMNh13Q8//ABzKRCicNVkVCwyN0IfBoFX55tbuV+tNf3ehJLW2ufnZ55X7TUBoygCbb+/v+fKV6sVbMnZbAZRiqONBb+5uWmaBrYuJD0KLe/evfvw4QMltNwrihNAQK3GkYMaPT4+IlhL8YySKtESp/KXv/wlcA0nmqIF5kv64Suj0Yhx9ziGzz77DEHJ1WpFi8xisUCC21o7Go2YzsQ6MOqHCAwTf3t7m/lZ8djM3XabZRksVaKxOI4/fPjA0hk/BPubb74hCuEg4/KTJHl8fMTl/zT38E+mxX0XRZF1VjjRD/3QD23Xdl2nlQ6nQkoJDxojhbV69+7d1dUVUGY4in3f39/fg8ElSWLtfoBUXdcYCyLo0agI5xBgDqHjohg5LwAVrHAURdYa4krl1UZ5WfBMbKMgvxXQfEga7LPRqGDi7jCY6XRqjAWkG/r+/PxMx5GUMklT51zTtsYYZyyPraoqhPxCQRXzRABL0sPBkL6FR/jAX0pJAvH09MQ5pOcrz7OgVRnioCRJ6rohntVeQ5j+I2rL1lqgocjLOkkp5/O5+kQUq/MzQ7SfBoynEVIKJZEPwTjmeT6ZTE5PT+lCwNOUZUkGORoXztkkScuyqus9MSHaq5houKEcxdvbWzwT7xVCUPrjITZNC/5GMzDGRWtd7qo8z5qmEVKsVqvJZBx463ggnD09wM6TGqgYa61hm2A6cS0URbGnbdOYYbDOUVydz+f90GPaoJHwOdw4wXL8Sf92XdfT6ZSKMUvqnDs+PiYAxGfQPcdz2de9+4Hf4cX3XtG2LEuyK5KYMISgLKsk0Z9//u75+WkYhnfvPru5vqub9uj46Pn5GW9U+7FCcRwdHR0+Pz1ppRn+vNtttdJWuO1uy6OE0UdpdDZbRJHebrfb7aZt28lkHMTaWAHi6KqqTk5OmqbZbjYQWIk5ENdLkqQfBiqCkddml1L2/UCEd35+Tg2SJTWDSdJ9ScwYE8W6qqq+78bj6Xq91kqVZWmGQQghEdAdjzCst7e3AZFHOTyUYfEEURT1Xc+hvru7Izomg2maervdfv7551SGZrMZbWjDYHiaVI8ZVUQsMh6Pr6+vMQ69H6FIPv3q1aurqyvQQuZO0v0upQSohKWqtQYs9al5TOPOw/0SQ9R72WmEZ1arZ+dF/WDfYQqqsqZ2y+00XoccIgYB8dnZGbbl9vYWoXisXFEUd3d3h4eHjDpfrdfDMKzW6/l8vtluq6r6X/71P/9T3cPfff17a4wUwhqjpERY0ZoQ1Nowpo5EBmYOpIKAPFDBZkXg/xIpN01LAwGHFnOmtc6yNLQcE9r70nGP5Y28HichpLX7oSgUJDDBvR9vF2IrmukxrEII6kvgLWdnZxcX530/sLnzPI/0Pj3qu3Y6nVjhMElt20ZxpLWeT2cwZzCC0EUIvYmUWz+5oes62uLxZ1B1ne8uxM8Fy9XvhzFIvgtnExogh8GE3RMSc6riJPiw1rD42k/BxQHzLTg56F8AKZPJpGH04HiMqZJeWZY4CyNIQZjH1HXddDbp+y6K4iRJ6mrPpPZZ1168xPqp7m3bjsdjBrWSOGqtEWEuy73XYf4SaUEcx2VZJ0m82+26vqNcia8FeSC5JFEwxmCwOCQshTGGLIGAEZIfzyXLMmeslmq1Xltrm7ZN01RHEaeaoI9VDcENmRCegMcEXx7X3vd9kiSEn2VZomCx2+3u7u6SJCEwl1I+3N1zNJgBAAdJSvn4+Egcg9YAyzifz7fbXdc379693e62Qojjo+Oyqp0Tu7JUnwx94sOXy+Xp6cn97R3I4eNyaY1pmqYbeh1FSHsS4EspoyieTKZE6x8vP1prh2HP8yYckV7c3hgzn89fvHhRlSWxNlx2cFqt9Wg8hr4R0FqeaewnYimloJOWZTmfzYUUbdsK4eI4HkzvnOv7rijGaZo2da2UKne7ruumwLN+bhJ1Do6A1toJR6oEjQLr7KyzxuR+FAfxaJZldV2RlZ6cnDCfue/75XJZ1/t5XOQHo9EIgUJOMQkcKAVwIl7k/v4ekAeTQlC43W5B2IC2jo6OhBCY+91ud39/z2NK0zTLcrxL78U0J5PJhw8fpBIkjsMwkEhh6CbjvSoGvQFCiK7rNps1y3h+fg6u9bOf/QyTAn+p73uS7+l0enNzMxmNJ5PJ0+p5Mpm0XWetnS8WZVn+r//6L/5k9/BffwdrxzmhpFRKxUmCVgFGiqw8lKcweXVdo+tJCBx5hXHn3HK5xLgLIZiRhNwgXppn45x1XrEDO8gJ3253ZLshwsIzGzNgjCAVsFKwlZSfkcAnwBm3Xg0Y12WMOTg4ODhY7HZlmqZK6aIo8qxYr9fGWOHsfD7v+r7rujTL2rY11ggh8nQ/CBockxMIz0/76XhCCOKOUKmGetT6Hm8AXNImjq7es8UTqi/AQZQWqqpqmpZyDjtG+Hbu3A91YQMFW8ZDJHZjifaUJ2fjOHZCGGMm06m1Vkc6SVPEW6SULFTqe+Zp95/NZs5rOfR9Nwx923ZxFD8/r0ajcdu2WmljDY9P+cG8PPFQdgb3e3x8pOtivd4Qe3KDROtd1w29EQLWbFsUxf39Xei8IQXBiIfgnQAi9UO52ZBMjKHIETDDKIqKPLfGSM9PjaLoxA9+ATYh+xFC4NQxOpAv2VRAK+SXvB5oZb1e0wuG1hPd0SD4e3vtsekAhXO19DmzaM658XhclXXft+PxKEmTx8fHohhVVVNWddu1uDHWCvPhnL2/v+v7riqrNEm6rjP9MAxDmmdplnZd9+rVK4oiaZoqJbu2b5pmt9vVTa21rqry4OCAtB6DBUoWRdHNzc3r16/TOEGKUfrJvlD4nNijsoF965Gobp98DwNn8OTkROuImmrfd1mWCYlmgWrqrigKa0zsR5QnUUxqwjUbY6j2O+ecEEi4o35BAUApFUeRcK7zw+MoIcRxnGVpKKcFISkp5W5XGq+3hoITQRgoFruUkiHFG5Rd9gp6UnLKyC+tp2/hIWgIIyrVvp1lGIb5fJ6lOexkdgXb/urqyjpDWBPSayDcSO87OvmW3W633W67vrfWdF23WCw+//xzugLxZ8arPkOggAW+XW+KokjSFAtT13UxKqy1/+Zf/upPdQ9ff/2NcE4KoaSE+qKlssboeD9naj6fw/ZjaTBwGFyt9WazwTFwxYlXSSL0YIw77o5dhXFMkn0iH0BPaCrDsB9ymXgNOw8x7QciAjiA2ECmCoxP6F+fFnUjL5nLmh4fHwEo4XeiKO77fjabdm0zHo+lViChfd874ay1SfQHcQ4MNNVFoE9WY/AqEQH86b1yEQYUfBkM0X6imp7nGc5PefFFvZ+JWHLLLFpAqLgw7o4ICNSLeww5MsUu51ySpoMxQgpjTZZnVV2TU9P8DBYhhIjjGGxqNBoF5ivVOVg9VVUXo0JKxcyZPINQv88bOO2hBAKIj7TkcrmkPgEXKyBveKO2bdumE8IppfqBmdiSWAQvCBGQQh/MgmEYkGDDu5DLC7Gf5B7HMQ2xoO1JHCdxnOYZXqTv+/FkL9VJOIl8gvTUQMwcexWlNmB6Wlg/DZzRUTg4OMATPDw8YHOFEMazDEmkMDSowBKYc3lkfkqp3bbMi3S32/6Lf/GXl5eXeZYnSbYrq9LzpkJ9a/ADcOqqXq1W1rmiKMrdTmu92W3jJFFeaYaI8uDgYLl8Xq9Xm82m7ztrbVHkGGJr7Zs3byhfrVYrRKfH4/HJ8fEPP/xAxBa4W06Isqqcn9++XC6B+5WKWHbMnxDC7VnpXZLEwzC0bZOmaZLEaHA5K4UQiRc6jKIojZM0TVW0J7tjT6SfeCOVwkYTVqJ1aAaDBAPGlGhsu91OJmPENLuuOzw8xJG3bdu2e326cChCrDAajR4eHtI0nc/nbCp4NM5r8MS+Ezb0eEVRBJeJmHg2m0WfzC8hUsnzvKn3RUGl1O3tbRzH5+fnzrmH5T0dDFy/9LyS1WqdpmmYS5im6f39vRCuaZsoiu7u7l6+fDmfzy8vL6nRshOiKKJexUjR9WqltZZ6r8QzGo3ars3z/F/+xZ/9BPfwx+OAIi2llEo6IQZjrLWYFTCTgOREviGL3wNKGwI67eckKz/1UO3nktq6rqJIw4Ds9kNOuqqqu65HbpqHzfmhARIaL5QDY4zQygkhlcp8v4/zw+gBgomIg5an9sOVgueo6zpLR0mSt+0gpYqjTO37jdPReNIPPXu9KkszDNbYvu2EEKhBuH2vRhNFOklipWRZ7jabjTGD1iqK9qZNek1APCiGCX/OsnCQSK1a32kY1BooFQAsglzhlnCKgaHEDQK/4Dn4hPD51PBNPwxdr4Qsd7ve9xK3XlQjwIbsft5IcxD+ZrPZOCuVipXSZrBJklRVWZblrtz1/X4W8eA1emFhsdoYKSSGHh8fpZRM1pBSTibjoshB1bIsjWKllLDOwEcktnD7VrUuSdIoihF4oMJPWi29miy4OWREJFuAMfnp+34wRnn673a7vbu5TaK4rqpI6aHvzTDEUWSN0UpHOjLW6kjPFwtstxRi9fxcbnd929l+SOJYCLHbbo0xr1694mHB15jP51rr29vbqqqiOB5PJnEcR3F0fHws1f65z+ezPE9hfj8+Lp0zWZY8PT02XaPjdLPb3d0v3372ed3148nEOjsaj6I4Yt71ar2SSiqldKTH47EVe0Xhqq6TLLNSkPYhEhe6NWez+cnJsbOm3O36ttuu12mc5Fk2KgprzI8//JAmyXa7aZumLHeTyfjqw8dhGKAPsZH2CthdZwezXW8e7h/SNH318qWzTgix2ayFcHG87+mldWaz3rRt45xN/KBWBobHcaa0zvI8HxXFaERDRlXXgx8+j5phiB2zNC3SbD6ZmmFI4qRr29FoNJvNyDLZpYweAui+u7tTSlVVOR6PLi8vh2H48ssviqKYTMbz+Ww+nx0eHjRNzXGr65piAAy0h4eHrutev359fHycpZk11lknhGSUt3Pi/v6h9tLFyDcRwVxeXhLHEO5gG1er1dn5GXSZtmuds9999+0//uNviTCgnymlUNPy/f9dP/RRHD2vnq+vryDmAoeQVP3+979HquDm5oZwebvdGj8R6JtvvrHWnp6d3t3fMUG93O3yPJPOdf9D2uL+7h++abtOad0PfT/0UishRdO2JLYk+03TQBmCaCi8ClDXdRAt+KggDBKqiNgFyv2+tiyds0IoKVVZVgzkGoZOKZkXmRlcQFoGPw+k63uhnBXOGqMjnWdZ17bOWKUUYRSXqpSChAPmS2jJPqDk++rV26qsdruqyAv0bZRSo1HRdo3SGnyAGG3o+6qq0qQo8kI4kSQx3T2z2fTpaan37XhKSqmU5I60l0Mhc8JjUZzA8+FKnXNee6Cj5gwkopTK9mLxPRi98OxA4ccoYeLZNGyU0KsVYhyqc7v1ZlwU/TD0XRdFsRRS+xGGzjnKD9rTLoWfwgSSQ3mfWVpaaRSNgF+HoYepEPtRSMQylJ3BgjabDYEV+1tKOZtN4zhu2ybLsvF41Pe91spa23VNmibGDMbsx6ZS5imK8cPDcjQax3G8WBwIIfO82O3Ktm3iOA76d03TQAgB4qAEgnhynufCubbroCkPtLszHrIsjTFmGCiwMVImTdPODF3fH5+cDH3fNW3f97PptGua+cTr/jqXJMl6vU7SFFSh90P6tJ8RG8XRYAYdRU4IJ0UUx9vdrqyrNImqqpRSZFlqzBDH0WQy7oe+7fvNdjObL66ubpIk0zqSSlZN3Q3dfDGv6irN0rIqJ9OJjnUcRc7aLMvWm534ZasAACAASURBVI2xxlg7nc+U1nmRs+XoLAHHL4rc9l1TV3mWNk1d5Hkcx+PRKEuz1fNTU9flbkvLpOl7OwxluRuPJ845EinlVaabuo6ksta2TRNFUZqk88V8s9k4Z52zeZ5Jqfq+r6smz3MhoLM4KUXgUsdxMplMe2uatjk8OnK+g9UJ13kp2f13Nc3bt2+11mYYYr2fTSKEyLO8quqjo6PS6zWwrwJcKYSLIg2cS8fVbDa31pRlmSQxwINSylpDc0/TtBSNnHO09UkpX7x4sV5v5/NF1/V5XmilT45PtI7W6xXhF7wjBmRJL2LNVlyv10x/M8aU5e7Pfv5nXddaa5xwURzRnlWW5Xa7PT09PTk5oXcEzk6cxGW5I3TQkTZmOD457rqWdDboCywWi0C+oEEaNIJgtOtbpRkeLJfLB63UeDQyw/Cv/3Rw6f/5zT8GVp/w/QQEsORZdV3DBCAiJhyDG0elhfA8YETaTyzA1vCvEIKQGVNYVTWklKLIrbX49jiOu26P41svK1RV1bbcRXEU0bOTJJPxZLfdPT0+aq1r37gI2kP1mLw+9j1f3j3Ur1+9paJIFZfjPRqNmqZu25ZGDeNVj6qqSpJsPB61bWvsYK0Two3H47puum4PKwEvxvu5x3pPEBJC+AIvUQCmhLCOUlUcx9aawWtiU5Zg5/WDcc4pqLp9n+d56+UwA9qm/ThuspPeD3Clbr9arZqqLopCqD/o3YpPxHR5ZKlXEwmFjYCbYfLIZOktCLkCqXRIg6wfC+ycwzGESj4bCUFWeAR93xPswwABcSLVAx0GUVwsDqjWRFFE3ShN0+VySfPtYrGgEmb8uM2joyM+gXo70EG520GnwXMA1lG0gH8ZvLW1Ns2yKIm6rmvbdjIaW2Ofnp5GRaG17rs+SWJjbdM0TdtqHW13WxYw9qNMqUOMx2OAJmBuFERgYTnhhJBN285ms7btsixv29Y6Fyfp3d0d06HhcSqlBN02VTWdTmjWo+wPBAEAHaTroO4kSbLb7ZhUAVn+5OTk/Ozs9va273sQP6IHIYQxQ9/3VEeF75YYhuH5eRXwydlsRp9alqZSiGEY4iQxxpyfn8/n89JrFJ6enlrrDg4OVs+QPidt19R1ZbysZEgHs2IvQGCNkQK04g9IIxW12WxGW0lVVVrpOImlUgh0N009m8+bei/wzmmiBJhlWdPUVVXPZjMUmJ6ennsvrgxZSyn18uXLp6enaD/uzHDegZUAzaaTqZT76dbWE8qJgepmzyBfrVaHh4dUqj4t76dpCk4lhKBzLU3T6+tr7CFPimDOWvv69WsMVO9H87J5KOVCtQh029vbW+Yro08RSJtU7Fnkvu8nkzGBJmOsUJXP8/xf/voXP8E9/BG4NHjKPAS1YG74AftmdwZmDlG28pzXyI+0DWUDmnqQ0wmZBCHqp8VkGCkshNtP8nHEAjRIE1qiV9f3g9baGNu2rUAIqN8rLNG5LYTAdHJglOeMUxftuu7u/ubu7vbp6ZHp81Gk27bZ7baUJR8eHniEhNhd1202m9VqJZXUWoOoEmvUfgKP8zJN/uAZoCScJUKHvDL8r1DJkH42OmEIA89YkyzLpFJd3ydJQi83K89zgWwTmgOER6Ww1DSvSikfHh5CNwZuEv/BV0s/+Sf2Y0fFPpcaaa2LooBqDeGYah6yPHx7qMNTdKEumnkhOeHHa7O7MBZh/IbzxOXeDzxBjgaRTuao0EAXkDrn3GhUECPDMiRjOD8/V75dBgfA8HdoTkQ8oM+BVApsDYwJcMz65EWBZ4XSWhTFcrmkrial4sGt1+vdbkvz4HK5pMhP3Zhtxum4v7+31p6dnQkhkGhO0+zs/HwYjNax1tF6s1mtNszAoL+BkP/s7Gwvo1SWm/XaGTt0fRonQ9cnUZz6OU6hQbL3vF7O183NzenpKXjL1dXVdruZTiejUXF0dGitGYa+qsoPH94bP8kcnxoeWdCPYcEhdCZpGkVRUYywFXVdk1aynYiakyShcNV7/eawJ4Ef2OR5nuM7pW+4ib1OBiogHGEppbX7jmJyBSFEmqY3NzdU4wgdAkNBCFEUoyRJd7tyNpv3/TAaja+vbyCIU5Ojo5MhJYQUzPmIoujo6AidjPcf3vM0oS0554iJJ5PpxfkL4aRzomv77779frvZDcOersIUT7x1lmU0cv3+978HBAN6ybwamFLq/v7+w4cPFxcXwdZL3ydkrUXyj7AGcRRSZOzD5eXl3d0dt89nAr1aa29ubjBZT09PFxcX0+n08fHxJ0vy/VH28F/+5u/xgfQuwcwLaQGybiGXIVrZ7XaxH8tDxCq9HE0opbLnSDW4W1ByMITdrvSBA2UMaJdF3w9KyfF4bC0WtW+auu3bfhistUpKpaQQoqmbHhlRa+I4Pj09pe2TvDjI/3EZMEYeHx+3213T1P3Qd11bVVVR5GVVRnEEcsJjwMfEcbxareIoUUplWSokdZQe+YG6bqwXdCLLAY6Ucu93lddlI/Tm/yov/Cf32m2y8Rqf+5AfFSZj4ySx1vZ9p7SWSmmthBB917GTIt951PtZvsEjCiFosrWDqetaKGmtNWjHCocP4MqpN+xxGCEi3yFs96Pcel+B/MMPzMUsy0iK4XUQRA++F8H58dGEAtTQsBQcVMp9FMCNn7xk/bgnfNtms0U7JGTWQgitFfiV9MwIyIik/FJKom/UDuI4zrOMPRygYZaLjI39TP6BqY2Tvear6YeUAdrOmWHou3673cwXi7Zt+2FI0qTtOkJIBhyR7HIoDg8PoSfiFcDHV6uVNfbFi5e7cjeZTOeLBeI5L1684ttJOo0xFxcXQLhPj49pknRtNxmPt5ttFEXO2SiKrbNEmsQHL168oKdH+JIV/7569eo3v/lNrNUw9Nvt9uXLl4+PSyGc1spaA70ysL+cH8kexwnmHpP94sWLy8vLUVFopZRSToih729ubr748gshJXn24+Pjycnp9fX1weIAQfKDwzlNf4RHbLaqqmgqenp6iqPIDENwQhzyw8PDn//853TtgF1Xu7Jt28rLaGPT4yiKowj7yJZzXuqGnfDy5UsaMhgeJYSjo5jqXZi645xgjCBhZRRFjITabLbEuD/72c9wlhQJTk5O2raTYu/VkiRRUm02+yG+JMehTg5lhomBRFEwHhlFHhgHrDPTuoiS+77/Z//sn/HL8fExdbWjo6Pb21tuDdtCvypeEwKV1hohExakLEtkX7Ms+59/ErH1j7IH+Hnmj8eQESTSkiq8/gSYPjEm6ZLwyYe1NvEDUXEbyvccCSEwkYkXNA0ZAwG+8VPJhmFf6XXOGjM0TY3YpbX76Tdt2zqEIYfBOTeZTvI8V5+QFwFGnJ9bByEBDyyltNZQOlNKCuGyLJVSlOUO3CyYD8wWGYAxBv4DN9I0bRTtG3Hj/STkvQKoF+t3WLTBj1Xh1jiByksEEjX7N7o4jvM8X6/Xq8267VqpZJzESusojtIslUrtDb0Xweb1oFKYIQ4VSw2bNgBHhOGkRIOXRKS62zRN73nl3D4NFiSqRO4sLPuBa+bAg1pACcNDUDsJQnXGj6LkYzmZ/EuHCp8pPHeAq6KwLL0qbSjQ7euxVWWMIWwn5srz/Pb2Vmu9XC5JyOARhWSfDayUArvAo2dZxhECE5NS8uxZovv7+6enJ5gkYGJlWZr9rBtNgxu39s0330BwFF7bnNLl8/MzsgrDMFxcXDyv1qv1+uz84v7h4ej4ZDSepFm+3my/+uormPWD1104Ojp6fn7WSudZvttuu7ZTUmqlqrLSWjGghm0DTIo4pvXTtzDNaZoeHBw8PT1tt5vdbnt1dZnnWdBDoxocOCa4t7AT7u7u8HCLxeLo6Gi1WsN3YsFXq1VdN7/97W8D445HANm3LEshJOkCByp0BcFAZYINuazyU6GEEChsgzYzFAFW5PPz89PT02azQfa487LB9AyGrTIqxlEUx3HyuHw6Pj5p6jZN8+12Z60NSg2MWnv79u3x8XHmJSA5m8LTrrBmID/n5+e+V7+7ubktikJKlWV5lmXz+QJ1NfYhmw1zR1Ly9u1bxHHjOL64uNBa//DDD33f/9mf/RnUShZEKXV8fGytfft2D3p///33SFH0fT+ZTK6vrxeLxVdffVXX9eXlpXOO7KSu6+12GzT4GGIRvCDryaP5Cb7hn7oHjIj0c2CAfYQfKRNsCsEgQM2nwkTB9YGBhBYh4CbnZSoCB5EVZE35SyDP7AktkdZaDUNvzADZqW2bwUttEzYOw57xSaESo4bhEEIA9jkv/hF8xjD0Xdei39e2Td93xgxPT4+06VN9IivC32DlhRCh9N22+0Yb9rf0iqrY3M5Li8deSSnLMqov8hP9bednUWjfu6B9206SJFIrJ4Rx1jgrlZJamb3GmQsJO/hP7hXX+z0v3rHzTk9PiblYZNY/tCtz8Xi4AHYJIehaCAxjIQT9U6EGju2mQtD3PQUA/s7WDH221je9z+fz1GtAVVVlrb2/v//xxx+xYkgksXRod+M1acPEQACAhAiDnUYUhhcEwSPpcV4pk3ERJEY80N53xtB5AAYiPbM+6OdApW2ahh4xzEFRjOq6HrxSE0rUQQwGpB7Mt+/7L7/8MvtEhR9K1eLwqKzq+cHBdL74cPkxTlMdxzd3txi48Xh8fn6eJMmHDx/Qys3SVCvFdDDhXJokxksoknwvFgtUBk5PT2ezGdeDYBTO7PTsdLvdWmuSJN7ttkWR13VVljvaS2GLcLPCj/AChSM6/u677/I8Pz4+hn0/eI2v6XT6n//zX/3yl798eHhgwb/55hsh0Fyxbdt+++23wnPzydsiP8v9w4cPPFkKs8FtLxaLm5ub7777jjD87u5uuVyyr5RSbdet1+vV6nkYBnQHMH9MOEj28037NMkm42nbdmaw4/FECmkGU1W1tfbq6qrrut1u9/DwgOFisgUwJncRmliR0NjtdhBeoyg6OGDu4X6UW5LsS0FS7OdYMKbi6ekJSLPv+7Ozsy+//HLws7lev3791Vdfff3117e3t3AFf/3rXxOSslXu7u5ev34thEBA9/Xr10zcQqCsaZrPP//cOXd7e4tkC5XUOI7v7+9JIA4PDzk+9R+G7zYerv/T3EMABNy+6y8L0B4Ws/KzfYSXjoj8xJ5AYiEYDHF3yPuUZ1vqvRhDFErZGBHtf/jdCVs3ddu1XdcOZlBaKq1CqVMIodUfJsm0TUu8D9wRRVFoV9Z+Kg4vaPfzV5VS0jnTto0Qbrl84Hf8H9kSWIfx7bKceWPMbre11nZdP/jUuPeDJdBAll5MiesxxgB0SK8wGJD3EE0TiZOrYcedtcNgejPESTKdzZwU/TB0fvA1V2i9mhBadcDWdPfAakAgL/6EY06GG54j6xOWBWCEaNp48XAeFn4otKYTb3LIMQH4GyGEtRZtS2x3qNmymyM/h4vogYoIqFrMoOMkOTg4oI4npaRUSx69zxqdiPdzyjoorQHVQUyCTBGHsVqtLi8vCVx4OnxOIF9wF74Svliv18MnYQS1n7ZtndiPN4+iaDqbxXFc1xUJCn3maZpSigTwgZMdaFT39/cUOZRWSmulopcvX15dXQuhnJOT6RQpaaAGHodzzvpoI4lj4ZwQAhDy9ub24eEBA/T69WtQ37Ztf/WrXwV/XBTFx48fHx8fI63zPAWh7fpuNCo+/+IzKUWSxNIruFCzxbNSKbFeoN45d319fXZ2xiez2lze82r1n/7TfyJB7Pt+s1lvt9skiVFaJEgKp5s6DRyWzWZze0sYLiM//qEsy9FodHh4+P79e0r6zrn7uzsROOXWEtwQ1D88POByHh4eGEcR7Sd2iDwvkH8H49U6br2oASJU2+2Wt+d5jpxo0zQPDw9kS4iXsDnv7+/hCh8dHUVRvFgcbLe7um4OD4+jKJFSTaczhgBwZDhZ19fXy+Xy6enp/fv38AzZbLe3t69evXrx4sXHjx/xo5PJ5Oc//zlPnG6P7XZ7dnZWluV/+2//bbfbkQ0Mw7BcLqmEQfEg4ODg/O53vyNCxYAcHh6GEuPLly+D7/8JP39Ue/g///Y31rn5wbwbBqmltWIwJs9HfdcaP3YGxIATbn1TmNaa9lHrB3DXdf3mzRvjRxhS5+T3kNRTyUTwve+7KGKugyDWjuOE4CiKYjMYJ0RVVta4YTDC2DzPCy+ooLXO8sx5ZiFNsPD6gzoeKRGVj3o/MGuwVii1FxpTSvU9XTk9fQxZlibJXkM0SZO8yJ2Dl7ZzzjhnhRBN00KcDVX9tm2N6Y3phqFzzvZ9W1VbshfnhLVDURAaJ3zCaJRrHaVpOgzGWjudTp2Txpi264UTSRQPXaeFctbFWg/doKVSShozCGGLokgSpjer5+dnISSY3kCzgnNpmkohlFKJn6hF6TWgXkKIMKzGerEThtQLIZ6fn2ENHR4ebzfl0BvhRN+b6WRujUuSVAjkMEkxRTHKtVab7YZlHaxxTgxmmMym48k4L4qnpwelVNc1zoq+79u2F0JOJtPpdEqSPp3Ojo9PkEBIktS5/RjUvu8hqj4+Po5GxWhUMDg+gF0EaMYP8AjPYrfbbTZbpC+Qz5rOZnXTWOdms1lZVePJZDafNW03GKO0ckJorZ2xBUOki7zruiiJnbFRHDspD4+OTs5OHx4e2rYr/ExZmNDGGDwc8gZd15H+g4YTCaVJ1LddXZdaK2tdmiRxFCkp66p68/r16nmVZ/lqtc7zbLNep0ma5FmeF3ESD/0QRVHrVWm7fq8wj5Q0wcev//mv27opt5vvv/v90eFi6LtIx5vnzcHhQkoVRXFVVk3THiwOpZRZtpckUF5EljyYcdxRpPueju7i48cPr1+/Pjg4+Ob7b51wVV1FceScHfpeWFtkeRRprWSkVF2VWsm2qbquztK0a9oiL5x1Sqo40k1dO2N3m20axX3TzibTqiyll9ElJcLx4ycI0YpRkWQpYOZ4PC53ZV1Vi/kcUIV0FsjXOafiWEgptBS+Atd1vZRiMp3sdjutI2tx82nX9VG03z/grmyn/bj73hhjh8FIqax1BweHbdvESWKcQZvk6OhgubyPI53n2Wi01ykIeQA5DfVX2B/UtwgQj4+PAWOBEz7//PObmxuaDWm2xQsKP4eR3BFwDP4CvhPnTaCptb6+vv7FL37x29/+42w2T5Lk/fsPxtiDgwOg0f/pL//8T3UP//dvvjbWjsfjOE6cEFJqKWWcJNvNmkg57Cfxx0qoZGrcDEFBVVUvXrwAV4l881fv6eGfRqbW2ixL2YthOlgcJ+PxxFo3Hk+qqjbGOiu6rt/tdsK5KIpSH4dKKbUX+sZMwGYhK4SMATAiPF1n2E+zGQXUiN0phADT3O12QjjUW6010+lEa51l6TAMWsu6ptWe6x8AAYK1tdYybFgIEcc6TmJnKYzrJNnXDLVWo9GobVuK0uPxRKl9Fbdp2pOTk8vLqyzLUL4ygynyvO+6SOu2abTW9Of3fR/H+6aKPM/X6w3IEtBcoEWNi5GUMs33fdeGzgJPI8H6B84S5w3suCgKeAdCCCFkXe1F7uhTY+WttWJfyGmHoU+zVCq5Xj87IfO8IGRbLpdFUbRdd3Z+1tSl2ZNfhZSSsVHj0WQymZC75Hn+9u3bh4eH+XxfvoP7i/n77LPPnp+f6SsPQGLiZ87wMo6f8UxEEMjXb15LpeqmbtpmNB41baOUss5JKdMsTbNs+bgUUhDlUI4mJXLOSSWdc23XGWudEFmevXz58uHhoWkbNr/wk+NIBInELy8vORRAN+TTSZLESRTFerfbCCfM0K9X66IoDg4O72+uszSl4VZJOfRDUzdOCBHpYjyaTGer9SpJEunlfZz/agJw0K22bsfj0eXlh6YphXCRjlarTdt2DHA8PT0dBuNVT0Zh54RGGQ7LbDY7Pz+/urpiQIIQgmhpMptudltMkpTy4f4+klK6fS+tUnI8HtV1Vdf1aJQbM0ihhBCR1gQfUsiqrNq2HZh07QnTnMeTkxPSvizLQHXwVZGXEiHunE4mB4vFbrulgBSgUdLBYRiclM3QNW0rnEviRGklhVitVkkSR9GeJrNeb6xXVCPApe4FWaau6/F4stvucJBRpK211rquax+fll989bnS6vn5ydrBOVtuN7PZtO1aqmXDMOBcR6PR8/Oz8nNBgKrAZu/v79kS5Ge85cWLF3/3d3/H6FmUrUNqZT0LkSoFpR2CP3Ba6C1JkiBzmWXZhw8fPv/8iyzL0Gl3zhXF6C9+9eVPcA//H8NEe99YGyIyHmQwgr3XFXB+4GiapjTH8gLqJODd4Q6NJ24K347LJiARoaYakE3lJzAHsQ1i/5COKK8DrPyURzwzwCLXz+YmaqNkCuagvTQxCEm8l4TaU0WFl+92ztHIHUAP/13CGFPV9Xa3J/XyRmCcOI6zLPf920IrDV0B9C/4VEqjsVc+549lWV5dXV1dXZHt4r2Ij3BySikpBSiK9pISwk8EIr9GPBnlpdArB5yaepkzHKH1Xe7Wz58AViKJgc1SFMXj42PXNfCMhRBKye1203V7qrVWiqcP5zWKIqV0XVaRVNK6UV7EOhLGPi0fu7qJ44Q4A5QJg04+DibG0Qo49Wq12m63UF8A6IhOdrsdjECE/6BX8sk4eHIg2n2llMvlEp4VjhPWIK9Ezoh7wb47z3ODYqf8pFjOP2SKw8PDX/ziF1VV4fwohrP+wOK/+tWvwDl5ykdHR1mWtW2zH9FkLSJau92u67q+7w4ODn/729+Cb0wmE98VWNj9jKkozzJwrdSPmuAY8jnGGCYz397c/uVf/mWapnd3d+Bpk/F4s1lza4eHh4vFoixLvCnni7fDESDmvbq6+nf/7t+VZfnq1Svn3HQ6Xa/X79+/t9aS++JR0nQ/2DGKoiRJsyw/Ojpu2xYRBBgKjMd5fn5GeMoYg2XkUO/HkNQ1zS5EMExbgzxJlVh6gY2yLI+Oji4uLliZyGslJH4WHjCAMWaLPu1mw+N7fHyE/Xl2dgZTiPnhnAIEH5VSTdOcnJxkWfr69WsgR2vtZrPpupYS13K5/PLLL16/ft227VdffWWsvb6+Lquq83q68IPDHnt8fIRXyqpiUn744QcKA7PZbLfbfffddxiQu7s7ay23H1YjjuObm5ubmxshBNTbu7s7oGztu5GIfU9OTr7++mss81/91V8lScIUEJbiJ/iGf+oeYGKQjBg/D875KQ74c8429lR4Ji+nnTG/vJ4czcNEe1YvCC8eyHxC+Q+4Oa8PAVeoT6R+wgz8Pyq3n+4b7VWdsTVEB09PTyFp4ACQ6GH6u6579+7dixcvhKdm8wvPBl4Bnkx7SdRQTdFaK7knRB0fH1OJ5dqw41JqKbUxTutY6zhNM6Wi8O2cHOA4ajOtnzYTcqDEt1izFSKvZuOcINjhPPCZSZLkftxQkArg8GO4Q98plQ8hBP6ApzydTimUhVZHlhcwtKqquq6VlkIKIV2cREJCxG+FFEpra23TNBQGoigajYo4ipxzSZLkWTabTodhSJPkhx++5zQGkhJ11HCG2QyE/8jR4POg+QLdLhaLptlb9lCF49kBJQshnp6ewCgODw/n8/l8PkfoIs/z7XbLvOLgTQN1J9RU2duDnw2Abyj8FJDn5+f379+/efMGs6K1Xq/XxLzEg0zJZhnhhQdPTKgBjwASHV7cGBPHUdu2qIESgQohrNtr0DonpFTOuSxNxSdDRwieuOwkSSIdlVV1cXFxdnaGc43j2No91RBXCuuXscyAkHhWGMzQviFQ/vmf//nz8/PBwQEH9vXr1+/eveu6/s2bN1Gk+QopAS+jYbBdNygVLRYHVdUIoSjCBcLx4LuCMv8TtLMobxweHgLSsvmttQiUKT+enU37/Py8WCygbxDGpX5+A5WeADmSVVCzxSZQY0eamypa6Nm6uLg4OTmBB3V4eDifT5WWUay6vjGm32xXD8u77XZ7c3OTJHsJht1u98UXXzRNM/R9FEUM+QEjQcSaFEd53uZ4PGZMxfn5ObRXOBdZlt3e3s7n88fHR0ZdaT8n+ObmBgwNtzGbzU5PTy8uLrBOGKKqqvY6LlH02Wef/Zf/8tfL5cMwdN9//23T1FGkqqq8vb3+ae7hj4eJ/ubrOE7iOFZKd12n1L6bwVlT+9HYgIAhbwjYEexP3AMugaEZ2tMfA97NUWEf+HhTEayhnaK8VrYQAmARXfXBq4hTq6He0ntVNTY67z04OBBCQIdP0xQpzZAfcLVMhZ3P51dXV2maEjZGXj2bK9deplEI3FiMH6aynee5EHtxJxAtwBkpJauEeyPO6v0PDFR6WNC0SJI0ZFfxvu9akTGAlsznc+yI3TfxaxZQaz2bzcz+x3bdXtUc+84F5GggIpxF5xfJvl/e8XjMKuEqAHwpJI7H40ARiaJ4GPooig4PD7n+OI7TJA14nRCuburpdBJFkZLaWTcqCq10U9dKys16s16t54tp23Zd13VtX5YlINVqtZ7OpnBA4SAtFovvvvsOJRyoTbiH9Xp9fn7+9PQEzQNzRs6Ov6e/iTHC1vejAfuiksa+jf0ILXLK+/v7YBqMMWRmKITPZjPy/ePjY2wZCeiXX37JnCtemef5/f09+2q73Uo/1fz+/p5vDK5rPC6Qwt3tGBdRjMfjLM2yNCXQC/U5wtUkS6Mo0pHu2y6JYyIGY22W58T7WZaR33dd1zVdkkRVVR4czL/77rs8y42hIKdQuqY0BS2NcdCwnHk7UhzDMKDq86/+1b+6u7sDE1NKvXr9+uz8/Pr6ygwmiZO+77VSo6K4vbsjooJylqYZUQt1kb7vaVFmkA4hvPGie1LKo6OjxovloY53dnbGX+jJmM1mVGVpZSfy6/0oC/raQIq01mcXF0LJOI63m81kPG7b9vbmVkqZpklVVchFn5+f03IIlK21RikPVu6HDx9evnxJLZP0MYojWtafnp8WB4vpGQU2TgAAIABJREFUZKKUWj0/G2PK7VZKuXx6evXqFf0N8/kc6vxqtTo6OkJbKY5j9ipW7quvvvr+++8PDg64gI8fP9I7DFGKyAzQCYN5enqKEcA+kExDvsKDov7gu+dqKQUvw4kOQz+dTn9a38Mfi2r8/T+mKbqYfYdEj5R1XVszAKQCAoSoP1hSwiLnlZaV7z8gqpVSUmQPJQfhw1jqM9wJu9b4LnznhfnIsIDLAd2SJOHMJ16AKNTDA1SF8AAXRrjqPM0ft0RbwPHxMRNsgngLyUoIQ7DRWkej0QgI0jk3m02H/XzfjuvhS2mCl1LHcWKMlVIlSap1JKUSQjJfl3gK+TY8Ytf1eAXthz3hHsJJgPYD/YM6CzcuvJpFSMgg6hmvtzGbzZyxu92uN0Oapk3bSimV1hwzIUTqZxhgf9l/uGQeH6Z/vV7DLbbWFEXOWOyua1+/eVvXtZTi6elRa11VO2OMc0YJzUPEOy6XS3zzdD7FZfbdkKZpVdWLxaLr+r7vFovFx48fyZmOjo6ur68DAUl6kW32wGJxsFw+CCFACdgzrBumh5gXN3B4eIiHo6GX3B/HGbJMogo2Hsj+vpTdNCQuKCg8Pj6GMubl5eVoNKLBmw44tORwORwK+GNsXcGIpKrsuoYVaOrWGJNlOSRdMwwkjiQ0h4eHm83GOnd4cvz09DSdTLIsK3e78XjsEDly7vDwEDBws9mkaXp/f9/UzcuXL//jf/w//v2//9//9m//NomTNM0HYw8PD1HAJFtl0mfnp6YTIgSNXhCb6+vrf/tv/+3vfve7LPvvrL1ZtyTXeSV2xpgj57xzTQAJYiDFSVJTbfWyZLXf9I+1bC8vuy3ZLZmCmiRIgQRQAOreujfHiMgYz+SHnSdY8psF44GLwKq6NzPinG/Y3/72jhaLRRAEh+PhBx/+4OHtW610HEdSSMF5HEVd30sZzWbzq8trznjXDWmaWeOUHgghI2yA3gUkPcBreZ6jf1oul4M3ZG3bFnece1cxREnsikJLFZ8ZIXv0DkHBFERRkqbT6VRyoaGXfNYEYpxzoFIoN8uy3Gw2iEWLxQL5G1S0h4eHFy+eOWdPp8oY3batECxJYs55PwxG608+/uRU11opZ51SyhFHKAWgB03rrutGVYXD4RB5RVtwqa2119fXv/vd70bLssPhAAi6KArgbCOOguFZ5t1ZgK9iWI0ngKOLv3U4HK6uLprmZIzOsvR4PDd/+/3+b/7qP37X9PC//h//FVcxDKO2bZ09h9ShP1sRABIZ+1njGeL4PqiggXejw0BkB79w3ADAEdTeUzqOY6UUHgQ+xkjSCr2jHDIqYDj8EGQFcB6wFYLb7ryJFaASpBn8TFxa4qcRk8kkCIK7u7vXr1+jg8FlBpHD+HVfdva6OBusaw1HkaCu60ENzpKxc8LohXiapvF7cEhRxpjlcoGBFdQagHUIIdCxjfO3MaaM32XME13X5fmEUkL8eN/5HYg4TsB9QuOCsohSmkRx0zRFVTLGHISRw2A8WOdLFQQjExG/dwy4IMJDo9t5+zmIVCdJslws+r5H6c05V+dVecEIG/peDQNxbr1a3b95wxhjlE5mOYTbAhnmed623fX1dZKk33zz9atXr3BytNZv3rxBNYRXZr34DNaywjBI0yTwAvqgDAFmDL0IWN/36Ou53xrDzAAJGBtkzjkU/pif4/QC1cRPQxUy/kzUK6j9MQsVQkC0gBCyXq8JIZPJBFXqy5cv8cEWiwX+CiFEClGdKoTIqjpxzp0j6GOauo7jc1G82+2wJrLb7ear80elztV1HWPXqSjCKMQXwfq6EOJ0OuleU0ooJWmW3N3dPT4+BUGUxOl2t5nPZ9PpFGxIsJ9RxuHQor+R74gOCCE2m83f/u3f/vrXv16tVm3bKq3WFxfPnz//L//7f1kuFmmS9n2XZ9nz56+Kory4uHj+/LnW+nA4INlEcYi3YK29ubnB8x+8+yyUR8cCAsDvOIpD9YYzybzaHYIjqhZKKTwy8bLa0TCcsfXlRVVV11dXVVkKIQQXSinGKQTAP/zww88//xyJB5kG5SAqyBGwklJcXV19/fXXGM6hgJgvFkVRhkH4+eefv/fq1aeffiqFTOLYOmKdxawCtCvn3Ha7He22Om8G5ZzDyjTKF6jPoj0tigJHBQUHDDkQD3ErgYnBmrRt2+vra+0XA1CC4/S2bbtcLmA5J6XcbrfOWaWG//F/+Mvvmh7+4Z/+BbCg0abtOs4FblcSR+NyNvd2WuSdBVQgP8h473Lq8Zphy46GEWEXd35kGeFo4g3h5yA9jAUdmg/nHCSecN8QuXA48DkB3YxgMTIHgoLxm7f4LUBUjDEgaVhP1cXrxNNATYEpk1LnxZamqQlxQgilFaWUEoYRIo4FLlvTtFKenQnGLkQplecZVtIwpQi8HEUUxa33yUD8Mt4D2XllbKwFEELC8FzRn+FpP/bHKXw3G1FCsyzrmlYpBb1CyphzLowi7RfxcKRGHBPZHaAc0vDY6gLp1t7FCPk1jOK6btDSHo/Hx6dHrF/FUSKFxAx5Op2CxgeR17brML2klBoDskcO/58sy3a73YsXL3a7HXIV9cR8FE1Iik9PT9PpJPaGKnhreFYghsPDgzGmtcnzDFPovu+Hvhd+ZxAYMSrocXQ0tuRohSF6ihgK0R70ZMj61gsaYu6NeTWQTAQ+DMwh/YRtKSmDq+sLtI+CY1eUCCGWi6Xqu7Iqhcf9jsfj9fV1ludUcGTKNE7bpplOJlEUPT4+tl03eD9wpZQ1VmttjR2G/uXLF59++suf/exnX3/9jRCB0no2m1BKxveLFzpaeIHpQAj58MMP8zx//fo1dLa/+eabPM8///xz4BtZniutF/OF1nq/28NKdjadXd/cxXFiz9uLoizLOE4Iof3QFUWxWq0Wi4XwUhOgJyml6rrebDavXr367LPPbm9vx+NXVVXjjWnHlhpnG08GpTSKG9Qx9TtGvNvdvuna5XJZHI/WmL7vjTZN0/RDhxuEuQUSQxzHmLcLIa6urtCgtG272WwIMZwzpYa6Pl1crOM4Ksui7Xopg/p0MtrcPzxwDzFRxtAH9H0PV1F0nzjDI6aC4OOcS9MUezBPT0/A5fCHm6ZZLpd4UC9fvgTJtfNG39DIQTO33W7BSgJsTilFPkjTNI4DpYckiVerpZSSUBdFoSPur//TL75revj1b74KRJDEqRoGQojgjFCnjWKEIm5iroB0aq0VIoiimJy1iQQhVggOvAX5FkMhVGd4/cQT2lBxYyaGc2C9XzQCNEoDRAcsGSFiom0HvQqIAa4xQFhQIwghiNSTyQQkDShVASKjXhvKGHNzcwMmma9SU0LoMCilNKWQZKkXi2XXNUkSl2VBKRVCEkLTJNPqzLJHgiSEAELV+mxVBtLROFSXMiCEKqXxf5wjXdcbYwG/Wq+ABBAAwA4hBA+Ec47oGccxY+faarVaoVlu21YpF0cJcdQYy6gA6SCO0yCS++MBi2PoQ6WUknPiByQoPUasj1KKhXO8F0BwaDKwDAF5FQiNGW3jKE7ihDEGQJJSRgiL4lhrvd3tyrLEPtd+sx36YblacyYEl1VRaqWjMGCUBFJqY8CTefv2LaUUkjh6UJEU1ujZbEqczdKkOB4dsWEYIHwgcFBKy7Lc7Xacy5vrO0b5crFSynAm4jhxlhBnT2XBOSvLo9U6CgM1DM4aTrlWyjmXZ7nkHCrfzjohJerrkQqMbHF5ebndbkHawc0PvXMGcE5QWvM8f3h4eO+99549e3Y6neD3hxZtPp8/f/5yuVz98y8/XSwWURRBJ2axmBNGi6oMopBQGkShIyQMw8V8EXIxzSfUkfp0ooSkaYpS97DfX11e7re7PM0O+z2jdDFfdG3PuRwGTZzYPO2NIU3dJkmqNFZBOZhaqOGklC9evPryy6+cI0EQLhbL6XR2c3OrtXl4esonUxmE3/v+B4dj0Xb9+vLqdGoFD06nZn8otbEyiMIwUcZlk0mWZ5SxIAwvry6Nc0VZvHj5YtCmbTtt7Hp9KWSQZvlsvhgG3fftOGcGglQUxWw+Z1IwzkUgnzZPlrgoiVFJTPIc1wcpfwwUyNmIFegywzCcTPLjbme1nk5yzmgghZSibs7ST13XFUVxeXn52WefoSVt+s4Y0/Zd07ZxmvRqgBWYEPJwKOq6DYKIMXZ7+0yIYBgU9Mqss1zwJE0cJaemMUYDWgQxqWmaqqqur6+dNov5XDBeFgVxrmvbMAi00UmezecLpbVxtleDUkqGYZqmRmtoK4BH8/z5c9QoQRBcXFwgo2MXnXNeNw0XIgiCOEk0DMrCSCv14Uc/WiyWxHHBRRhEaZrnWf7BBx988P6zf0d6+DfMpWmWRzKAUDbzUkXIutQLMwC3jaJICDk2p8MwwCuUC84YxX/RXqRhHLyg5EddADrT2FKMKDPqNdQRaK/AydNaY9bNGGOe9gpEi3ozHIR+UGzBWACrSp8do/px6oC/rrX+9a9/DTdBwO4jkRf5g3iNKcYYNuwADhJC9DsaG70388KX4pz1fQ84CM0WalvkJ3AYUG6MM/D5fD5y7cFhJ4SgCDXG+G1hlyRJ3w9jZAT2dZYtekdsFXYCUkrOGab0xPttSSkZpVmWccqYF7JF94NxDvDDERYDtAKyLHItPkxVVQBYZrN527bYxMnzCU7L6XSq6hPwQ5AxqqrijNWnU5okaZJMp1M1DJQQNQwQXqyqCiPl7XYL1mYURYwQYmx9OhHnDoeDsQZtFmIEXCUwH2aM/fznP7POHI77p82jDAQXrO2atmsEZ23btk1DCe26rms7rVSaJHEcz2YzfAZCSJZlcRQxxhg9++UBokT8wncZVwFwrowx2+12sVigKN5sNoyx1WqFRqTrupcvX+KHcM4howRNUOz0Gr93UlUVGmLA09h1r5sGnl+L+TxNEkBb0BLG2AMd1Xa7BU7yzddfN00DRTYwW3C5iuJAyJnzrbXGV6CUfvnll1EUrdfrNE0xPcb1/Ku//uu3b5/CMF6vL+um/clPf1ZWp7PzLuWciSiMppPZdDqbTKeEUkdIlucXl5dhFAVh2Pd91/ePT0+z2Wy1Xg+Dqk6nIAizLL+5uf3hj34k+FnJYzqd3t/fp2n67bffQroOqEvobW6Louj6/unpCTd98ILHuMjEr1iBGPL27du6rqMwfP7smRqGtmlGId5RSxT39+npqe/7r776arPdXlxccCEwJICrz2w2u7y+gm1j07Rv3twrZZTSeT4B9Of8wsHTZkMZS7IUgwEUqeiijsdj0zTr9XqxWLx8+RKkbbB4rbWbp82ghsurSzRSeGhtd/aHl964G9/38vIS/Cvgcs+ePfv2228557PZrG5qQilYUqg/2rbd7w9a2Zvr27bthQiUUkrp6+vrf0duIP+v7uEPX9xTSrkUbdv2Q88YO5uJWye8rIIf1XLGGCE0CAJKSds2hLiuax1B38DPYKuUOMRlWQJRQWTHcGwMSdYLt6EMB7iBiTE6L9BI+r5H+y+lRLiU3mkHkY4QAm2yOI5HUU88aOiXxXEMCiP1FiJIQuObwxgAn3zEZLMsg2Eo4CDnpUfenbjgU4XebFJrA4AeYAXeHEYIiG7IZ/jukMrBXzReGwqhHNlRed8LFK2QtnVeRt/TSdM4SjD6HvOW1jpOIvA0AM3j/yRJorR25LxQkiQJnA/GvyuEgLAM8buE49R6nMoIITgXRVEaY7jgh8OhaWoklXM3rbUxJs+yLMv0oBhjTHCITyC2ntEzSqezGcbXAHBx4Skhl+v1MAxt301nM+fc0A9N26A7RAoHcwN/cTabj0aSeOaEkLIssySBkYvx+ijoaMFQBH8BWuXM7wGEUYhpGSFksVhgsoLiA8Ptcdo0EjGGYQBHpes6HD/keOBLqAO01uv1Got+I3cFcRwXGBqC4+guTdJAyizLzgiStYBWULUcj0cMxoBcr9drShiOxOi0EYZhXZ8IsfgJxnsO8rN0bjYy3+q6BjE9S9Pjsbh/8+b25pZRerG+iMLw69df31zfFIdyPpvPZwutjOBSisAa2/cd8xLcVVWtV6unx0c1DOdmnVJMsPEtAinKohjUgEiC8E0pPez3gQzUMKhBZWl6/+b+Yn1htD4eDs5LXuJsOy9NhvIf3xfmItI7pcfebh0dA24N+m8AMsAzrDEX6wvBhdF66Ic4irTS9ek0m86AVxNCQALETKLre844JYQzpgbFKN1utnEUT/IcdxPFyt3dHcLXcrVigs8XC23N7rCfzmbELyFZY05V9f577z++fYyjiFFmtBFejt543fuu6x4fH99///39fo9rhVHz4XBYLJeOkM5HFSllcTzmeX4sysvLS6UU3NWGoR+GIc+z9178ezLEv0kPv//ijdZaYq1BK8YYFpKJdaDloUDGnQR2HAQBpayqSsZ4XZ+Ic4ye8SVEGcxnRsYkQH/iCSeI1MMwbLfbcZaAFIo7CQ4SegvsamFIhQCBYpx7Fiz4G+jyjscjJj9ghkAnEo8Y9RcCLvjC0i+LQYYTA0k0JQA9J5McJQaaCQwtmXfBBb3SeHFKSimIsNZaBAv8TCSnMZdYL+EXefFaoG24MxiZTiYTPIeRgI/Zr/U+dFAQOY9b0kQb5YhVenDEBoHs+jaOY5AokB2B4wshKGOUUchC4AnjsQNkk1JieonXh6AJWXIsM2IqK4SEpZ21Zhwh4h4aY9w7a4bXl1dlWTLPckaIxDdN01QGAdY18DXRe/V99/6rV0EQbHe72Eu3VqcKRwu22Cgdzqx/wSHuGwQScuuQn2tOZwlC6mUH8erDMMIYAEgpygsZBGmeIQ3gP2Lwjt8Veod38J0wJwN8X5Yl8fw3wMr4IYfD4e3bt8aY3W6HSTh6iNevX4P6iVA+JmnjdQ+FENbZxXyOkzAMw+FwwCAXRsfgboAnc+Z3JZEjNs0SQl0QSM5ZdSrzPFPeDYV5g684jqfT6ePj9mc//blWJgrj46GYzebL5aptug8/+vjzz39vjL28vIyiaD5ftG2bJCklZ1V8xAGspIEUxL0UOfGW2sDBUNjikGNdqSzK46FI04wzYbQdeuUc2W33aZY473aAxR3Oedu2YRBgQwLIPvIBntK4hITfCPYOri0OLXprvCP0Utz7mEkpUTJiPInLu1gszosvwwBMGLl/Npv1fS+DAG8feZcxBmha+jULPBnOOdoFVJwAw8uy1FpDhffy6gq8VXBht9vt9fX14XAI/XIfajtwmUB8AOmOcw7Dc0LI1dXVZDqFvBXAJavN6XQKo0RrRQihjAzDACGssix+8Wc//q7p4Q9f3GutibcnY4wxzqIoglsnpiiEEPCR0SWEYRhFYVEUYRgMw6CNZpQ5d17KxR9+Fz6azWZQ9VJeXQ6JHXj3GGrx36094/Jo3hGeMIFBdY9Mgw+jvL8V4t3hcMAjvri4wNYi/gC+wnhQgHuMtipan0N/7A0GkNWWywV6VQQv/JnA6wNeXV1BZhyVb9+f/eBG9AZNxri4MNIYRkR4uVwq77yEIwLYAfsZyCjjChisNJ1zwBBQfnIupZSjOSgSBs66cw5bHZeXl+fhqnNpmg7qrJgGbAdZk/itK/BVENHwtzDUBZefnYVVyDCgJh36vocgFRTWGGNGG0rp0PdKqdlkGgTBoBRMepEmT6fT8XiMk+R85LzjwvkmMx6H4atXr05NDacK62xZVYBiUBjiveOJaa2///3vH49H9Y7diNbaGWOtRWpHKYosK4SEebV4Z4U+DEM0VUiZzrnnz58DwEHZgWEMxtRYlAWqAMAEA0OlFMod5Ineaz+AqMb9AjbOPKAJhF2AKih0lFJJnGQ+LgDUxo9CfsINGrk3RVFAt3ykUY41RBBISH4ifP/gBz84Ho+z2dwadzwesR0ihIBpGqW0bXopZNu0eZbn2cRok8SpGvQ0n7VN27UdZ0JwwSgXXEZR6Kztu14r1Xc9JYRRRglFlc0oFVw46+pT3TbtdDK9vr7e7XbOOZxnJDxjjFK9EII4B2fTp8dHdCTIlLh9KCvRiIzpGbd+pJ9UVTU+Q9RbIA7ANw06bKDwnimCWnOQFT3jGZQnME3gr4fV1Pp0ghWrEOL29tZa++zujhKy2WxwxxEuMKLL81wGAfxa+r5Ps+z+4SHL8yAMi+NxvVp1Xbderc71bhAIznGwmXdwOZ9G516/fo38xBgDx9paK4NgtV6rYdjtdktogaTpZrNxhDpn+74XEjrejlK6223/81//d981Pfz+D28opUzwrusG+NuBW0bQK5wrREjxMMaiKEb3UJaFlIJSwhgNghAAFIJI6/1eUFWNAAWiP94f2mGMeQH64+bgrWPJU0oJgTMcrHEzDmxoSin0eFF3SymxnALNcDgr4TOg4saoagSdENGEEH0/4E0ANABiq5S6vr7CijzOKAINjmDf90CxkDxwTI3fcdNeGXTcZhg7Feu1TnHbES4JIchbgKFQ2qBlATxNKem69t0SBvEly1JKCZz1nLN1fQI//nSqUaFg5Is3eJ5DBAHm/4h66O24t51AX4JqBYcVAXez2YB9L4RI4qQoyuEsfKK0Vs5ZkLvjONZK4TFyziFkpK2Jomi5XFZVhUuOuyfEmaeAC4yTxhmfZOn19XXTtnVdt12nBsUFH8kn1NPbCCF1XVen4uJyvdtvjdF93zVNrfQQRkFVlIiVQCdQjKdpGkUx1JmMl1dijKVZxjiXUr548eKzzz5DCQzGAd4mEE5KKbjzYyNovc0izhWEMcaFbYxSAQRVVfXFF18gH6N52m637+IewKm6rpvOps5YvAWcFkQrfGUUDUqp9XqN4w1CMmM0TROlBimFUgMW0hvv/Z7n+Y9//OObm5vtdntz82y/P3AuGONYEAHjY7vdTadTSD6MtIVhGKx16EGJV0cnhDDOnLM4kNwTCPGv1nNnuV877/s+juMoiouiiOMkDCOtDcTmqlPJ2B9NtBBzUenjsgzv2EojgCi/vwkuA1pSMDVGdiX4BZiuKS9KhhSbpikkRpDd7+/vYVb4+PhICKmqCtu4nHNUV6gLCSEodq+vrzE1vL+/bz3bylp7Op2AkaJhstZCUwRxH2IbeKSo/Jxzbduu1+vNZoOQhfYRrS1jDG//448/7rru9vYWPtXWWiHlZDLp2hZH12HY3vbL5UIIQQg0F9q2bYNA/tV/+g/fNT188eWDUooheDnb9z3jTAhhtbF+mwx4EbgHcZzEcWyMhuGaMYZxRhkj7iz/eebweVCbEIIOGlWD8WJb3KvwY9EcFwwDiTiOcUbjOMbOKvFKy/g84w4XTg9C7fi3sDeIvn6cnaAnHat1IFSEkDiOT6catQzx2mdYg7q5ucbnQVcI2QPAFMr7qaGYReHp3lGj0t46lBCCj4SlpPML4BxNNKLPyIfjXt4DudYYg8MKHWZcIfxSpdTpdIKeF34R9Y5+Qoinpw2WaIwx0F9DZ9YrBYI2IWSc9WEshMuAIhTNfhzHSCSopwghmK6nWV7XDaWUUmKtheYu8T6XzjpCyHQymc1mgZTL5bKqT5DNwblHPV4UReWF0NFvQZGYURqH4c3NzWa3BcnPepkWxAXkLSgzE0IcOduQINriMTLGiD3r9QLsBuKR53kYRiPoNHhxTSll23VaazAUkbSg/IExT1EUiJgYPODaMy8KAqo+wA2gEyAyYdqBB9g0zVghYao0zkWOxyM2ALj382CEAlMdNd3GEImTjLEWdBeklJQScEBwOLXWuAjW735/8sknxpjb29s3b97M50sw5RFAsaBa141kQSjDQARSSEqoHjSxhDjKCOOUc8rbpuWMM4IuwTlrqCMwdYRcu7OWESq54IxZbSghxDpnLaO0bhrU5tQr8/d9LwS3SlFHpBCC8bZpBONWG0apkHJcgwXqMI7BEO7HZhHPEM0ZIiz3Ej5AlUErQPDBNUnCKI6itm4YpYf9njh3sV4T67qhPx6Pu90OFNWz9rNzTpu729uHN/fEutlkmibJermyzsHxCacXwyHOeVFVEKm0zlanyhGntF6v16obnDHTfILFoKe3j1EYFoejkBI+lTiWqJNwbZGrJpPJ27dvgUolaSqDgHMGKV8cQSFEWZ3yPJtOp1rD9jy11hqj/39ID7/8599qrQmjbdsOanDOccGTOHbGjuGScw7PHMipEkK0NsPQM8aN0VpraxwhZ/watSHKhyCQXdcSQlDpoL1wXj8A425A4SMYMtaemEkEQQBaHrI9msRxmAOSHD4kjJ0RHwHrg+kMLAXh+50R5dkMLkmSoigxE+69ViBO5Gw25X6lGeQTMI6BSkkv3WHP1my07wfniOB8sVwO/eAIgTM2iheUvah84ziGXimCFybz1u/34YsTQkC19F2FxffFW0MuSdPkdKoIccboYej7vmvbxlpjLcF3p5SOKrbDMCitiCNCCjUMQkp0Xagxca8Qy/q+L8sSTBsAu2i5sixLkoQz3rZdHEdccELcZJKhwwPOxjztCq+AcU4ZhTMiSDhYERqGgXtXauFNGZVSkzyP44gxdnN3WxyPbdu2bcM4B7HqVJ2CIOiHvixLRiklVAimVO+cc9YIwZM4jsKQEuqsHUFzoMPnpJ7lddPAGtN5IeEgDGfzOR7R7e3t4+MjcDaUEcy73iOYuj9aUJyNXYF047mBw57n+ajtjOJ9OpmAiYQrgHhXn+owCkF/yvNcCB4nsVG6qWtKmbWGEFgonh1b8Q9CCY4oKNFZluDVa636voPNA8osxth+v//FL37x2We/ff/97xnrvv3m4fLquj7VjIu27WazeZKkeZ4vF0tEUpxV3FOjNSFIgnQsXxhjjJ21GvE0iJcnGCsY55w2xlhDGeOCW0cJoZSyNM0Y41EUp2lmjJGC9/0gpUBpFYbhfD4rymM/YPuaKXU2R0GpPpvN1us13ibwPYyRGYNdI8TNjLWu74emae7u7tr9f6zeAAAgAElEQVS2HRmu/Cwl69YXF2eF1yDYbrf4yVwKBCVIdSFMH3b7JElWqxWCmNYaXL7buzuopaK2456+fzqdGGegnPR9r5UGcDybTNumwbxqGIYsyzBfJIxiloYJHO748Xis6xpKWVdXV4+Pj+eJiDFt11JH0JqEYSg4d87VTVsUxcXFRZ5ncAnE5t1f/sXPv2t6+Pt//OdBDYQ4pTSjTCsbyDAQgdaKesUklDwYTsZx1PedENxHK9q1PbICISgnBbY6jTGOGMqctUZpZY0h5CxqhEoHcRNXdBgGQAcjuI+cjMoU/4uJK/XkTudsFElUBITQyWTaNB2lLIpC4c2TsUkwDAMWgvDPOA/A5p21pm3PKw7OWbiZQrQE0wi8icViMRIxgSNhFQOFGGeSM8mZCMNYisBa1zStUpoQ++74XXk9vjxPYVcnpYDFKSFuGHqUt9I7lQJ2kFIac56OoufAF4yiqG27qjoRQtu2w/YGY5xSxr1mLVx6AN20p4YRQgnpuz7Pc0apsYZzrrqeUhoI0dT1crEw2myfntI0FVJCew71Pr6vc5Yy0g+tlEJrZbzHETJEHMfVqTLOpmlqCTHOSiEBG758+fLNmzeIL6e6ppREccQ5w0Y0zDbySXaq62wysc4WVZmkKea9nLE4jDD3joIwiiKrTRgEs+lst9lTwuIwcZZQwoilnIm+V0EUUSYo52mWEcqEDCnnxljKWBhFjpIojgc1pHleNw0KUlQMQHvOPKI0RchAXp9MJrPZDINxjHkQggFe7ff75XIJvBhELHQtzPFpPl3MFpQ4a0yShH3XTfOsbbqyKO9ubh4evj1VZSB5FMowCKMwa5pWcFEcS2tdEqdG2yD8o/kSugSsQwvBuq5P06wsK8Z407THY6G1MWqoT5WQouv6+eKibtpsMr+4uPrmm2+jOArjSEppiZNhkOaZkMIRJwPJJddWE0YIJZZYR511jktOKFFGdX3HBOOSC8GFkEJIHDZjLCF01JKxlPaDUsYyLnqlHWXc7y6gqPJpO3aEKm2NdY7QpqtXF6vvff99RyyCBkpPKaVWVikbhsFyuby7u0PNBPyNUtp3Ko5TYFaMcUY5CPfaqDAMr66uUCMCoAvC0FgbJlGcJFVdw1ii7zopJWHnISjwbSShzfZJWx3GIROMMPL26a0MpTbaWgeYEbkHTNOyLK+uL5hgh92u7wZnXVt31JKhG6yzTPBTUwspj2URJXEYRUEUnqpWymDo1bE4BEGgVB+GQV2fdrs9iuw0TT/44IPf/e530+mUcW6s0dpobxIcBWHbtm3fnupqsZhPJpM0TavqFEWxEOI//OmPvmt6+N/+/r8SQpMkgYsyvJaiKNJajduzKEUBw43tIUE9wBhCXue9L4ByAKykjDhn4jhp2xYWDmiQCSG4WmOw5n4xFcU+csConIGYAnkW3GTnHGM0CANnnbVksVgwxpXS8/m8aeqLiwtsAgfe0Qz6B9ZrSCD9oCLGaAt+OGNPiiYDgRVl/mQygbcfwiXmxmhoOOdReN5xl1Kg3gfj05E/fuWRtoQ/hgEM5xz1l+ddnEt41CkoT2C2BoQB2Q5tjbV/JC/ipO73+77vwzDyVR4jhKCGHYZB9QOjNIpj55zSmlLaD30URdQRiHBqL+n18PBgrE2zdJy3D+8Yho9wPBB55S32rLWLxWLE0NIsrU4n6gjC7nq9HiPm6XQCwR/w7rl1mEyEEKe6Xq1Wm82mLEu86L7vrTEvnj0/VVXXtphb4iVeXV1hixiHdpzrdF0ngwDKiYjdhBAhhAykNmd1emMNoRTSHegjsQ0DfAmTBhw5cGcZYxBnBDSEOUTf95PJBD0ToPOrqyvclLZtwZ0d2sFamyQJ59ikcYGUnNMgiA77fdu2q+WyrmtnIS5AAhkVxREtb9PUhBBjNKOUcUYIwa6+8D7YhJz7+5GDD9ZQEkfWmlNdz2bzp83ub/7mP/9P//P/AspT3/cXFxeI1EBmjDbUnxPitQ+MMVIGzp11ds/UA2Pww6XkoI1FUcgYrE0k/FSU0Ywx4NWDVkopyYXxagjaK+tg6CiEFIJTSoUUNzc3L168aJo6iiOARUkSa20Ioc65sirn89n3vvc9tPIAk40xfT8IwWezKSEkSeI4jjhnWmtjdZ7nWA5XSmGfjjImpFBa51mmvExWEsWbzSbNMmR94BDgeuR5XtcnXD2c9t///veXl5eQAABcgWPTdd1qtaqbkzbGaF0UJaPUaDt0nRCy689IMmPsJz/5CeaCk8mEc9mf5YtYVVVd1+LCwSyPEAKJQ2vt/f39+uLCOme0iaKoOB6VUhbdRlEADlkul7gFh8MhSdI/+9nH3zU9/OMvf8Uok1IqBTPFs9kIDN2ACIG5gdAMqqj19qLMC7Jigwx8ZMwtgyAIoyBJ4jhOtttt3w9SntWWAKSgGiVe9gRHB+AyzhDWu9DlKb+CgIOrtTZGU0qEkJSy6XQaxyk+gBfIIDg9+FEg147EIfjenCN7FOEmG68dPzJQ8R9RF+OvAJVG4Ea7g6FFFEbelEpyvxfqnKOMAC8eI5rPr2dNcudlSPD10WDh9IMbh5ci5Vl6QXuzUuF1SpDw8HyKotjtdhB9G9FqtH2U0kBIY0wYR9ZabQznvOs7IYTwVFQ8AcBKSus4iYHqGGPALFoul3ACQG2FFpD7fb3j8QirW0wv0GYVhyO+I3IVBAOUUpSdFUGapsHUFNcS0TZJki+//BLX7ww2hiG+OF4uDsNkMiGEYMsUZN/5fB5CMoQzDIepzwHsLK0jxxHUqK6D0wg4CKNF1I9YC3h8fGSMrddrbJMwvxyDEh4zSXz4x8fHFy9eAKgcM+iprCklzrn5YlYUBaUkz/PdbpumOdJPmqIMMkBLyqK+vLxkfv3TneXOuskkByqF0Rdifde1CNy4g+P/b5t6NpsyztfrCxlE1tokzYqiOOz2ztr5bCaFFFwwepZoZl6pzHjhL6WUEDyMAiE5GCgyEMPQD6qXko90O+NtZa21xlgpJfULSUIIxvkwDCgRiCd94h+knCzL8PUZ4+v1BQgCbdtIIbuuD8MYDrJCcmP08Xj86KOPoEUfRREWmLBxdXFxMc4hcMCM1YSQZ8+eAerEzuNkMuFCCCmssQBnKKWr5VJKWZ6qEQKBFU8URUJwQhxgJeC0bdtixo6pPp7DGX5ME6XV/rCPo8g5Yq0Ng8horZR21HVdC6AyiqLLy0somF1dXis1EEKCQFprlRq6DlpSBjV0VVXIEEDYkjShjlBKLS6Cc1mWJVk6UgxGyEdr/Rd//iffNT38n//0L5SyMAyRovGCnXPGnAlkOHbITqiCQ28chMAE9BxjQ0C9OMRRFDlnpBTG2GEYjofjZDIFyRUVPcpz5lWbRhrAMAxJkgDpHqlBKLTxjsEMCwJJGcvSzPnF7CiCrN6ZdTBiXIjpCOUoEAA6oaYGB5d50SfUm8zvCmDea60FzoAiHSRLVP0490KKYeiN0VEc4iIR6rq+ZewcpDBqk96jzdqzEMjI+sDzDMMIAWg6nXLOn56eurMlxlkYylqLj4T/3/f9aO/cti2G9pSyEZdDYkDASjFOSFPuhUIt9Mw9Ccd4KwhCyNNmA6NvILadt6AIw/DNmzfKWz2PtQJjbLvdYhX8/v4eGTEMw6oo8SK0X18Pw/B4OAgh+6HHTA9LxcMwzGYz/Ebcc+jpJ0mSZ1lZFEiK8/kcT9Ja+/z588vLy81mc3NzM55Po/VkOg3CgFKKt4ZYjxEXQjDaL8wScBiAFg5++RkD+Y8++gjIGGIudrYBCo/BFFUw/gzuPDx2rDde5pQ3Te2cC0JZVVXT1NPptKrKKErg/+GcnUwm0Gzvuo4LMajeGF03J0rJfDELAsk4GwkdgFjBJUPGBbiBZ472lBI3DMN8sUySNE3z16+/doRqre9ub3GGMUW372hJEb9mjPQz/ij+juRUEAS73a4qj13Xtl2LEq1pGy5417XEUSnloBUulDGGUCKlzJJ0zKnOLxZYa4VgnLM4juM44VxAYnKxWJblYRiGqqoxqOv7bhh6CCKFYfDy5QsheBgGw9CfTpUMpNbq9vbGEYtJetPUnDOg36vVCk8JjG3QTAghzp4l/4qi4IxNp9NuaBmjjFGtFVBfxijcxY2n3XddB2FdQiiIlxibjQFkuVoMSkkhAxmqYSCEcsYE50Jy0A0IIVAJxGnM8ymYVFEE619ozgutDQpumPuCzdi2rTYmCkPsvUspiXVhGAZRyDmHhwpqMqyb/PRPPviu6eH/+qd/ERzGnMIYY4z1lAyC2hDxC9ybEflh74hvY5iD+hcEUzyFLMvatjZG13VjjBkGlecT4ReYcfLwZFEkInSCPwMmEjIzYCXQHkZxMc65MVpKmSapEFJrnec59AQZo0BsEBnHPhGC6ehXiNdzHSsOIM5olYDzUE+RQj+OrIksOJ1OR8MmROE8n8JXFb3CqJuGF4YvhXSCUFXX57Vt/BDr/2GMA0wb0VLE5ckkxy/FMvn4o8bfgooMX4RSJoSATAUSD1a6BONBELRdx7zBX9O2s9nMajNykTE/QCOC3gJPe6QYRlGE8RqehnMOJTDoatZaZAis9iwWi7f3D6guwR5GSzoo1Q8DoQQRfBiGq6srUNS030KaTCYgh3DOLy4u2qbBkUMjiKoiiqL5fD6SWBDC6roWUqZZBhgQJJBRKwk9K1CUyWTCvXjnOPEap6z4L+DdYXUZfEdCyM3NDeQ2kZIxl76/v0dsRQbCMdNaByI4nU7YESGEVFWZ5/liMbeWgJfZNLWUcjab+oPthNfpQ8mFn4NlHVSRuIzT6VSI86txzkHnDrcG0kOOkCRJgyAKw6goyizLppPJdDIR3jcedSjEasYGa7Q11FprPVhruq49nSpoUzZN/dlvfo1lSewAnk4nXJOHh7dBEMLY3FhrjMGObd91eK2AepCK+r5H4SzPmgUhrLyNtYTY+/sHtPJYAjdGaW2EEG/evMEBG20PmqbF+AdWJcA5QNuDtA8UT3HwqqqazqaMMUZZVVU4SMfDAdMXFF5pmkKuGI0I+M24QQ8PD1DZggRkURRoNZDmwyjM84xxXhblfL4YhqE+NdPJhFFGGYuT+IMPPjhPpAl5+fJlmqa///wPz54901q3HQSVKZgv6/UldE7hVAH1Pca5da7vujAMry4vZ7NZ33bGmCRLZ7MZHCBGZbk4jn/08XvfNT384y9/JTjAcQpdKkQE7MRyv2WeeC97OKhwLx6A1gG1ABIpogxSWXUq0T0QQqqyiuNknLWCPgQOKDBc/Efi59u4YziycHoCfwZh/Tx7CCTn8KfUi8UCeplIbNbacR91DFvIB6BboABkXlQKsDi+FEpd68UXUVOPmBjYn2iAWi8cG4ZB09Rd10ZRSCnp+w5ipc6Ry8tLd6aQn7lJ1lrnzkxWEIdA16OUMsYR0LFev1qtiqJwzjrnQI7GGi1GdtyLcSLFMi9L3vcDIWRcBkSq0FqrfqCUEnYmCGitHSGMsUBK3AGgQAAAi6IY1ADfFWQ75AYsdiIZ4CDO53OIPQBnQBkLrGmxWDR1PTIycd/QdtRNE8URWkbQxo/HI1zXtTepr+sa6y9BEBqtR/hRnoVsz620MQahGcgVWvLpbAanJnClhPeGG1VgAaOBm2+9NiJ2DFESgR7z/PlzWDl2XQeCw36/v7y8HBESSDvgzUKndhgGHDAQrAMR+gq0t9YeDvswDOM4KssTCpfpdFLX9WSSp2kaxdFms4GcxzD0cRwJwbMsPZ1OsH5CMzeSqZyzuLBIJPP5HB+AEpckMWX8xYuXaZprrW/vnnHGE0/pHid8Y5WDszQuygghgkCUZbHf77bbzeFwqKqy69q6PlVV+fqrL/u+2+93Xdc2Tb3bbauqJI49PT1dXl+VZTmdToMgIAAVrcOED9USqsyuO9usaq3jOGWMSRFaS/uuT9LoD3/44vHx8eHhYT6fQTyVEItm5fPP/zUMA3yYtm26riPETSb5arU8Hg/D0GdZejpVxljn3WWEEEgGXddZ56bTKTY3ceybU+2ciZOIEKe1CgIJwgjStvVmwLe3t/CGA9YCrQscYISFNEvjJNZal0U5nczyPD8eCvB0y6oQUgghICeKv1jXtda2KIosy7quDcOAc0YpmU6nQRDleQ4+CDSXmqaJ4rgfeqPNMAzoD5IoppT2ahilxUc+WxRFf/LJ+981Pfy333weyIBSaozlnGMh1jkHrj26TrQRYJeiJZde+AhFAbh6OK94oBgE1c1JSqGUZow1TZskaeDdZhANkQNab9Y6DqtxRQGbIEIRb/eIKEMIoRRK1zrL8qZp5vMlIWS73eZ5hg+Dj4eQRylFmMOVBkSAV4vpNLIFwgQImvjtIyIBJBr4BvcK58hAxOs1oY4jvhuglFLKIq+Ijm4XMc7as7kCwjQ7LySfuwfkKiDjZVkKIRijk8lEa73f7xEH0dJhSty2Le4evlHTtOgGEEDDMFwul6fTiVjHObfEIfgOw9APfdd1SRwjEeI6AUoalDLWYmKMpU1gepzz+Xz+9PQ0CjRhrAIqM6V0uVwC38eoNs+ywevR4mEO+Hd9tjdHvKPevAH7aM+fP9/tdtTvmUdhGMjAz5wMFn1RyIdh+Pr1a5Qa+/0eeFF1OhlrEM2B8Di/Zm+9eyDeFIbtSZJ8/PHHeCOIvGjy8BuBZeH14aPudru7uzsMzwF24emBf6k9/RQcm8PuyDkLw5AyMpvN+r7rui6OI+coMOXr66thGJyDkI5w9tw34BPCFbmu69Pp/H2Z33MkhEBNAV8NTfZ6vT4c9karLEspY1qb65s7KQPORRiESRgxQp2xDPi1c8Q5TlkUx9RrnuNXUEoJcVEU/epXv4J/MsZ1+/2+KorTqbq/v8fZA3hwPB4Ph6MQcrla4bNNJhNHXNu2RmloV4DkIs/yB1ypASC7MTaQEeeSEMo522yeAhlUVXk8Hu7v38zmM85ZdTqhoRmG4f7+/s2bN4fDARUtYIwkSdbrNdSutNZN06LwAsCLfk4bo40OgqBr2yzL1DAcj0cpxNPTY5ZnyFuYMyFoGGPAtsDlury8fHh4ADyQphlmTs45bO1YY2UgoHhRn5rlYpGl+fFwiKLIOgtQHQSNEV3X2j57dsc5r06l323UhNDlcoWV2Lqur6+vMcnTxjhCoDI5yXMhRCAl57wfBvjuGa+jhTb3z3/+yXdND5/97it+3qFVXddBHFRKCaqM894DCFUAkXH0cY0RTdBhjfGo6zo4ifZ9yznvup5zfjrVsKXT3lvUePMclKWo7JBpcHudc8Za5AbcAYzHgRIoNQRhUJ/qPJ/WdT2fz0+n+unp6ebmBlgBRib4nCObdrVaMe9JgAsfezcS58VQgS303jgXB2sYBhxxhDD8XRCZhBBCcEpJWRbEgxJ++BkhA0FU4J0hnkZ0Q8QBOoH0gOTfti1U3rqu09pYa4C2bTYbBCzQLtFLEULgT8vPnmgO9TJwMON5kBJlI3Gcc3O2L+0pIUgeQGAQ2s5igl6aN4oiVJcIgnAZQ0bHX8FDwEgQelOBV9njXr82CAKotoxVG+pc5CrtN9EAHwFVOy8GI9ESYp2Lo2i5WmF8farrOEmklJvNpm1bqGYul0utdXU6QUAMNUHoJaRw+elZkoRjKq61vrm9Wi2Xs/n017/6Df4YzB4wkwRCjZMPYASoGk4+apqqqh4fH7Mk7dp2tVoRz9JZLlf73aFuqjiOIFCf5xm8bDk/99Bd1y6Xy2+++TpN08e3j/P5DATrrmuF4LBz6HsF5avlcon00HadGoaua/U7UnQgNfT90HdNEARhFNV103X9bDYHKTEIZD8Mgxoo+eOiZRAETFBCXNd1VVUEgcSsS2sVhuHpVH/66adQUC/Lsm6q4/Gw3WwfHx+32y2M3oAvDYNRSsdJ8uGHH3775g0kvJwXI7i9vQXIhiEEISRJYimDqqogmB8EISGUMRpF4eFw3G62zjkp+ddfv/6rv/rvX758WZalVooLNqih67tjcWyaWnDOOauqEuIC6MsheFNVJ+1X3JHvKSHT2VQI0dSNlLLvOsaYYPxUVwbjQErCMGi7dr1aD8OwWq3QGY+143lb09ogCDAhR4MeBAFjtBvOFkNhEDVNE8gwCkPnLJdnDaHtdpvnOVaFKKXE0aurSxBtyrKazaaz2azrurpuGGOvXr3a7/dFUfzkJz/5/PPPV+u1c67vOkLI0+PjarXq245z3nidWhy5vu+yLE2S+Mc//N53TQ+//fwbQijnouu6YRi0VlEUSimYd95Ao829U42U8t0xL8YMWBEghCBujgMJSmnbdl3Xa6UppWmacM6GoReCJ0kspUAf1zQ1Y/R0qhbTeds02hpCCBPcOqetEVIIft6JA0TgnIMknBBcaxOGYdf1QgitTVVVWp/dqQYvRQDQA0UuOgAMPNEDAVPCnQcMopRijDhimPfMadtOiGC5XHVdJ+V5BzvybreEECECQqi1rqpOYH8Pg8KFpJTO5/P9fo9PhSDb9x1wcMwPENSGYWCMhmHQ9x1jVKlBCF7Xp2E4O16gbMGmBWAxQgi6GfxYD6yfhZqllBi3YgVUGU19cLTGqGHAJsF6va7rGkQsVNbGGGvMfDaDAmvbtExwSwnW45Fx8fXHbA2iJAptnASgwEIKxjnlTEhJKDXGUkaVVlEcj6kXDBYk1ChNjLVlVRHr3j48nPepKBNh0PVdFMd1U9/e3e12e0cJFzwIwyAMi6KM02QynTZdO1vMAfig0Q78/i1esfar+2iSjDHTaU6IubxcZ1n8T//3P1FGKSNlWS2XS2Mc56IsK0pZGEZK6SiKnXNd1242G8BcSI1fffUVtWeSLiVEcD70Q1M3ahjSLKLU7fc7xiiEycFrSNOMnGmy7fPnL/pePTw8KqU5YWmS5lneNm0URrvtbjadySh68/aBMiYCSRkjjCqttDZNU1ujgkBaa3wAdHmeVeVxOp1aR5bLlXOk7wfGWZ5PBqV6NRhjrLNd33MpojAMAkGItkZnaTz0nVa9FDwMhFbGKJJn+Ztvvj2VZd+1bVM1dVkcS+hX48VhIkIpZVJSTpUaVqtlGMjDfhdHoRp6fCqlhul0Utenuj5xzpyznAvob7Zt5xwJAmB0MmDhbDq3xpbHQisTCMEZ+/nP/+zbb++NsXEUHY7HMAwodX3fNdWJM34simHQhLGyahyhg9Lz2RR6M8ZosF2tNVIKrbRWajGfn6oKQUBIeXFx2TX9enXhDHGGDN3Q1F0UxIPSWpuyrHAMhZA4BoSQ06kIQzmfTxmjjDlrNSFE9YY4CkXYoigIcUEYFGWBwRUUvVAhXV5e3t/fgwyGsdPpdCqKcjqdTSaTtm2++urLrmvjOOq6Vms1n8+qsojCoKrKoe8wEAqioKgKShnq5r5vnbNgm1Jm//Qn37l7+Nc/fKv9riB0SwDVob8G+jF47YrQa4jijkHqBC9YesXE7XaLq6iUopQoLzMJdqnxtqAQV8CAFDvDnPMkSiilFpsBQaC1HtQgpIR87MhtBQofRSHjDGMAxliaZpwLUEcw2cbQf6y18VbG/gYtS9d1o/03pbRtW3DaKCWU0elkSohL02wYemPsdDoFokUI8Z6g8l26KvIipgJxHKMP0Fovl0vIxo10dSG48+xvzBiwdTHOezAfxmADDweEKxTv1BNqkcXxeIH7A9ECMIKGY1zUOhwOgIDQjCMuSy9jjvpAeCsrwBdRHMdxXJRlEIYMq3ZCknecJICD4a8vFgvUBFLKcYkaBTLgICmlsQafDwvGeNrSS503TdMNPdQRAimrsiKECEyP++54PGJSheEH9+Z3QRBYZyFi8+WXX758+RJUVPIOlx881P1+P/4u0Ciqqrq5ven79tmzuzhmdT38/vM/LBYrrbXWJgjCkbqGThd1EiZGONh40caYsijwu/ClgAU/Pj4qNUDLKIqit2/fAqFljMFeEKMXQsgPf/jDL774IknSU1Xe3NxEUfT69WscsLqu8+n0VNd1XS8Wi9VqNXaonNG6PsHRATDX7e0t5/zp8XE+nyut5/PFcrnOskmSJM6dtTUpIc65QEo8Z86ZHvqhH4a+l0JuN1ujNXHOWrffH+I4evv2AYVR37eMcYxPoV+U5/l6vcZtury6zvP87u5OCPHq1as3b96gqQ29FrKUcrlcfvrpp4QQkAJw1OM4qcpT3/d5PqGUaqU5YxeX69ls/u2bbzCB+7M//3MQBB4f3xJCuq6bTPK2bY02SmkhZVWd4iTpuyGMIhkEp6rE4pQx5nA4zOdz0PykDEYqHW4c8C7uXZBxTyGIQAhdrVeUUvxJ6Td5nXOM8aIooyiezeZCSGxVgt2OpSuceVzerusQ8QCIoVplng53f3+vlLq9vYXIW5Kkxpzl/TEhwxgc9uaQ8PMTYowzpUensaOKL0X//Offmdj62b++BmQMABGTInSF2A9C7DZeqR+1ofVWoMMwYPlIeOdnRHxPdqSoK/ETZrMZ99KV8/kcGC7i1Hmc4GgYho4ScC611m3XQQASoRy5hJ3tujTn8CAyeZ5TyjAnFEKEYYBYgHoW3QOaA8xFsXOYZRmWidCUMMaqqrq5uWnbFjOiMAibpuFcoDXBi0FwHMtepVSSJEFw9gcfEwZjDANha22e58BJMFCy1u73O4TOPM/n8znaC+PXCOxZqIOM0QH0nlHdc5xynyfMzgGuAdoOBhcgHURhIDZPT09IDyMLAB03HsI4SGCMlWWJtISWpeu6JEuiODbGLOYLEF2qqkJug3deWZaQ16XeumA8CWCDEEJGibSxH8f8n3hETmvd+xk+sW7oB+HldIIwxO7Fer3e7/evXr1CtsDdwxdH74KBM87VMAygNoHrDEwW42j8a9M0xihj7GQyTdPkvfduf/nLXw+DiuO4KMrlcrnb7QCaoUtAU4KZKsQwwHa9uLjYPD4BT41pOTwAACAASURBVDN+qR5qV8acyZT7/R4tF96Cc2SMTfv9HmX44XA4lscPP/6o7brpfPbtmzeXV1dlVVnrrHNN06xWK5ADUT04a9rmLKCCDh7hJs+ytu0++MEPjCFJmkkZSBmOZQFxxDkXYbh1LAhxRXEA1QJjud/97neTyaRtGqV033eXV5dVVUJWRCm9Wq3gjD2ZTLIsgweOUmo6W7x69Wo6nV5eXiZJMp/PP/vsM1yHkQXQdd177733+eefA1eE/IOUwSSfbjYbyCdzGVBGjLPL9SKfTLa73UeffHz/cP/JJ5/86le/mkxyxigwQ0JcIIK2bYMwDILw62++kTKom2Y6nTpjUMPN53NrrVIKNh5gf4DhgzsCsD5OEiZ4EIWH4th0bZbn/TBwwff7/dXVFcrcMaz3fbdaraAcgyQHXkDbdkghIPhdX19jVIaqDsAgEEt8BvDQPv7444eHBzDTttstZC6192UJggB8CngoQZJnJO/keW7t2a0oCGQURTBna9vmL3/x7xHV+DducYCDRooOakbutYzGwhBhF2kD88/WWyU7L/YZeUt3pRRuMki4iEEY3CHogLuNCZL2Cr2MMcqcDDgh1hhljBKCcU6tPVtGv7vRhlkFZQzNAaW063p79ru247fARyKeqYIvQghBWYE+CTFxnK5rrdfr9TjQRizL8xzqm+N3xM8cJxbok8bIiFAVhiHCHyiACEZY8xntgBhjs9kMdCBCiHrHuQX/Oj7SIAighj1iZdQLZFKvrYvUguS9WCzwrPCRrJ/i4NjhmyJ7SSnLssQIDimK+j0MoItZlmltkLHqph6Ht8avwqLUwviEEAIWDfoJ6/d4oyi6uLgYFZCaprm6usJlAwsgTdPZbIa/CGQGvx0fGzgG2twkSTAZBh1lfOAI/ciC40QHJw3w9zg+GZtIKeVue+Cc/f3f/71SZLvtLi+uy/LUtj2ITJClQm5wnvOKk5Bl2W9/+9unpycISk+nUwggoncMgmC5XCKaG2M++ugjDMOgK4yXgroEKe3bb78FE/fu7m63261WKxw2vG7Kz9M+bG7jk4DjC2lS/GR4yb3//vtRFNd1HYYxKhjiGCGkqmqQ9/AKwjDE7nfXdX1bN6dy6BqjeqP6D7733n/79JdJkgxDfzjsoyj80Y9+9OrVex9++PH19W2WTe7unr98+d56ffnhhx/P58ubm7sPP/yYc35xcXF7e7tarVDW/MVf/EXTNG/fvsWCJ9DUzWbz05/+FOxnEAVRzVxf34Ap2/etOi/MD7/4xX8UQl5d3goR/sM//MOPf/zjxWKJaBNFZzVPDAUppavVClPf+/t75CQgHKMdwHq9xpNHHMBkbrPZdN5eAjkPG6kQMMYeD8rKqqqGYVgul4wJSvlqdTGbLbbbfd+r7XY/mczu7u4wWAUBDNvRx+NxvV5LKYuiuLq6+vDDDyFrhku02WweHx/hYnA4HJIkadtuhO5xAlHJ3dzc7Pf7L7/8sm3b6XSKMarwO0kIDtiHTZJktIz8//rPv/Wa/u2X1lpwEvSZdm0gHQNeICgxzOtc4rOOw1tgCJiCQpdmrGSdcxDbIn6UF4Yh8Af4ammtsSuA0jIMQ+KslDJK4q7vZCjhkigEZ+QciYQXbmSMWWuIw4+nsHmJohgQGX4asuuY/M47R17/HUcBRSVjDEgUAlwYhmVZSinyfKK1TtMsSdJhOLsQw2UI1QQizmQycY4kSToCa5gWYpMTUAmiJ/eL6BhgIJqD1GytxWEyxiB7ocQAPoZNpXEmNs51xyYD3QAUXvt+AN8UyDsyH2r5MYvgPwLRwojbeG/Ud9MeIQQiHG3fpVlmrU3iBCr5+I2IVuiCAf29ffsWHTo+P3pQPFju96tBTMBmxjiBQFLkUuDO9F23mM+NtVNsblsTx/HohAGeG4AmlOQw2AIRYLw2wIJQBABKGgc2SOd4JhjAMiZ222PTtEIIRvkknyg1BN4GanwymKJBxdYYs9vtptPp4XBYzObcS6MzL2RCCBmGHmUN8ZikJ7xRPA3saSOEaa2VUVEcX15d1U3ztNl0fTdfzJumJYQURdE0DdIkRAECIcCCQxeIMdKLFy+++vKrQIaUi8uLa0eIFIHStms7RxylJImTMAzBe2mbpiyOSYz7q50DcsJms+mbb7+l3nAU2S4Mo/Vq3XX9bDZ97733lotlPsnTJE2zdLFYGOvKsry7u0P6t575hvIIm57gs6B0I4Ts9/vpdDoMSnDJOQ/DaLfbrS5WSmullZRSG/3y1XtffPHlixfPfvvb3/zd3/3dfr9br9dJmhwO+x/84Add20VR7AihlM3mcyEk40xrrYZ+pMzhqAA5x/4EqnJ0M0EQKK2dH1ICJMK5DaOI+CJ43L7inF9eXkkpsM0DiCWK4r7vX758iYCptb64uHh8fAy9BQXmi4DXGGMI5QjCyHDYysqyrCiOwL4AgoH103mzGTiZQ0I/TdPNZvO9730fosJt2zw8PCyXi6ZpOGd/+tMfftf08JvffYWpyEhEQSuA9SjAhYCVKKUIN9QbMoPThmaKc14UBWoTxFxjzGSSY90Mt4tSit2o4/EIbAHrJwAZgiAgzjpnwzjq+8E6Qykz1kgpnSGEEAxj8bcIIRAN1Fo7BzhFAv2HsRrnfNxRJISg5ETvggRGvUYerihoZ6hzZ7NZVZWMUSmkMcYYO5/P0Z1ofd51UF5SGKnIOfr/sPZmbZIc15Wg7eZ7bBm5VBVQAAQQLYkStbSk0Whmvp7/rB+gN0kUxRZFEQAJoLDUmhkRGZuvts7DCXei+20E4YEfSVRleniY3eWcc8/FG0D4QMuFqSv0kkiot7e3eDmUEjpuppsWUgLymlICGwd98ZHX6zVUIjhkMUYEPsjVEZhQTHVdj1YP9a8aB5un/If0M/0ojElD8t+2LSS/SEWMsSRNpZJt1wkpOefeu/lsjsOKszEJw1Hsw+ICKknMGCIzsdG2gRAylmBsUoWGcQM2lPLeuaZuZtWsKIoiz4WUIV5s2JED3rx5g0FWVMHe+ydPnrx79w4PBsNLNEx4KmT9MHpDQWHFOQfss3vcCS7fvn3X9yaE2DbtyJZ7NMqEEAgN27bVWqGma8ZV5Mj0q8US3xQ0V7j8jLG+h36vL8vy8fERYFeWZWjI8BVgp5BzTic6yzMfwt3dnfP+dDwiwBVliaYcKcp7j9XfIKIhZ4T87JNPPjkejz/9oz/ZbrfDYNM0z/OSc9n3pmnaED2llDO+XC4ZpcMwFFlOSBzakzUDBoYpJc4aSglof+/d6XQ6n2tKKOdCSrVaXb179y5Ls9VqNatmSZKQSPIsV4kehgFw3ITrSimx2AAza6gV0BysVivv/eFwyLIcwH2SpFLJzz//7MnTJ1mWWevrun7vvafb7T4E++mnn0BAlST6gw8+WC6XH3/8B33bG2OKsiSEVrMqRvL02bOu69i4cwWgNBDCEIJSGi0LSBEgt3gk591ms83SFOtmBzNoqXCjEcQBKNV1zSgXQg694UxIqaxxy+Vqt31knM5mM0IIqhZUq1AGSinh/Io1cFprtLmImehU3r59i63jMQZwpYgewDkQS93oPjutwHnz5u2zZ8+MMdYaSmkIfr1e++D/8sdT07/81ReAjNwPNn3j1CKmg3lDUEPPPmFKEPBOcBOqYzEupo4x3t3dbrdbxAW8BXxJkzP2RDw457jgWskQPBOcUiKVxLy+c94OFsNoqA2HYSDwKePMWUcpU0pLKTGTGcJl6SACwbStBbJCVPF+3C5Jx9XzuG9ACfM8r+uzTjQkCiHE1eoKmssQPMoiREM+LgJyznPOkYqAyGMP13K5xGwXMgomawghlBIIQPGesdYCPk4A9ICSI/ARQiZXKNSMADEgHZmQMeDySinOLzgMCjSkAcRNXBgUg5TS0+mE66HGRcp4CVLK3W6HFK60lkI2bRMJKcvSWhu8Rzs4jbB1407gMO4Igb9e13XwEmCjjaMfZ5txecA68nH4I0kSIWXTNE+ePOm7Pk2S+XxuhiHG6LxDmY+chN4ZWmrMZAzDMJvNdrvdu3fvphYbrRvKYUR85ONJkQy5Fxx4GGOEUMGxSE4NQ1/NquPxCMgefTPKJufcbDbD2Ua2CyF0TQtNBB6mHz1UILVE1ufjRpCiKMCWoYBdrVabzUYIMRhzdXOV5Rmcix42G8qYdfbps2dZmiHxuFF8TCk9nw7LxQJ6DXi4fvzxxw8PDy++elEUFWViNpspnVDKmqYdBqOUUEoJzpMkgeT0eDjMZ1WiuRzXn41FtE2zLC8KpRSljFEeI+H84s13d3f7+vVrzi9kHr59JjiUoPgJDw8PAPcJIWVZfvXVV7gyTdOAIQdefzweh8EILvM877peapVkSdd1zgfGuE6y3ePh408+/Zd/+WdCvJRyu900bRNC+Lu/+9uyLOfV/NWrV1yIsqyyPJ/N5s67sizLIgdYN8kXGWNpmoZwiWPo5C5ieudc8CBWu75XWrVdB10pgCzgPJzz5XJZ1zU2LuBTg9hDSfrV118WRXF9fT11yS9fviSErNfraR8zcIu+76+vr3GPAN1jwOXly5dleVm9hbMEbDNJEjjh46puNhtsqDXGKKXBWxwOe2ut924YhuVq+ad//KNNNf7j8xcTVTvVrXH0UNNaoylGUJM/WHmPcA/0hlJ6IbtG5+pIyGDM7e3t6XSEyfvxeIqRwFalrhvGGOcCzpQY+SORBO8ulstanc4n733X9TRS7zxwCcjM0QXDFQrBKE0zIS5Wi4j1Ukp0ysYM/dAxSpIkjeFSLyut8zyPMXg3tpPeG2Pm8zmCCGPYVhRCCHXdXF9fm7FXpZRMon48Urgsaq4AJVVVNXl5oqPEaFuMEStIrbVt20yKjq7rlssFY3y73apESyEY50WRN3XDGIshKCkppfIycnmwxmqliiIP4346xhhlLHjPOMeQIAA0yMmA/+CZsQcYLxNOAFgHnSSJxyLYti2KIknTruu4EAKdn1KM87o+S6UYIUM/xHH/No57P65NRdzHfyJpwaSPjH6CIKImRRzKNFDWaJYPp2OSJM5aKWTbtKvVylgLl020KXzcLodmCLEJAOmTJ09evHgBoAndWz9upUf3wDlP0rQ+n/OigGjNGKMTjWDhXNAazpdXQvDXr18niT6fL+v8CCFICefzeTlf9F3P4cfAOaNUK9V2rVZaSLF73BV5EUnkgqdpghmXyewLlQruSNf1nDNr7frm+nQ+JWkaYmCcp2mmtSaUkEi8D1mWxRBn5YxzHkNo66Zru9PxZI2dz6rj4XB9fR1CwEouY8xsNj8dT5HQ6+sbKXWWFZTQ8/k09G0kUQqBnMooozFyzr2zeZ7oNMnyPC+K+XyhtCqLSioVAtE68c6fzmfORZJozrlUsqxmq+Xq+5cvYySDMVolhNCiKJMkvb25+fLLL1EIdm0bYnTOKCXvbm9/+8XnaZoyRoe+99598MEHL7///ic/+cnvfvc7Ri+MsXemLKvffvHbTz75hETS9T2jnDEuOHvx4uu6rrMspQxT3+ynP/2Tjz78aLd73Gy3ZVlFQhaLpfPu5uY6hnCZ3CaEjFaMKJ746CmC7fTYjWGsabu2HBf7oCEbhh45jFLSNM3xeCzLwnv/1VdfP3363ul0ur6++eabF8Zc9JmcU5xAXEyINTabTYwxz7PNZpskGnacs1nVdS1klnVdY87myZMnxpimablgdtwtgeQEk2DEYaTw7XYL510hZAjeOQ8GlI0WO3/1Fz8aXPrnX/z6h1UDSOk0TePoc4e+DCSBHH0X8Ofd6HPACC3yHD8wkCik6Pq+6do/+PAPhsFAXGTMpbNDHhZCtm3nnMdm3CzLT6dziJEyTgjlXJ5PtbfBGaekMmbQWhESOWdKSWy/QWiOMRISlZKERCFYjAFTlM45722WJ13Xem8IjTTSVKfOeyFEOauKohic9TEkSqONsKMv9Oir4RljxmBz3JMYozEX7dZEJ6CC895nWY6SH4uaEZGhAYMtD9IY5M9d1xljldJ5XiglsbXbObt5fBis64ZeSelD6Lsuhhi9V1IxGq0Z0kSfTyfsftFKkug5Y5QSxniSpYwJJoROUzsYwCAgdSDnRVJHTuKjRjaE8PLly8EaIeViuWi7rh96xrmQMklTH4LSSicJ4yyG0NSNHczUOeHH4ufj1EIbBpwdDhkYY0bRgEYT8A7nfGKzQQyA3AohLGbzpq6DD0VRDMZwIZI0xTwdqmbY4OCbwnZxMLTouKFBQIUOP+123DibJMmxPi+WC+NslmcUq88YNdZivAMHLM8zyogxfde3hLCyKK+u1nVdU0KkUMMwmKE3Q+esSbRqmvp0OqZpkmilEy0TOZiBUNKbnnIaYuBScHrh3nE28NrTPCM0GmeEFJFGY4fZvLLOOm/zciak1ElCKdsfjmmWzeYL0/WSCyUlJTRNkq5tBefeOdP3nHESSVVW77/3/vFwDC7Mq1nTd9Vsdr2+LYsFJZxz9vi4Cd5onWRpWpVl8N4b662jhLZdK3M9uECYYExAS8yoEEIpIbx1GLA/HPeM09myUkonaZrmGWXs8y8+50JIrShjRVp456UQ89n8s9/8R5rooe+Dt1oxYzpC/Hq92mzeem+lZFrKtukWi3lT18v54nw+B+8ooYnWWvKyKL7/7nvMJTDGfAjPnjz77DdfeO8IDYQSY4ySej5f5mnBGb9/2FhrdZI4Z/MiM9ZILhB8ceTQCg/DMJ8vJkIOyjoo8rUSzhgp+Kwq6/MpBs8ZjTF477C4m3NelsUw9LNZ5Zz76usv5/PZMHTGDufzkdAolUBniXyAIFlV1W63a5oz9AT7/V4pqZREe03IhWvsug5GlsvlMi9yKliaJofDAfi86QclpHVuqimhhESEubpaCcEBhBpjtE5mszlj/L//+X/G0Pt/US6hcQYBgnYbL5QQAkU2kCIkQyBllFI1GrhOGhtCCN44vgzKKGKlHx1JMc+JV8Y5B88D1gjI2qULIWRaBT6N5mI+AH9mWtAGHgIiGeDg8eI0EPnoVYc/CQDBjWvHwZriv09sBII7UGl0CWE0H54Ujfi3QJPwfia6FTlvelp80rZtt9st+pXpVaNyEeOmdeA833///atXr/quY5TCUAyxsus7SukwXLLL1JIDl8TYnVQKIg2AhEDkwX8AWMNbxTUDMhZCwMWoqirLc1jqYtqREAK9MmQtZHT/x0tDFQOOio0GdtMEHH5L3/cffPABPLTx6UCioL3o+35ycJOj4SPOA2MM/TtwgOvra2zpmLj0USRGQRhMDBMdTW8OhwMYJjhv7/f7+XyOY4ZfdD6dYox5lglsQqa0LAp8j/iMTdOANsiy7Obmdr9/7PqWC5Zl6Xa3fdzvjDXH47Gu67quwU7jxCLo41/h58RxqnSaJQJ8PAm9svFGcM53ux3ep3P+3du3Xds6lN4+PNw/NHUN5ygk5tlshp0HRVGkaco474eh7brD8fjhRx+1XUcohdkDvmicPRyM6cRyzvf7/YUq7zspFWds+/DQNs3Qdxjb9s6xi8FwEEK0bXv/cI+oiqOyWq1+9rOf7XY77GpG543q6tmzZzCAopT2venafv946Np+tVynSSaF6rru1atXUsoXL15gaANyDDzwfL44jnMkQCy7rvvoo4+aph2GIdEJIbRt23/8x3+KMX700Ue44OBjML3kvX/37t16vZ7P57e3t9iFMJ/PcaJQ5CVJst1u5bjPEVEFw0N4mBjju3fvcLUnF86u66DBu7+/B5mE025+sChzs9nsdjvQsaiWsLUeOq7tdvv69Wtoz0BFPH/+HFKuw+Fwe3t7d3eXpGnTNIf9YbVaoQ7TSuMNCyH+8A//8Kc//SlsXfCiMJusRl8GOe5y/v/7z/9myfcZYisZZZSIkoAOkFondwH8ekQ3zEng2HHGlZT9gO2JwgdvjCWUFFkOEhhF04QDdF03n88hJ8ARJ4TUdV1VJT5eMtrAIQBRSiGsBuyOAh+ZbGL/x+aOTzxK33dSAgcz6/X6eDjBTpIxxgWfz+c+hK7rBLtIDCFwYqNBGwyT8QXgjDZNg0OAFBJGRxDvfd8PUirgmEiEdlwoCJu5YXT/h0wFWqkYI+cshNC2Dec8klhUs67rtFJKKQLqmIsQYogXuyE2miGT0WV6Np93XR8JyfOCUJKmaTeClZODIfIQngdcEZgYQOen+kwIubu7wx2L405TlN5TIkRcw4AIJhOnbBFjBGkxVRLOuf1+/5Of/ARXbsrKdBQcY5MzIQTfPpgM4NQThR7HxQMQoeKdo+fAIgcytiy40kjzeKq7uzuESDo6wDPGKCE0ktVySQkRjAfvlZTN+dLH4OKBQRVCeO/evHlrzOCcn8/n2+2GkJgkadfW9fkMvBhHEbxO13dcYHvCxScYIl1vL18cTguqv+12V81mQl7MpvCQ6/W661pnPSXksD80dTOrqm9evKCEZGnad50ZTbqQgRhjeZELKZM0GYypZlXd1JTRSOJ6vS6KUqtEqYQQaszQdW3XtUmalWWZpakQYvuwAYBJSGScLWazb168wDYyEj2JnlGWpGmSJNZa5z3jzFgnhFjMF2VZ4RlAIkKzd7O+4ZwbM1hrMaUMwsxffMnUKILSMRJKGSEU0zMxRCFkkiTG2NHrniZpunt8LKsqhMiF6Nruk48//uef/6NOtDE9ibFp2r43zfmcZdk3333b970PIcZ4Op9jjLfXN69evcJKdswHoGxaLlfT2M1l2JsxsMGgCWFuBk0miOgYI2bcwHIXRbFarY7HE2jn+/v7cRhiwH/Bpne0jECuTqcjApqUEiwjsrVSGpUTjihiY9t388V8MV9QSr/77rs8zxOdvH71quu7dqRAlFJ/+qd/CmkcFID44fhEAH7+8s/+8Memh3/+xa+R7ti4hxkJ343rOwghq9UKvQx+K2ZrEaTwUjhngnMuuFJKatX3vXFWa13mxUROiHFWzo4216igp3mZtm29d1M4QNzBpQLcHEfbE9w6dDPAdnCCx6EVPlqxuzzPGGPO2eVyeTqeCKFlVYUQQgyz2cx51zSNYHDFuKylU+O+ASSeH07nnk6nydAC0VmMS9lijPBIsOMKqsPhgE4TuDC+PKQHlJYQMwjBy7Ls+y6EMJhBKh1jlEIwxiihTdMwQqUUUDJAnIcUiMEoQkiWwcaLaJ1aZ4W4jLwj0k1OQShmEWEhVcLPyfL8cf84uWlOsw7gD+i4Ks6NruAQJiyXS8DoU/5wo1keVK13d3dI59Ovo+Po6Q/FV6g/drsdFPEQCEAfQSkFb0EphXMyGe0aMRuPugxpG4pDMrYUKCPQSeBoXcoOnVhj1+sr9DrOujRJkjQlhGDfr5Rys9lM500I/urVyxB8lqXn8zkEX1Vl8H4iolClglcYzKC0nq669369XltrTT8gME2vtCiKw/HQdW2aZciL2CKwXq8548bYrm2NMXBahBchA6Y0zrqr0WA1kri6Wimty7Jw1oUYCKFt28AgOU1zziU2V3PBjsdDluVFUXDGvPdIOU3TCMnTNFlfrbebh9evX5IQvDPBu0i885f1R1yINEuNNcZapRLOxcTYzWYzuOUroSilWHxWVeWXX30JfK8sC1SBuPWAHzjnlDJ05/SyiTpfLJYxEmtNNauEVC9evFiuVkqpGAkltKqKpmk2D2/h9RsjTdPssH+8ubnZbLfn81mgJ0jTzWajpSqKYrvdoh3B6NJ+vxdCTns+Hh4ewIFJKUPweOdN09zc3BwOB1QezjmIHTBFBLxU63QSHb3//vtv374NIeR5DoNhsMpkxGC998CjpjsLXKSu62G4sGKHwwG8YAhht9t1Q1dV5UcffvS4e3z58mWRF33XPR4OYF5RP93d3ZVl+fDwMM2ZPj4+CiGWyyV2sv3Fz/7bj00P//br36HZRNiaGAhcZohSIKaGjhN6SqB1uBVt28ISPEkTISVWz1PGCCFZkoYQ0MDiMKE7QWRHBoKcAHU6oiSywoRrTUId7z3EAHq080MExzeNv4XIhi5PSiGVFEJYa8qyPJ/qECKcKZ13s9lsMEPTNMF5oB94KhhUQKc8KfCklMDc2DhnAHIP/8paG+Pv5faQeKEMQe2AD4UoyRgDKAHxuxC8KIph6K21SZY4Fzjnh/1eCJHA0kdpzhkhl+COI4u/jpib5nmW5T4EIQSFMIlfpL3QVtNxxTkeGIF+sVgopexoywpwH172hBCkAWgxkVrQNJhxqx3G+oARQa5DCIEvN2aR3nvvPa01cEJw4BN4OIxrwKdtvag85vM5ZEKIRyg5UbhMpknoyZCNkJgR1tu2xeyrGNcfQTOO+gMH1VrLCCXxsn3TGotcnmjtxqlJIQTspxhjh8N+Pp+1bdM0tdZqsZiDQ+KcBh+89yD/0R12XbdYLq1zbDTuHYYhz/PlcumtgxCWjv8AYW26Vowrd9S4+rQsy67uKKGQ9gbvszTdPz5CwovuCi07vqM0S8i4PKDtukiIMcY6xxlP00yrZBis1okxg9LSOat1AuWl915wrqR0zoXotVKMsfls9u03L4ahbZra2iFGYqwPIfgQfAxVVUkljTGPj4eiKHHd8G6TJCnL8rtvvquqqq7PeZ5LKY7HA74v3HqtE60TQi4jpSFEOqqfnzx5Cm8CxiA1RpBJznXNOM/znFAaXFBKZ5l+9foVIVFrnaaZtZ4zutlsfvbnf/bll18maUII8SGEEE6HIyQMUJGGcYpwu90RQrTWANW7rjscDhAfA5uFIBBGOJB3E0JgjolKses66AUQiPDpoPTDTmK0ERMT5r1PEg2xGeQqOMlwEZ3kiBgPjDHOZrPtbjv0fVmUy+XifD4/7h4n2hLyPACV7733XlEUm80GZTTmBAghV1dXu93ur/7iRwtb/+e//xZ6/yzLUAmiPETbAt0LxjInBSQCExhX1KRSSCklE8IYwxhvu1YITgjhlGF8bOIb8AAAIABJREFUTFx2E4pphg6/AsELlePxeDRmmCZ7p9ozjj6AU1sz6gUjykzIV1DNOecY44SQpmngasc5D8HP5/P63HrnkyxVSjVtWxRFP/RsNKPHax057YvdPzpffGfoqxArJzQcGF+MkRAKaAJdXjIuogHsABgK/yeeHDzN8XiEKSbnrOs6pIcsy/aP+yRJEq2VUozSGKNzFlcRFBkyBCEkBM84X61WPgRnPefcOiu5QKSbjFIQVujomoVWBrR58N4FD1x7Pp+jhUJePJ1OT58+xdQxbg5ue4xxvV7jaiFMI63iHqrRGApqXaQWeG8gTWLGGO0asC9cxadPn/Z9j0YB+QZ3DBka498TYQOBKT4dLOEmSB0vHLA+krcYLUOctUIIZ60UQgqB0xu8p5zhaGESdbPZLJdL70PTntMs1Yk+nU/L1XIY+oeH+yzJEGjQUk/dIWUskjieh4tUbLlcCnZZy4H4DpV9kiZmtCNDe3dzc4Nepzk3qITQJetx/4SzFkoYDBKipeOcueCzPMPA18PDg3Muz/Pz8bRcroqierjfzucLSgkXiCEsyzICYpZz7AVKEr19eKjKKsuSvm+8c4LHGFyItB+c954y5rxXWqmLlJG+evUGT47KDIDqzfrm5z//OdY+C8HLqkR4jZEIIZMkxQn0PsDHO45mvZ/95rMPP/xIa825IIRwzoyxhDLGedf3WV5QSjnjwfmb2+vvX37HOKOUKqml1CTGt2/fcCmKotjutmVZciEIIaYfJjAHg5+APYWQF03wKGyllL5+/UaIy7gl/kHbikITxwxIGs7e8Xi01qF5xZX03qNGsdZut1sUQCCcOOfeX+SCwFEwACGlxOdFsp/YXOedUIIQ2jbN1WrFGT/s94zSNM+h+8CfH4ahruvr62uYhQBx3Ww2KA2rqvqjTz/4senhl7/6AvXLBCCOcSeEcXkh+q9JKn44HJDz2WhFxyjFCQZsYqzBfeuaNowuTHz0D5Hj6iUySqEwCbLf752zICcQnnDHMMOF2gqUMoIRGREhlLRmtEzA+cPcYMBUXQyr1coOtu+HuydPQgj7w8EYY6wtisIOxo2ie0IIWBYyankn2b4eHTLiuCAIGQKPF8JF6MV/MKarfmBqHUYPc/xdyGkwyZIkCaWk6zqpJSFMKRV8cM4tF0shRPQhxtj3HX4vQjAKFkJImibW2tlsobRumpYwaoyxw2W3xJSDJ8wa6D/GgJEMBmNQ1yA9kHFB5tu3b3HnJ8aPjLU8pfTq6gpVOW4OPvswuglNvDeuB2Ps8fFxugOIfYQQpN7z+YztOjgJIKsQrFEcIFAiAUy+gWB3kRtCCEh1KOjY6DCI2YhhtPqglDJC8Ej4n5djzDllF+d29Kbb7RbojbUDdtNWVYkNVyEESuisqtI0RW80n89R32x3W2hL8jzHk2DCGelhGsAErtL3XZqm1jk0EJPOqus6Ti99XghBcB68r6pKcI4WBEcII4R93xNG8zK31lZVdXt7ezwdkY9Xi2WeF0VRvXv7YIxdrZaUEs5ZkmZKqTi6WgXvlVKL+fLbF18H75bLmbPGDK3WUgk+GNsPIyweQ4xxdbXK8nwxX202W4x9gFJG3CSBSCnfvn1jjCmKfH29tta2bTv0g3c+TTMcD0oIBimMNWVZVlV1OByzLJ/N5rApdM4yzrHe+FzXQgjOBY3UWrtczt+8fV03p7KsQFeURc45a7tuGAbrbJZly9XSWquEhCHSZA2HeB1CZIxdBt+GAbHFGLPdbnBl0NG60XUUXwQuFCKD1vp8biDLhjAPvwgQy2SPhkoRPTrcA2OMmJCQUj558sR7v1gssVYIXS9Wu1NGffTe+6auOeNFnkshh74/n2sU6GiUIU2End92u4VheJIkMIdmjP3Zn/zouYdf/Ot/xBgl54nWzjqsB4kh9sNlg+akUkARl6Yaq34QrCH8Msb2A/puTighMZJIyAhlIIf7cZEOpTRJdNe1aE6tNWmaUkr2+0cEr0kjNLFGIQTMrE6t3FS54CsUQsByB2QvIYFxFoKPMcAWuygKymXb9zd3t8aapm0JiQhhNBKELVT3mEzO87y3RigppGzaxodgvQskEhKhLp8CJUKDlELKy/b28/mMxeKURhQdRZHDPRDQIaWUMcIYFYIb0wNGp5RSQijjkvM8y0gkeZZJKbu+o5R5Hxhlzrp+GLRKGOOMcetMkiaREK2VVrppa2eH4G2MMYSLFmsCDKdhHPA0wzBgZ+QwDJTxGGLTdlO+x5kmhLRtK4XEkdBKSS6CD0KKSUsK0EkIoYTs2i74cLVaBedjjF3TrhbLQCKldLfbIesj1uOaidFwyXu/Wq2AnA7DAJ0f/GdQwUEZhVZj4ma997e3t8i7ZVl+8803SqlRrUBDIEgzbdu2bW2ssc4GF9CDQpWL8RRrbSRUax3CpdFp245SJqWo67MU0jvvrLeDGfoh0cn5dAIOifjCGINFmpTifDwyQvOsaJpaCgy19HmRHY8HQmOIgTJ6Op+klk3TMC4YpV3beh/q83lWzdI0EUyYcXNi3/dlWVCK8XjgZgRjWVmWFkUeoy/y3Bqbp2mRF1VZFWn24usXnNCnT5+0bQOxwG8++81777+X6IQzURSlEGroBxJJiF5qKZSAD3HTtoMxi8XMec8Fl1oX1SwvihCjscaHwChXQmqdWmvLsthuN3VdG9Pf3NwQEtM0ccZAtfn4uPPez6oZR+IQlzU+GEHlnDPKpJRFWUqlpFR/9Mc//errF7PZXOmE0EjglkNZjOR0OkmltE588JTFNMuapv3u2++VlEWRhmAiDWVVtG0tpbTWWOOe3N4tZrPz6Yx6FF/TbDYrLiq1y65WfHfo9W9ubs7ns3M+TVNstyzLsmm6LMuwrsZar5QyxsYYkySdhh9h2DfJHx4e3hEStVZ9351OhxjDYjGvqnK328C7G+uy9/vdhx8+L4pcqQQ+YxiWQic9mMEad7W62u0em7oeMbeybmvv/d3dXYzx2bNnUOjd3Nzc398vl8t3797hjmB7o7X2//ybP/ux6eFffvEfggvBRYyRjL02iRFDUsm4iof83nUkQxwnJFprKb3MmxBCEfpRG3IsTPjBwh+UPIj1lF72nXVdh7UN1trT6diN2gzEXyhewPOgHkQbO0EHUxsYRt837z2hMUlVVRXYmgDZZZplPsTD6TSbzx0sJMeNslmS4ifLceNbjFEoGell0hJZxznng6eMOvN7a9909McG20YpyPyI1ooxNj3tqKeKWHdhzNC2DecX4bOUsijK0+ksuMBkLCpx6Km0VjSSEGII0Rqbphnez2AM5SxG4kOYzWZ913Vto7RilHcdbAovReukclPjrj1w+wBnGRMhEqwgLYoyhDAMRko5DH3btkpIrXXfdc7aKi/6YciLAj3c4+PjhOrQGIe+v7+/f3J3570n4zSps45xDpkTPikODJoqJAlY6flxym8SFuMPoATT42o8Oe6TABk4XX7MG4IB8o4472KMSsmu6/qhIyRKyaMnqDxwIIFEKa1DJFJKEuHR2xljvfM3NzeH3U5JJYX0o/sso8xaW9c12k2gZ5cDIHhT1z4EMxhKafCRgWbLUiEFZTxJU6XVuT4LIbjgphvyLD8djoxQOxhKSJakeEsgeIdhgE0CziFqoKLIkyTx3sF1BxiRGQwlhBLy3nvvvXz58nw+z+dV2za3t7eM8V/96t+ePXsvz0uldIxUKx2CP59rqQQXjFIqlCQhbLZba4dTfVosl8Z7laRcSkhrKGVaaWudlOp0OKVp2nTtarXqug57A1EyY9gty9I0TbbbLbp8dEWg9EBUxNEEzIwyudls3rb9ZrtdLFfWWUYJF5JSFkM01nLGhZCM0RC8FGow9uHh4XQ6pAlLEjnYXidKCNkPHWccrAD2G1prIQIGxgUirWnaOKrDceoQporiYqu8Xq+Nser3i8sSUOh5XoQQMKoF7SzKFFRaOFFd1w7DgA2vw2CUkoyx6+vr8/nonccpnc1nJBJC6fP3n59ONaggzPNDzei9r8/1Yo4tgW3TNDpJuBBSCpDbMHThnMO8DngXdlaHEE6nExYY/x9/9aMNvf/1V18EEr33lNFIiHU2EkI5Rz5AesBNoJR678Ddo+VHZco5HwYjRnk1wiIKQ0IIssXULOPnQNsAvAwVJRoxWHeRcR89fjXwpSnT5HmO9gWiKdhXIK9M1B/noiwrxJeyLLVOCSHW+sPhUJaV5MJZV5XlAJ+JcZMoG80/KKUAHLz3yGdhXMxb5sXQDwhkSAlkVMuQUctECKGUwl5JjTv1UKejL1mtVofDEaUKpTRGgsHAw+HwwxQFWxhKqZCy7TtjjU50iDFJUx+8D14piU4rhoi2tGkwlJABgbmkas6hx8Bvdz+Y+bi6ugKhDVACUnoMpWO61TknOJdCADlUUnZDHynBnceKNLSGzjml9bmub29vnfeCcx+Cd67tO0DMINXBLkDYBxQebybLMpRRE6EHOAtvFWofgHuoNvDmvffz+Rx/EsISoDrOeqyxRHrwwR2PRyF4cHH6giaNHCb8lVJSSfDkjNFIIudMKzmxx3iMKcyho4X+BN9aVRbL5XK72yVJOl8sOBfxsp/ugoAJIVarK1gjeO+LLAesCvQVknkECBy5OA4uAONdLpeQeCyXy/v7e7SD0O/jyrx58+bq6qqu64eHB+/d06dPCaFlOfvVr/69KMr11TUEo5RSIXjbtoPp0zRBoeOtPZ2OlJLtbnt7e8MYN4P1LoQQKWXeezgXAGZxzgklkZ4ZY9vtFhCTD9GHECJRWs/m81evX/kQuZBmuJgvIZiiVLLWYq8tikuw1rvdbrGYM0qCD5FQH7yxVkgJDJ8QkiRpXZ/fvHpVN8flalbk6es3b4ZhkFJ577MsF+IyzIsZY6wizvMcehAhhPfBXpbVXzb4gs5Bx4Y/hoYb2grkCegvYLqDSnQ2m4GI9t5jv3oIoW0bdwEMyXq93u12Uiqt9dXVqus6rfT7770fQri5uWkbHNoCaDauPJpjpFWwF0CiYMiRJEmSXMAb1CWopcI4TnR7e0tG5UKe53/9lz96avpf/+1z/AKpFGXM/MBQOoSA50NwROY0Zpi8GBGDGGPWoi7uQQXHGKSEnfKleqU0xhg4Z4vFIsbovUMJRi6LNS52oagNkyQBMY6ZSVDZOJqAX+GZqkYrcuhn8FcIIdCSV5BLc471nNa6LCussWoEbS+WDISocZ4OZBQyTSSEcQbQSY6egwD7+rZDQ3ppkxmDXhM5gF22y3UT/YDyBEwDkt9yuTydauDqjDGtU0KolNI5DwxBjA6j4L5QFGBAfDCGMso49yEUZamUllJZa9M0h4WZtQ7rQZAp67qeTD4A6UwTD9Ble+/TLD0eD1AohuC1ln3f9X2XJIkdGSY8PKMsUmK9myYoD4cDkKi2aVAYQv6EcDAMA6EE9xCtzPl8BpoEHDZNU2hD5bin4dWrV1proHD39/eTPzYdV9tO83c4hJiLRPIDXSyE0DoNAcZQgVJq7DAMA2M01VkcxWZ4D13XMc6NHZxzVVXChtOYPk2Trm2L/MKW4SHjOF+J64DHRkdrjJlV5Xw+f9zvAa3HSBhn3vumbpTSQsiirChhMRBrbNt0VVGgAkCORM0BxgIRH5wfzh4GVyeSrO97AAgo4CYaHDpsYwz8FT788KMY6OPj4/ffv1ytrtI01ToJIXDOGGOH4x5LsJ1zpu/q+tz3fdd2Qkq4v6B2cRaZKiaJds4ppShjgUQAvHCB3G63nHPnPBcCT5ik6Xyx+P7lS8Y5u0iVLkwSYEbnnIFSwDlCaJKks1m12WzbtinyTAhJGVNKGWvBbwI3TpP0fD4Pffftdy8W82qxmO0Ph7ZtsyxPkmS73eV5gYCG6IEGEZwNGAJKGZITIQTUEcI9oGbEZXzL8/lcKYXjyjlHDgAbByINFmpxHJDGMgw46JRlWZallOrxcc8Yu729lVLWdfPs2ftaa0q5UtoYe3W1DiGAqBhGS1CkijzPJzU/1Js3N9eYVIVMAw59017SyecNGaKqqv+KuYdf/keIIcTIOMuLwjrbdh0XHAWMGndzT4V/2zbgWKaMxxizdogxhOi7rpVKhOD7oSc0OmcH0ystCSVccCH46moJkgBR1Y27B+jocgoCAHcYM/EXSmOsIqHN5+NaBYhngEXgzzDKpJCLxdIYQyKlhM5ms+PhxChrmoZGMqtmbdMgKwTv8elwTLXWSHuDMfPFAnGZXnxh4zAMs9mchAuYgDs8DeUiz+NJAIAge6HPwNNCSjSbzdCuEkKSJCXj3JkQYhh6hHVI5sU4zDlYE2BzwhnjvKoqHwP80rMsCwGs9aUHwjR1HJ34YAY5DAN0B5gGAvHQdR1lbLlaDuMeU8boROTGGAmhjDGtVAjhfDwppRjnUivUVqhYcQMJJZHEEGPTNJ/+t08fHjZJkljv9vv9RONPSmi8N+hGEAdxk8GOoAgwoze4GRfBQxaFH4LESQgRQkw3HIkqTVPvAueMc962jff+dD5KKdu2uVquJxYdDAca8xCDMUNRFGVZCMF3u11ZlonWu+02TVNok0AGiHHvKU7vxFpxzk/HA2OsKMv9/sA4o5Rh3/X5XCPDgTwviiKE2LZtcz6v12t2GZrJvfdQbeG34BSB/ETzpJSa0pJSCoYiQDnApqBue/78+Zs3b/b7RynlYrFkVJRl+U//9M9Xq7W1VutE68u24BD96XQsy7Jtm0RJGEoLIY2x1+ub4OPmYYd9ebiSEM51XeuCT7McECvWAW2328PhMBhblDMfAueSMZ4XhRTqd19+mWo1QtCXDVfQPd6P+6n6fkiSlFJWFMXpdPTOJmnKGHfO98MQYtQ6wWlPk7Rrm8fHXYzueNg9eXK7e3w0xmCDk5Sqrpv5fA4XL1RpUxCjF5sDgYiBzn6/319dXcE0EzAAOmyIOXFNlFIYDELrg3XQQgioKsK4Gw3BSuuLjAIDxYyxruvRjkPtcnW1BpmcpukwmPV6/erVKzZOF+FRYcuxXq8xVo20lCQaMQQ1nxsN88kI0qAryrLseDw2TfP//j9/859ID+x//9+M8XGAFnVrGAdWp6G2sYe4WCzgSuP/J4QIKfuhn3AhQgg2FmGIlIwr4MllzZzAvUJInar+ruvwNuPoL4+WYhLq0FEbilCCXzQ9Axknb70PzgWlNOeKUk4IlUIxJsD9ItZAbiSFdNZNdoQ4uIQQuHTggackhN5CjhNJCMSTznUSYiG4y3HVz9RSTP9qfA+RcymlFlzFQIMnIRC0sVgZ1DSN9x4aUGxu0FonaYLeIlzcUn2MJAQSI+37AR/c+8sqJELIJJZHYOXjwvBJhIrM54ObzUshGWUkRO+coYw4b0MIQooYo7E2TdO8yIUQGBxBAXg8HoEaTba4SqlXr1/VdZ0XOTxWQS3gXmGlARAz5IyyLKfa/PXr14CqQQWR0RltemkTlWJHN3UcBj8OwTjnPvjgg81mgwCEJI03qccNRdC8AQEACiGEiMSH6E7nw2B6QmOaaetMmiWILAiOGL6BVThq9q7rQADgZfZ9DzMPNARTu6Z10nV92/aE0LpuQyBFUT158gStG0Qsi8WiqqrpIIEwRz3Yti324uExYPevtf7kk0+Au4YQDocDJnXxDJMU+NtvvwXD9Pz581evXp3P5xcvXsATDNBikiSwFB2GbgQehbX+eKyvrm6KovI+ap12Xe+ciyT44I6nPUhpMtrICyE++OADSunxeN7v94TQLMsQ+6rZ7L333nt83CNiotlCHMQ1ORwOQNXatkUuf3J3l6bp0A9sFAqzS1IMOAkhxqZt8jxP0+Trr79erVaQXMIHGtjDyNMUMcZutIVGfm2aBi/QWguLdewOmpraGGNZlkjYKAuWyyXcVb33p9PpeLxMVOCqItPgLM2qxd3tU0Y5Jayp2xDIx3/wE8a4Ma7rBsbEZ5990bZ9lhVap9999/L+/v7h4SHP82+++YZg5w0hjDHwOpvNBo4deJ/OeYAZWut3794BcoRkvKqqm5sbTOqBZYF79H/in/917uE3vwOglOd5NZ9FQk7nUwxBcIGLgSAypd+ua8EcTjKSqqpiDOfzMcYAIZMQHOQMIdE5p5Tk0CkrkedZXdfHwwkiMOQG4CeIuUCrKKUYsAK4NmV+PVrLoXxDykVvhauolCKEee9Xq6sJ/rq+vjmdTjREZ+35dOq7njOe6iTR2jvXjQX7lK4YY5EQpRV+IERTZNyOywjFXgeg5OC4pkFZctmCW03FrB53DaJ1gNAlS8umbvqulyO0xRnP86xpazqqhFFConK0wR1PRy54iIFQwjjngnPKsUDBe386nfE8ODTA6KZ6syiKSckH9QjugHPOeS+UmM/nXdcZM0DQNZtV8/ms63ozDMGH4P2lcu9762zdNEjVuBiQyXoSIyXL5Wq724GJhWYpzzKMj+EdAp1A6Q3BEkoBKeVms4GVN0q56Yji76I0Q7OIV42bD4h5v9/D8O76+vrt27dpkgHnHIbeex8JyKEYPVksFpPGF+dHCGGd9cFDWkPGAe/FfO7sRf8+wYZg0e2410RrjT1fSZJYM1BKCaVKacqYkloqhQlbxn6/fhFAaJZl0fvDYf/s2bO6rpfLJUAhfNIwekBNiCuAkdlshhrfWgssDhAZ0L/1ev3u3TtCyMcffzwM/el0atvub/7mb9u2K4ryu2+/V0p5H47H49OnTxhjfd+tVsuHh4fz+URJmC9m9bkVQpbljDNZVTNK6WazSZIkxtD3XSRRSu696wfbdj2mc4HbAIqxLnz3/avFfCGkkko5H07nuqoqSsNuu8URgk0kbsd8sdjv92VZXl2t0yQzxhpjsyxllEDYiqxAKJNScs6EEHmWn8+n3Xaze9z85JOPXr787umzZ2ma3t8/wHwCRh2ozAAxIS3hrFZVVdfNJLwGLQxtKOZpgOcgEJFxLSClNE1TGJWTy0wVgzMCukmU87vd7snd081mc3t7u9ls0chaa5+///x4Onjv4QwEVQgMqd6+ffv1119XVYUpUcjhmqaBVdS7d+/yPMe22vv7exxmMBN83GcTQsAWCtCBk5JTCPE//u+//rHp4Ze//q0b9QOYbn18fGSc5eml60fBBRIM1Dk+4YVPuKyN64fhEulQKSMKKKW8d/KyZuty1Z1z51ON8m3iiGKMcD5Ai4fog2ITI0UTzkPG0VkxLrRBYsDdI4RIqfDbcT4IuWgBvXXoWrq2hbaBEHI6nQZrpjlwfP0hhDTLQgx8HIiDKgZgq+QCNTh+BQ4fStQQAm4LJApZliHS4XWJ0bejKMr1+vZ0OjXNZaCGMRZC5FxYN2CvCyIIAlNRlYTS/X4/Ud+XHi5OvUgIo+cEGw0/kGMmZQX4N4CkY5d9WbMqFYdph/eYSY6ERCmVMdZYG0OMFzdscT6dAiU++ClZgvgRQrjgpxLbOgfWLsuyRGlUNOgX0aqDRgohXF9fY5wQtbxz7urqyjkHizokkh+OF4lxZM+OU5Pn8xkUCwblEMG7tk8STHuwpmng0lhVJY0M2RFV2GKx2O12nHPrB2sRmlVR5MYMIQQheXABNakYF+0BY/GjIv76+hprTauq8s42TSPgKUQiidR5YKcSSn/rnODSGNN3AyEkegvTQDJClBPthyMHYAQho5uWS8eIG4FfioKpLMs3b97Atq+ua5j8dF3nnOdc3t09IYQSgiGv4csvv/zJTz7BC7HWLJfL+/t3wRkAI2maWWul0FKqqpqlSUIoWSwWSaLzPM3zLEm0TjKYL81mM5ThiAmLxSpE8rDZLBYLrRMoR511qVZKymEYFosFLjXyRDmrMGGjlE6S1FqntSYkckYTnTgf+r7nXKB9Y4xxzqRQ5/p0/+7t7nHjXL9er7743W+llG3bWWtXq6urqzV6LARQ4CKQDooLL5JihAWhBgcVldPpdAKugIpzGAZUA8iChJD7+3s0DRBBoDRERYjWvGv7q6s1IYRSVpZlXTeEkDzPiiKH22NRFE3TAItDIQKRDm4xGhrG2DfffIODAcs/XFhoeSDKAGeOyZvJghCWNviYV1dX/wWmGr/45Wdt22qpvHXz+Xzo+sfdY7AeM9y4A4B6nXOr1erdu/vj8dQ0bSSBUhJiEIIxxglhIcQYiLMhS/MQiFbJrJq3bc+Y6NpeSEUio5S3TYfzpJRC7YbxK7TG8A1FVseLkKM5DwDc8HsRVBoCoZRJqaCviJE454dhmMyCyLgx6nw++xAGY5z3zntC6Xy5sM5Z72ANi9eKPEEISZPEOZfoJIaopLLGaKmcdcloX4jYNxUawASstRiSQrA+n8/eW8aQ5xmUqSRyzkSM/lyfvHNCMOctZVFKLjgXQmZZjlSRppkxtut6RnmWZGaw0UcllBssjVRySWIUlDBKBGPWDEoIEgKJgTIKuTC0gJyzJNGMUXjNww6dc+ac5ZxBsXY61WmaBR8opX0/ZFmuVcqoGAbDKHcu9N3AmXAOC1wDibEsSiXV427HGE/TdOh7rRTMUI/7w3w2e9zt1ldXSAx0dJAF9jJtTEP9i/gLgTLmzNEA4RvEORRCeOe1Us7a4H1bNzSS+Wy2f3y8urrilDFCvXOC848+/PBw2lNKsuyypHcYzHy24Ewur9Z9b7iUztnH/SHJtQuuKHIpZFM3V6srSmjwgRJmjY0uIq0itZORFUegATO83+8//fRToGFJmn373fdpmsUYtdKCMzMMiVZcqm4wPoQ0y+q2kUo1XUsJKYrMGvPtt99eX1+jfQcCNlGRiAW4C6gAsIvX/2AHgBBiuVx+8cUX2F2DQf1hGIRg0Olvt49//ud/4X2sqsVisYQN0Wefff6Xf/nfSaRDb0Ig2HthrCeECyF9CKurJRes6xuw60rrNM0Yl0qlWV4tF1dt3Qx9LxjPs4xRagfjnYskXi1XTX1qmpqSmCW6Pp8YpZTJLK90AKfcAAAgAElEQVR8iMfTqZzN6rbmkgulvI9aJ8+evffVV1+trpb90EnJuWDOe6kUJoq9C4xQzjgJUXDunOm79nB8tNbsD8eymnPGj8ezNbbrTJblIAy2223XNZSSw2Hftl1VVd5HY6xzvigyQmII/nw+gWwTgt/fvyuKHDditVpaa87nU55nr1+/tHbI86zv+yS5mMOjnyuK4vb29ng8AgAHI9v1XVlVQnKlJSGBMsIJ4Yytr9dv3rzBnDNjDDApKm+Y9wBIRKTCIPTxeLy+vj6dTvBxyrJsGHp4fkxVMsLmpHuGqg111el0+i8Qtv7il7+hEJISulwsQwibhwdgO8fjEWUvCmR0ao+Pl6EHKXiSJtC5aZVYY0m8MAGT3gnaFTZuTIyXOeGLuA1OfEA5QK+RcVTq9vbWj5axyOSgH1ArAZdQ6rJNeuJzUGJMngooHOLoR805B5sHIhExHYAVmhWU1dDkcM6tMd75GIIUkrML1oE/jAIWiWQSnKDS16ODbFmWm82GcyaEJIRwLrz3JDLGmBByv38Ugtf1WUgB01YyUhfImhMwDZQJ0IqzFnYIeFTJRQwBYJQfV9cxxnSaTMW4GZ1lJ9ofv4Jd/Kl8IFEIFXwIIZZl1TadGSxWptRNIzh31mGBmlRScBZDiN4775RSZVV57xlnSOTI9JxzeGLjAODkoMMjI5AohAD5DGJ5gkcmqHASRFBK83EVubNOClGfz5wxb53kAvgY6oZpOJbjzwgBTREh5Hg8YmHvcnHFBVw6wuP+UWkppayb0+36LvjoXSjywgyWRCK48N4niZ62sYbRLH061ZhCklKu1+s3b97gSqOuxIAenm2xvHLe9UNPKUmzNMZASIwhCEbzLAshgLEAa0JGLwDcC/sDm9tJsnF/f1+WJarLftxJBaeHp0+fvnnzBrNpzjkh5Lfffv/hhx9RyjkTEMP89Kc//Yd/+If3338/SRKlNGM8+LBaLnbb3STxZIzhSh4OB3qRUZLgg/chTVKMCiul7u/vJwFYkiSMEiX58/efffvia0oCib4qc2sGxrXWerW6Op/Pb9+9yfMML1NKBf5psVj8/Oc/f//993EkpJSCC+e899DrUdxlKaX31gd7Ph/P57OQvGnqZ8/ec8475xFbsZ52Pp89PLwDpmetOx6PEI5XVXU+n0B1TGZcCE1xVL5hDS1ChBDs9ZvXdPQLQGRHrbxcLtfrNbIC/uGcu+D3x/1quXz16mXXtkLwPMuE4Eprzvnj4+PpdCrL8ubm5ng8itEAFFrKSdR+OBziRe9LYI0MMStUowAqYJmutc6ybLvdAgPEMQBnPp/P/wumpv/pX/4dL2I2m01IHBj/yYIJQ3396GWNq8sYJZTEGL13s9nc2stuZMjapvCEKh6RDu0z0h3qIzLKN4Hb4C3gpdDRhX+6G4jCEJ6GEKRUiIwIzcgfE4OPbh1xdhKr4N0B/5lUs7iEz549izFCWoCfgwAHwFSMzqxinOeYBG1m3BOJ5D/xrijTrDXOYcoJNtceArgkvYwuA6QKo7UUyAmodcFwAANB6Q1FDUAwzrlUcjADYRRD3YRR42wgF7E80D+Y2anRmx2I3ITwINNwfhGJYsgWHy1JEu+DD8HAKZ0xRukIAfOmaQkl6J+8d0+ePMEbRmuFlICx5JubGyEEeD/EVnwFoKzB8YJVhtZ5atgnBgVfunMuhoscQEkZvB/6gQtBGT2fzyBCCCFICW3b8tFCB7KTC7lnvU60976uz1JKOJo0TU3CRRGnLytFiRDCWgOlIBp5vEm4qkFeBe5ku91++umnYEdDCLA0QL7PsuzxcTefLwXj1phUJ85aJaU1Jkt119SggkAO9X1/d3eHH+J/YC+GuIO4KUZ3GWisERHevn17d3f3zTffQI0aQtjtdmVZtG07m811or/+6uu/+7v/63g8xBgAvX788cd///d//8d//MdIRdYazi4+bofDAS8ZA4YA3HF/ydjEgBJHCvntb387m80YY23bwlE4SZKnT59+8cUXIQSMC4QIVz5aVdXjfnc8Htu2vbm5lfKSRCmlq9Xqyy+/BJKWpimh1AyobNh4u2mM4Xg8HA57pcT9/bu8yLquReuPWAFWSWvdda2QAm8DRuKMcVS3oFs452AOUCMi8WM7G6AnDBZorQ7HQ9/1aZrleaGU3m636ABg3YrjgdBHKWVc1E1NCamqar/ft22LTVM+hKurK3h4vHv3rq7r58+f44dACA7X2yl8YY4BrxogPGOsLIuJjwTAiLiBozsMQ57nY+8oGGP/BeASlomi30dYxOdHasWZAAIAyNV7jxF/pWQkUQhOKSmrqu96MjLvhBBEeULIVKiC66ejphAhDNQ0XgpOGP4i/EQRkTHbPFW7blxRlyQJpRdjPlAUdHTgieNUZBzN0cBqArBGUEMgiDFiEAysERxDJ1ALgRXxVIwueHgGBC+Uveg24jgtNT1DmqbO2b4fYowhRLwYfHNKKcpoCAHi5R9mQUTMiZ1GpYCyDl/B1BVlWYKnAgaILwsfGfAaiprJuxS4zQ+/oBACF4Ixjh5oQjbpxdtOemdJjF3XMkatHaQQMMtknDvvKWNlWXZDDxAc3mRIeDhC2+22qioY7qPaQEcMmSYy+uR8CbEN6v0pGKGMEBfv99Rai9WqgoumaSK9bPYe85mfVLBd36GlwOdFqAo+IqvtdtskSeDxxRg1vb26ukKpPlWCqIHScbEEvgLw8LPZ7Hw+j567AvD0fr9HS4Rfig9ojSWUwr60aWqMBGGxV1nkgMjF6HX/4YcffvXVV5hvwBcB+dPEAKP8l1LCgGHSU6BRw+VarVYPDw9NUzMmoGR9eNjMZrPVamHsQAkHbrzf79++ffv8+XNKqRCc0ei9f/bs2ddffw25FBIqZvgRLqfQI6VEVsM39fnnnzPG9vv96XwczHCuz4zRm5vrX//637Fg0Qcym1VSSusG0CrH4+l0OkOSx8YZndlshteb5zmJ0RgLyoTSi5uetX3bNqfzsa7PjBPOmdYSDDa6bbAFI9gbYoQsUzjnTqeLvaZzdj6fn04nTAzg0E4tOAazUOo1TTOfzwYzPD4+KqWePHkqhES358c1VnJ0gtBaL5dLQqkP/nQ8YWboeDw6Y51zgUTOOWQIAE6QimKMp9MJINJsNsNhQEqYymvg3ihDCYl459i8ggp4Qm611vP5HHN23vu//euf/dj08Iv/+RkwAeyU8N5jqJVj/7D3i8UC1mboG8RlzZnlAsPxhlJKye8n2iil6Jvw7uQPNpKimAUpRMcNMxcVR4zL5RKlijEGCBoOJQBfVM0XsY1zxpiqmsEcuCiKxWIxsr4Fsiu8rJNx850c97WFENbrNb4YpKv1eg17HxS50B3hu+/7frVaoYpHWzDFLGRNCBjiuIiCMWZHXwdCyOFwgPek9z5GYq3lTGCeLk0TIQWqmDCuMwRcEEbzPuROUCne+91uh6e6TKh6r7S6ul477+qmiSQyzqRS1tmh7xFSxehUCuqsbVsUej9U0PvghVRQPiDQLBYLtDXAKxCO0zQVnNfnGghVkqYhxgD184XUTdDWTK8CRVnXdRjlRYHGRkd0ZF8Ou+bRIA/UOtSu6NXQ/pPLrirYWlDGWJokxpi2azEWEGN8/vw5Kuurq6v/j7U367osu6rEdn+629+vj8iI7IQkwECVChtDUWiAGWWXH6jfwK/wo/2fagwPDw9XYbtsQQFCSCUJAalUZGbE19/m9Lv3w7z7EPajU/GgkSRffnHvOXuvNddcc801aj3qESEVSg8cDykzIbiU8vn5iTFmHUJ5NKMlhCyXS+ybw2i0EAIrYnCG8fBpcurGIjB8yJ///Ofb7fbu7g7UBCIO4ru1llJW5IWz1miNDaCYEDk/P8OBR3k0m82maSwAT5QOU/OfJEnV2dlZ3/eYPIfvISiL5+fnly9f3tzc/OQnP4mRgELBXPSPf/Kf/+Xv/37XtZydMMSnn376ve99bzabrVYrxmgMp52MWZbtdju0+rEhChENSBx2L2iE7nY7dJWccz/96U9DCM+75/1+//DwACWulPLNmzfWWi4E4xSfpOu6qpwXRXl7e7dYzrGZADcxyzJcvTzPrbPBx+SjCM42DkO/2z+N49A0NSERIBX/XyR1sMrYQfDu3VfqtIxyJqX0PoDxjzGAXkaVjB7PBCUxGg0WwRhzf393fnY2DON+v4d3yESrsPQHSjY4uzDGhJTeOWy78t6TEAkh1jt4IsDhQwhxfn4uhLi4uOj7fhoEnrS/6Gcg7iGoCiFi9ACXsOqB5vvm5gbVM2705eWllHK32zHGfv93v/N108MPf/yPAJ7QTVtrEYaytOGrqiq0zhFP42kVouecWWe999Zih1dM4tyTxAhcEOIyggL2JeGBgjbBxBZyY1VV04Di5eXlSYIyDCgjkKIQShC8hBAYoFdKwWIFdSX+likhIaafePYQcHQQdBCCUT1kWTabzUDtIaxMDHhSr2fT75lOkjptNTk5NKiTjdocf+nhcJBSwOOFEAr0mibAvfOn0U3E1ilfAjlOQyQubURApEMxEU7escFY496bCgR1IxIXN+ltIBpBlJmwkkj+xjBMjsk/EYOXMcaub0L0cFijJGJ1MOf8NCuXZ4Lz+XxeVeXT8/OrV69csr0CHqzrGipbxlhRFLAMy5KjKo4HPieErfjidV1vNhuI5fCp0ICx1sIMkVHKOZ9VVZZl/dCzZNwEKIofHrX24WR8QggBRGWM5XkJjTVjtG3b2azy3jtng4sYu/FJPg/mAUp8/LfT8QDFmqeFhhiSatsWDV78ksR/yqIouqEL0Rdl/vj0QEgsq8I6Y6zerNcXFxd3d3dwE8H8OZwYlsslpsdBySJAi+QThdMyDAOodkzFg4Zer9dN01xdXY1pQ0bTNHmePT8/URo/+vAja07wM8a43W7/7M/+7PXr18vlQoqTAgcybqQWMLRKqefnZ5omN29vb5GKnp+fP//8c8xhNE3z+PjYD/1u9/T09Ljf79q2GcdhGPrHx8embWMMFxeXSinvYMa1UEp99fbL8/NzHAOSRM/gJJVUhNAYA2OnroPW/f7w9PT0qPVgzDiMnVKCccqYkFLCWZkQ0vf9er2GHScWhZ6dncNNBLceFvoTKAGfieuAy0UpBTUkpdR6lFIQQrW2u91uuz1D7YWK6sMPP/ybv/mb7XYbY0S/bb3ZID2M46ik5JxLLrIs08ZYayErwIrc2WyGflhVVZDFI+XjQ6IMQkC7vr4mhNze3iql1uv12dkZFNUkmcog7CDBVFUFDF3X9R9997/+uunhr3/wU4QJmOZjWB/pEZEdzv6Ye8KoEcwYpBTDOOC9xkg4F8goU6FXFAVofVC6CKCAxkXaLQo7BABJlJaIy6vVCrpgyL9wXsEj4RAzxoqiZIzjl2OyH3pk/DzSA0u+T4iVaIqCK5uICFxCoOamaTAfQBLNCj4Ht1Qlg1gIZIGp8WMoqqaQgS+rtR7HAWtPYAxgjeNpONMHj+g/NdIBZ8Cl4hWAcyeE4Eni6EzjbCEGF/xyteqHYRgHylhRFs57Thm62eCs8bGBTPEPCHl4UM77k4+HEMiF+FLjOHrvjDGZyqyzjLKL83NGaYyEEGqsBbkfQpBK9X0PRyAkA0gy8LSR3q6vr8HGxDQPj5xB0sgk9JEhyZzAE+52u8Vi8fDwEJNZrx7HEGOeZZTQ2WzmvO/6Hr8TTS/8hkiIsQYQHncMX5ZzqTJ1PB6rqnx8fBSShxBC8IJJoHX8PDLEkNb1iDR5i4P67t07oHgks9lsBh9yMGbDMEDtjsY15zxTSo/aO0ditMYu5ovggzUWw/9gqK6uroZhgKvo4XBA9YyZgN1ul2UZGEKXTDUopU3TQO4CjCylRKQoiuL8/BwEOlZuhBBWq8Xf/u0PfvXXfo2zjNLTtiV0+H7yk58wRi/Ot3jCYMayLIP+YrPZhBBAy2AeEDhyPp//+Mc/RpWDsbLdftd17ePjA8yld7sdKJq2beumGcdhuVxst2eUMEphvrKhLL558wbhFbkc6byqqqZtlVQhYG6fc873++e37754fHwchsFY7b3lgiqlrHE47SStSUfwef361X5/aNtWSrVarcqywku5uDjHD4DJRDmIIAAeG9nCOdf3/fX11d39PZqyx2Ottd5uz/I8B8UCEDPhs67vjTGXV1dlUXDOj4eDlLIsijzPVZ5Ns5NKqdVqBRCAKARrgCmw4H/ZaZ88pZTiLTBGUWVeXFzgVDjnUOwCkWut0cy4ubm5v7///zc1/f9ND4QQxth6vX58fEQaAJCRSoYQNttNVc2OxyNlFMwjqt3Fcg4Ay9iJP5lYI0xU4tEjW+BLInlgUAXhCacQ2Q9cMIrEyWEGOTzGiEfJk/MSRiIpZYjFKOsQCCbSFqUJ0PrUAQ4hXF9fw7pkUvWgt4OZI0BsZC/23uB0SLtrEPUA56cZH0II5kIhYV4sFrvdDtGQEHI8HiGikOLUZiD05KgzfV8UxWg6YUgNJQIeIG4+2qHIJSEEQokx5vz8bBw1XhyHUjV115H40ZXBW358fMSYPkmbYmH/oLIMPbFJy0EpidEF70kMSslx6DebNSXRWKeU8sFLpcqqAkEvpXx8fERYnKbeZNqggJQGh1cMeSKzIhyD7MKDAouKNImEEdM2oRCClIJSaiFPYjRTGZajIfMBJaDQFlIcj8c8y/K8GOF/RWme51mWwzUkBF/XNaGnJc8kEJKcaqB9IIQMwyikkFIYbaRS2pgYgrXWJ3tgfEh0lZ1z46iF4G3bERLRhEBVFKN3zqLg1nosijzLFKWkqmb5yenSt217fX399PQ0m80IIUDoOCoIeXhWGG8ESEJ9k5C1BpgAp5dlGWPCewcN1eGwQwD6xS9+8Svf+DXcFMaoMebFixdffPGF997oEWE6JF0DBsEAswBowB4vFovb29tXr169e/duv9+jKHx8fGzbtmnq3e4ZWSGm5bXOOetOA6er1Xq1XIcQCWFSihcvrtu23e/3oN3Dyf6AEkKkkFoblI+o3na7x/v72+PxOOp+HPv5vLLOliWe82m1BsiA3W5XlhUh5OXLl7e3t13XE0LyvABOatsGzTDMVE9RAscSZAaUDmhBXVzAjCRzzjdNg+G1169f61F/9fbth68/1FrneaFHjd5M3dR5liNq1XWNYpdLkaqT4vHxEQsbmqbZbDYmbUxBY/99VIqolWXZ8XjEiPsw9HgpZVne39/zk7/W6XYzxp6enrChcrPZ/LPf+ObXTQ8/+NHf40Ognuq6DimUSl6UJWGUMHr94ubu4T6Q6JzruwYjlLNq1rWdd0FJpbXBUBIuDDgiDIKWZQlpIOIR3H4mYpokNSeaulKxEEMkUUqxXq0ZE/v9AfNZ8/mCcyGl4lzc3t4RQouiREBBfAdLi7COugwmCii18ILBYsMcFBQB7CtOuvU8R/U3KWWRISZ5g9aac4a1Ks5ZTBKg2eicxyUBhbrdbm9vb733UmaE0HHUEK4JwZ23Ugo0RUFSg5dHOY/Ain0yk2bXO0d8EJyjIYxgKqQIIQYXFrPFw/2Dd15ySQKhhMLxE0WPT5uL0GZ/enpClwVfilJKIvHWKyEJIYzSvuussSGGSIKLjgnug4/B121dVWWW503Xaued85EQLoVUCl0KYMbJ1AzpTSRDJHNa5xswzOXSxhU0WjnnExM1n8/B0p6fn6NORU3mvZdCFkXBEsA8Hg6UM0IIXAfce3Z1MZJxGKQQi/k8hGC1UVI5a2fzyhgtpcCwC4Y5QoxWa0yKYEDEe9f3nXU2Lyvn/GistU6pzIdorNNaD6MJMTofVFZIqfKieN7tzeiMdllWNG1TFCXMnIsiY0wMw6i1YYxrbXB0YyQheb6iFYQC6OH+oVBF13TW2FzlQzdUZVUf6vl8oY0ty6rvh2EYF4vlMIzOnZy+lsslOHFkuBgjYyeZ+Ha7LYpyHPVyuXp4eKA0lrNcCILhGGftBx+8+vzzX3R9b12o5jOpMh9CICEvimo+e9o9Z3kOf3suBBciL4oXL27u7+8JIe/evfPeo2723kcaCSN5WVSz2Wa7nS8WZVUxzterDWToMZLNeluWxYn3i+T6+sXt7T2lXAglZRYjpUR4R2JkkTDYxiBHDkN9//C27erjca+1FkJu1tv62CiVUUoRJS8uLgDP9/v9bDafzxdKZTg8hMTFYs45A/ECJIHUjgQcAtluz7qut9aVZYW2x273fHF+CWxXlmUI/v7+rutaa0JVzWMkx2PzjW98MwZCCDse69EMxpiu7QghwHx92xVF4QMRXGGIqsgrzvirD163TXd3//b6+urLL7/I8yzPs/1+RykZx0FKcTjs8c9Fke92z8PQV0W5Xi7rw1FycX15pYexbRpGadt0mczarl8uF6PRXd8cm+N6s/rOb/7q100Pf/U3P0FMx4z48XjUWnPBl6vVzc0Ncun5+fn9/T1jzFnrrNVagwBJNtpMaw1OZgIdaN8BIaK1OyVMYB+YxGFhGXB3iH65XDpnMTw5n80JoaCbtNZnZ2dKqfl8DokIqgegYzQzkW+mUBhCABAjhEDuBooAeA3VA5zAgby89xhrJIRMhUhIDtJAT0C4dd3gloYQYSCM7daoCpGcIFHQWsd4MqfCL5TJsw/ABKAYXxAtEPzVeCz7/R7nngthjbHOwQGprCosd4wxWmObptlut+hToZbi/DTtQZOqCkAelBooHZVW2RRFwQkjMUYSlVIGTiSYxogBph2cCz1qo+1ivoiEDqMOMQgpkKjQ1ptIRURqgCAsoIXqASOswCJQEADvoBgCxwL0gIOEtSdY0Q7KEUAb/+d8Nm+aJlKCB4u/QmsNxmYchqosn56eoBcErKaM5XkRYpgq8VPLXUhOqXNuTGs/wK2NxnAhSKqArbVSSSR1+KntDwepZJnGMkikEPzA9pVz5r0virwoyglhsGSSoZMLN0AVwONyuazr2hoLHEPTyjycz/liPk0C4ZA8Pz8jB0A6IaUEH8U5R48XdNCnn36Kh6C17rqGUppnBVahGGOVyi4vr7/88stx1JRSKRWlBBUhMs3t7W2R9jOiWMnViawDHgKvuFwu19uN83aayYAL0GazWSxWmDE+Ozu3FkvXKfpeIAa0NjY5F3nvki8cKrQ4DH1dH5p2vz88AUKBsuacY2gGso5pSAsPFo3oq6srsHCQkCAmgAUBvwQnfCGktQ50NNrymGngnD88PJyfn0Owi70Ofd/vnvco7zD4td1uQgjwig4xNnWttS6LoqqqPMvatlXZqfkMDhOdhuvrqxh927ZQx4DMB44EkT4tKATH+HB/F0I4Pz/f7Xbe+9evX4N3rarZ09NTlufD0JdVSSlpu3bU4x/9q9/9uunhz//yh8DIxpjb21uUVHi1VVnunndSyFcffPDFmy+C91LI4L33Hs33CWVjyhzlglIKugWcEpo2G+P9YbBzYpnTgfC4dbNZqbW2FoYHG6UyzBlVVXV9fQ34D28TNKB4Wi2Cu4TgO3VZAUsR2V1aiQzqY71eu+TUjzICHwnZBZ/ZpUWPE2GCEIzLiZYAhoPAfkxda3w2c/KMO7XlEQhAjuMPoiH+PaqHmLx80WLFScJno5E47/KigDteWZZ1U8M/T2sNITlKNPAGE+0+9brx2ZAwULGhVBrH0RnHOaecsWQJJ6S01nLJCCHBe2+d4DwEH0Jo2h4VDHghfM2pXY/0g2+Bx44oQ9NgCt4I1KiTGGxSPeGDgaLFy7q7u5tia6YyGAljp+FmvXl8fBRSggVGBJyCqfceQQTsASK4dXa+WECCASkOnoYSJ7eYIv1xzjVtI4QEtQJmAA8HRChC0vF4xDJIxpi13gfvQ1iulpxzysgwDlIJzgRGzPAtoLYA5pBSDsMAN39cByml86Hre+f9oEfKmVAyL4pIIm7KcrnEvCru7zj2wBBIzIwx+EBgNAHUHKX05cuX6NidnZ39/d//rCiKGAjngnPhnM8y9erVq88//9xa89Xbr7JMFkWBc4sa7vn5GQpOQBwmRCCUcn5+efnu7u7lq1fz5XJ7fr5er3xwWZZdXV2hcHn58uX5+cVquQZYXC5XUuDsZLiwCJrAGRO5FELE7tIQnNbDfv+82z0NYzuOPV4xBA6QA0BkjEv99u1bdTI3PTlROueur6/xbOEug/uFRA7zbRxLdAe3222WbJuhU5qgFa4/jAvzvEQbIIQAarooihii9Wa+WMyqmU2GvsvFwjlnrJs6BCQtl+76dr1eNU0DSISpLJZ8wBAzIVyEWudY15Sx1Xp9dn5+OB61MWfn51++/cpoc3l5WTfN8XjA1iBCqTH6v/2jf/VLSA+43sfjEdN6oCMury6FEGDSpZRfffUVHoG1Zopc7zei4WSJ9gN0sev1GqtafDLKx0OfojC4hSztV4kxEgJxFMmybDFfZFn+9PQEMD41kBeLxd3dHSwqyQQw0gYVVDbIECLZE5GTvpCWaX0CwvSko0VAmQ4rUG1M/lG4YDFG7B+nlPX9UBQl1oqFEA+HY993JFHn0zCttXYcNW4DDuikQdJaA+oCxUxZGW1qqGPRplZKeeecdcdjvVwt8WWLsjzRu84jXUEel+RAYWIk0TMHpEIvhFKKMITNJ977GGKM+J8IzCWwZiRGISQh1BrDufA+CC61MXAOn8ZNRFoFiLSKoI+nh5cCwtomm1VCyPvtcbBeEFDGZPpN03TkL37xC/wkKjMkBkJI33VVVe32O568ryFLnSg18PgijaQYYxjnMEWHlhGvFdGE0dNBEmmQVUqJZZbQ4OIWTL0iHJvf+q3fevPmDUq9tm2tce5k/D6WZRmiN8ZoPWKaD1CDJN3E1LpH/sa5JYSsVqsYCcpKgBKEdRwJ/BsQksgWRZEDpQGpSClRUuR5Dp6NMfbw8AApoNY6Rn99ffOjH/1nShkhsSxLEglAzNXV5eef/7xtW4z7bWMAACAASURBVJAbNOnxMBGy2+1ggoLaF9NqIYTVaiWE+Pjjjy8uzpfL5fnZOaVsvdpcXl7FQG5uXtxc31xeXs3n881mM5vNq2q23+/n88UkDMX2Apv2A8YYvAsh+Bh9CK7rm+NxXzd7742xZnJIBZbHc5jP5/f396gMQOlApCvS2igMGRRFgcERRIO+79HkU0otl6svv/wKgBJNCADKSf6HqtellcMqy0L03rth6CMJRZEPY68yKYRknG03Wykl3MC8dUVRtH3LOctytVwtslw1ba2U9N5rPQohHh8feVIbyrThB8QjPi0OQ57n2BYFtAHXshBC13SUUspYCH6xXFprIokh+H/9h187PfzFX/0Ib6Xv+/G9pWkvX7yUQu53u/Vqtd/t26YJPsznMxhhYn4NYkRgoqlPiwuPQbMp8IE8QfQBBYSjDKiIYsp5WxSFUhJrfJx1RVFixCMmeTJu1N3dHfzQsYA+plX1iLw0eaniiSNSj8mpH/8vcLVPT09ISyHNsqFQRSFMUmMD1DmyOnzNAMaR7XG967rm/DRtC5kjXuc4aqjCgHqKtFUUJ0Alb3DUHFM/E3USTa7XRVl65/phuLq+LstSJwtbyQVOIUk2hRiY8N6x91abgeUH1ROTJDemWdzlcjkOYwihHwfnnFTqeDyWkBWQ4L3njOlxnFUztAqtddoYlOfn5+dwoYEFC4I4GtExRvgBQAIIJAt/8vD/ngXF24fYBkU3TS72l5eXfd8/PDycRtz5SbO72WyM1pxzLqTzDmkA42D4grjnADTILpDQeO998KjtIJtmpzFsURTFdrtljCFeGGPyIm/ablr7jnQO8LTb7Silr1+/Bs2NhNT3IxPMWKPNWJSlypUPvu1aGukkx0JwPx6P2OcR3tvHjmS5Xq8vLq+stXVdL5ZLa22I0TpHCZnP57hchBAM7jrnYgz49yiF4fiGwggsn1Lq8fHx7u7uW9/6Vl3Xxuibm+u+73/wtz/IsyLL8sVyTiJhjIcQzs62x+Ph7bu3eX7yPwfGmkrbkxqQ8hCI84EQWlazEMlms91szxaLarPZrFarLMs+/PBDVDOQPoK751xQyoqiqOsGClrU7kiH/GQcAj9zE4Jt2uNu99h1TT80h+OeEOLSeq4sy169etV1HVAO4gN+J/rMCC+IHpvN5vRO81ykYRqoq4ui6LquKHIh5N3dHdpXJjlJSymVOuFgwGjc2eVyherZWguvzLIsPRqWjDprJ98Hwbn3/urmGt9UKQXGD3cBTqagwpDYACsxNgHAnef5aQguz+aLuXXOWLNaryljXd8vlkvBRN/32/OzYeiFlKvV8nA8GGP+zR9/9+umh//wf/wFDi7irErG5Zi9woAMiJQQQp4XWo/Q3k2U8RQukWZp0tRvt1shBChChMuYNqaBbkLQByShlHZtq5QghIDswTiMTf7+sOo7MS2UPj4+CiGgQIAIHelBvOcCi6Q1xVm8XfAGgLdwVQO+QFAOadEF/iugm4l6CiF4H5wNSmWciXHUlDJrndHWGD0twwDbdn5+HmMchhGpSCZPZvxO8KS4dQgZQ9prBDSENIOPzRmXQvRD70PIiwLdCDRdYoi4HgAduJBYeQ38iIQ6VXUIx9OUMsptwQXnXEgJUWbTNPPFwlorlMqynEQSfBRMMMYpZVmWj/rE0a9WK5Zc85Aw8PwRLvHA8YKgjMRfmp2MdU97GlCkYuwRvXqkT1zys7Ozzz//XEpZVZUxugC9lueCixgC43zUI4oAvFakYbRV8MuxUQdTAoAyOOEA7KemlBSILJvN5mc/+9m0cquazSH/BwtH0pAKYnrXdd/+9rfv7+9xsAlhQvzTrHWWKQjnY4iI17hf+D0oSiannbIssQ85z/PZfLFcLqHZ18nNkBIyHR6kQMQ4xijaaXmeQ2WLuTaaWiY4SHd3d2jgvX33ZfDh4uL89vb+H//xs3EcsywXQlAC9BOFkEWZv3nzC9Q3aL957yEyNMZst1vrAg4tcMlisXh8fFyvV5DMzWazPC+EEC9evHTOAzMhwuZ5ESOmlyiIHaRbVDksSTmd886Ztq13++fn3YMxfSS+61rvTyAaPdH1eg0/u6enp7OzMzA/+AEcOSQSPGdcDUgkcKNRaEJz33XddnvW9z0iWyr9R+9dlp0WpQDWgMV99+7tYjGHAo1S8vj4UBR5jAErAkkkuEpKqeVi8fz8TFi8uro0Rltr1uvV+flZnmdd10IHBJYFQQaEirUWXTe8a1TblJ4WaKK/iy+otSaRSiEGPc7ns+fds5RCKklI/OPv/t7XTQ//+3/8y4kc92k+IIRwfXFJYnx+fPLWRR+sMcF5paSxhqX9J7gVyLG49ohEMe1Jxu+klKJ/gGCBB4G/9OzsDDO6wAsqw5K1EELIspxzgWQzjiO2IiPin52dQf4Mfsknnwxg/LZtQXFMxw5Es5Ty1AMPASpjqGWAZN8vOMhpJ+0peMUYsfIbtyiEU/4Yk6+6tbbvuxgDohiO483NDWMMJktINsgHuLdABMglMS3VcMkFfSrhT4S+c866YRh8CFVVEUoxqQCjwMPhEJI5FcgrSk8rnsDA4iswxiDbfb82wieRXDrnIiU+Wf2cnZ8LISIhRVF454L3mVKQnHnvm7aZzWagRK+urjA9Ow0bIjoDuU9TtYAXNPkIQdmMTIA6BnO/u90OqRpJBd7UGBFaLpfjMGAyjnPunddaG2uMtbjkWuuLiwuMrSEBAE+ACyqKghLi/IkcAGRDhV4UBWcUle5isbi/v0cZYb0rywqgAawOjgQ6sdDpfvTRR8iOwzAEylSeFVUZaRzGsSgLxnlR5iTScRhCMhlkJ2s5jxYL1H3TjM6729vzy0tCicqyUWvrrHWOC0EJ2T3vLi8vQWqDSvLeO2dpcr+PMWLPLqAS/iLMjtR1/fDw8ObNLwiNePZgKvp++Lf/9k9++tOfUgJXmDHEUJYFY3RiBXhajwHMG0LI8tL5SBmn7OQxvFqt67pdzE874QkhOOrw5sEcDNCVEFJrLaUCdkQEB2tHkteLtbZtm/3hue+bYWh9sCGYEKLg8vLyEgcYrPUnn3zy8PAAdWJVVVMmAx8ITIlYD3oDfBF9b738lL2sdbPZ7Pn5GZ8KbSFK6fF4gNkiCl90evI8f3h4mK4tBlCWq5V1vuu6PMuUUlhxEbx//fr1Zz//DL1PXEMs+Zl4RXidOecA4F6/ft22LYRCNzc3UyDKivzh6XG5WnIh+mEYxnG1XkcSHx8eJRcqy7QeZ/NZ33dCCiHEd3/va+97wNwDXj8KLjzlzXqNszUVFowx5z0XHHhnoiYRCJDTaLJaQgggabYOydwYY9IAPWLN1K2llA7DCTl6HzAor/WJqoY6AkAb/Sh0vzFei8eHeA0SEycDMmSkByRzlORZMoaq6xotUBTsuLTgvhC8kG9CCFiL5pwjlFHCGGOcC2tNjOCsibGaUoJxDXzO9Xotpey605wkPgnGMlAwQuuFQCmEmJhigA48JUIISgSjtTaGi1NfjzMWQpRpZTTOGR54VVVajwBQ0JjiqaL6+eqrrxBKcElOUiKhlFKBRLBzRVFY566ursZh9M7FEEiMzrrNep1lGeNcG42DLoS4ubmBBQ0wLGKuEALW+SjSwS2otAfUJm+rsiy98yH4SAgyOj4kCk3UUnVdO+fGYZBS5VkOLiuGwDlz3sfEZGJttU1+kTALYskjchK2d8OAbAFAg4MB75NAImUsz/MPXr/66quvyrIijBJCkXJc2k7snDs/P0cwQhy8uroCj8fFifSAW2oIsEDgLFKdghHy5alKmM2m60Mp3W633vu2a+umhSAH6eokZQkR1d7FxQUSUl4UXddlmWKMzWYzXDQuhLN2GIZ+GBilaNsCkyX8K8ZBa22891yIw+H4B3/w3R/+8IeRRHhlC8Fj9JvNBs7YwzD2/UCSu/7JlcAFlPg8GecAZnVNy4UkhJLISKSUchLJcrkCL+S9R+suhGCtizFSRketCaVYSB5iZIy1Xde0zeGwG8c+ROecjsQT4jnnUmRN0/z6r//609MTxoOHcfj000+/+OILNCYhQ8DrQKHGk43xkJaGoV9ikjUWKoOiKGI8CSswiksIQY0ipcAkAXAYY6yua8Yo5xRTLCF4pWRdH7MsW65WRttxGEBnHY/HMi8IIVfXF2/e/KIsi6oqj8fD8/PTYjEnJDLGQRLgQmVZ9vz8rLW+urqCwc9+v7+6ukIbMsRQVTPBOSN0s948PT0FHzBl2XfdbDFfrZbDOIbgtR7Pzs5+5ztf29D7L//mRzEEY3RyixOr1brrOi6ZD36335VV+bR7yvIs0siVUHkGuwLOefA+eO+sZZSuV6sISWjwlLH5apEVOY2nvTSg6VHEAd0gw6MWRoe27wfvotbWWY9V3SQJfuDbgy2JMOVG6QdMh5oxJg8lzPSBoPBpO6BMEwAoHYCsIXiAmIGkJjYh5DT5QSFm4DGSw+GoVLZYLJ1zlBEfXFkV2oyjHooy98E7Z6VUELmGEKVUdd0wxnFGUTzidsnkHoxvga+PhKHS5jU8XpIQt3V21JqL03I6wTmj1BoDmEPShlRUCVVVUcqMsUJIcF9SqrKs9vuDEHIYxizLy7Ky1mVZTilzzlezGXxeSWr1t01zdnYmOCEheGv7rmvq4/X1FefMeldWFfAgqDnv/TAMKPORYIDsEFBQHgG+4feDNG+Phxh8ninvHKMkOEdjLIt8HHtOqVKSxsAIoRAsWMsimc8XIXiU1XlZhEgiJf0wjFovlstqNouUSCUDIVwIH4MxxsfggrfeGWvzssATRhOIpK1K1lrnA+NitVofjsePP/7k7u4hUoo1y5988gmIqZgGaxhjZ2dn6Igge6HvEmmUUlB6WtN92j4yGsUlnvxprM9q5+xsXiqVA3ji1F1fXw/DIIWKIZZFMQ5jCD7EYL3vx6Gaz8A1Z3lWlCWmcBnnZTVr2rYfx+3ZeTWfd/0QCSWMHY81lzL4qLVpm1ZK5Y1zxuWqtNpGT4ILmVSXF+ef//zn3/j09U9/9iPOiTZd09ZKZkplFxdXRV5mqrDWG22984RQax1j1Fg3amPTVi4wzFmWGW3HbvTWW21ppDSSYdA00mKW9V2LNgZjLAZCCLXehXQYyqrSxjDO9/v9vn6um93u8NT2xxBsJJ6zGLydV3MSYcOXcSmqaq6yXGbZLDkpTLU7ukcYJyLJ8zzGOG0tROPaew+dfer9UCE4xl+8dxhpQugIaat5XdeIErvdDhecc4EhCe9DjCRTajGfDX0nOCMkZpnq+i6SWJaVMbZtO8b4q1evf/7zz7uun83mPC1FNsk+K8aIFbBwKkNiQN0TvJdcnG3P1qu1FEJwcdgf9Ki324319nDYF0Uxn82d9cv5cuiGX4Ln0l/89d/CVD2ECD1ynud934VwmgNYLpeI5koplSkUNZnKBBeUkHEYKKWz2Ww+n4cYj8djiIExtlgulFJWG8QL9Bvoe/sRYYMDTQvMMDjnIZzcW4uiAGcK4giosG3buq4xNIi2fowRDmgx7XR7eHgA2QVf4inegd6BAnri3wEH0ESB8gffpa7rYRiqaga2d0KOUO6ixoS+jST7GpCeLP3Bt8DLNmloPtGvOahPiNsgtxBCOOdgIWmtRWMffxE+MCae8AvzPEdoRoJETVakFYMT4gajXaTtmIhrsEzHhqyJ1578U/HJgacIIdfXV1iE2zQNLslyuTTWCikxs4LBEUBUfDUcmDzP3717h5Yp2AOtNZ4hHrjWeux78CHIK/iE08Mp8jxL+/s83HaNzfJMKjX0/aj1bD6vZlXf9ypTxhrn3Xwxz/NcZRkqmPV6bZ2DJgT9M3DTSKUTLkF+xSkFyQMToclD0BhzdnaGI4EMB9uu29tbkODQgJVlSfmpzY7aGqqHYRjDP51hX5blYjFHyqeE4dDyZMH70UcfVVX59PQ8am2MaerGes8411p75zjjjDGSzIMjiYQSHzxaWT5NtIA9h8Inz3JCSFPXUsrgPGRa46gnwfd8Pv/ss8/+9E//9H/99/+L1mMIfhxH5zFsEYuyyrISi5cpZcFHo20M1IdI6D/12xE9Qwib9dqZkwBpaviFEHzwnAsM6QdPKeXWekqYthr14lQsjuNYN3XTHprm2HWN91ZKUZVlphQlbD5f7Hb7fhi//avfbpu2KMsYY9+1pwFPQtbrNVgsjOLigoC4h+4Zql+AG8B29NuxiG2itnBWcXHwn0M5hruJsAADLhDCCGhoIKPJAZp0Ej4JIVCyYE/chx9++PT0tNvtUHu5tA0FbVTO+d3dXZ7nNzc3x+MRYz0Q/sBp9JNPPvnBD36AEOG9R/wEYzbF27IsfwmOrf/xz/8a6WG5XM1mM9ibNE1j7WnOCw6x4CIiiZSx4IOE36y1Q98LIU6ezFhGIWUIYb6YSym9dZCLsLSWZ5IGIgHQZME90Uo4cAi+EJZsNptxHO/v7ycyVAgBqYz3HqJj/MsQAgZGsM4+JtEOmqVTszd1z3J0JsE1YVaepHWnfd+XZYVEQpKv+nK5RDjGBD+aHFP7F6EWXBZjbLFYgFIE+4lPDj1JCAHpYeqYQeAsklNT13UQWQkh0AqD6vH9AQIc5Qm846kiyBJCEPiyLFssFmAeklJC46G9fPny/v4ejxrXGxl3olCcc6vVEshFJUf01WplnYNkAF9kGIYPPvjApjERkHWYi4ZMHskGLwj+P4ih9eGA+wM1BHvPYYYxhn4+7ttJgUZZXuR5nlPG2rZlgmPkFeQniI5JDgc2IMuyu7s7NO1xHuB/CRYC/WpCSFEU8BjGyXl4eOj7HlbbMcb9fv/JJ5/c398DnYB9rarq9evX03Jg+Aov5nMlpORCCkFiXMznwE96GJM2ocuyDK3OsirH0YBGQ2Da7Xbn5+chRO99fTxaa513mH8ETRG8RyabgvsklqfJSxi+logXXdct5gvMmVtr4Z/4fNhvttuma7M8t85Rxn7lm9/89//hf/sf/6f/4d/9u/9Z6zHG6Jxx3vR9p2TmPYFgDw0GKVWMJ+dji70jjIGBhHIjL4p+GFC0ueAhuwo+UsqN9iRSQhiljEQaYxjG1ns39F3T1MvF/O723XIxf3q8b7taj4Mxo3OuKouz7RnnLMtKxng/DFKqr95+9emnv3I4Hp1zuVLglBAcEXDruvZpfQBuJe4sLBLytBcLsAwztoBrk9wRuWHyOny/aQFBB2AlEMOLFy/QRYAGAXUz/nnaz4Fbib+FEIKxu3/4h38A/vNpLTGgYZZlmCX46KOPkO8RTHDYYozf/e53/+Ef/gEfdbVa7XY7xGEMZ+x2u/l8/rv/1W997fTwvb/Kc6zHqcqypJQRQur6wDmn9OTOfzweQR+HGCijnDFGGaMUTH2mFAJBCOHp6YkxCsamKMuuaRH1SPIDn3qVeENoaaIOsNYaY/HDaNCD+8MoKU8Oycg08IdBKJwEVCDiCSFoyU5vGhAeUBehEHgZwRQfHhQtORlH87Zty7KaPi24kUnjhIYn2AZkewjmVLJ7yrIMNpzA+wBHNE0ITusu0DNHM8CefF0c/jqUCFBkKaUABkUaEJu8enAr8clR4aHXisuA6Nx1HbbWoKBGgfzy5UvYUIP0RCAGEkSffxiGxWK+2Wzw8/Ayms/nwziChMO0Ng63MQZdtTQ0lHvvUV6gtYMW0alzEGOM8bjfH49HlBr4FpNUzBiDOA7FMCFEj1pwked5WZVSiK7rQog+nHanIBpOauMsy7QZtB6rqhzHwVidZcp5W1aFEHKz2QBbQS8X3lN1o07abrdIOaDyCCHgfzC/5r0HiizLEvfCOXdxcfH09KSUQncd7pZFUZyfnzvrGKUIE8fjQWt9fn4mpRz1yJkYxxF2QzhXX3zxRVHkTdMWRQHiRRut4EbOOBTMJO3HhV4LEAogHdUnGtEn3syY2WwWvO/7fjGfSynLWQWrH7QWEFPq+vh3P/u72ayy1oXgD4e9MSYGV9dH70nXt8MwppqA5XmR5TmMb6crA/zrLM6ve//AhxAIocET5zwhNPgYQ0Tgbvvj8/PT8XgYht45+/j4MAz91fXFw+N929WcMUopiXGz2VRlpbXp+6GazSB5+vLLL/OiIISQGJDvcYB1cl+e5j/wSaCZRB2PWDGZSmF2miaHZq21TH/ww7hZwGSTzSh2XEI48+LFC2Qal/xjbBL07/d71Kn4GCjysEro5uYGMrOpBQVQhV4IYwwuR69evcIvgboShcK7d+++8Y1vHI/Hruuur6/x100RIM/z4/H4x3/4tZVLf/WDH2cqA5cihBiG0Vo7DD2lJ788XCd8LOssds8WeY4eNKWUUbpYLBhjoKQjJYyxaj4TQuhhDGmKdeIWgGrxlNEmenx8hHu+SXYCyNiT0AgzU+Cj8OyA9G2atIJECpF9ErMic0xvHeQjyj0Q/cBuMflbAOGmRGXyvECVg2cthEB9hx4GSXsxJ6kZ/gHpAd17hGN8MNALYEhROU2XCtjhNI/KGLbKxCSgmjKWSeMOWmuIhqeOOgIEmtuwmcLvh6YFRc/xeIRIRkyi2DQiAOAD3ME5x1d2zjVN/eLFC/jo4Rnmec44r5vGJleDqe+NWAAvT55W3Mg0QenT8AFyDyGkPh4QcSB4m+o/fDBUeDFNyTHKjDZCSiGlVMoYM4wDYwwu5Uj/k/Og1rpujnj+Uko4Y+PhoDzFzwMAZmk7LHg5IIOp6AwhALUhCU3+80opMBhIbOfn55TSp4fH5WLRd73R2mjddx2jFOCan4ae+OFwyDJ1fn6eZcpZj79ru91OISyEAJt6DD0UVTWMY4xxvV5ZY1GJApbh4UDWjK+/2WwAYoAzOOfWGEIIidE5F7yv5jMhZdt1eZ5D5uC9N9ZeXV/2bc+F2O/36PbVde28abveGktI2B+eRz0AhHl/UrLNqqoqKyUlo5RRZo31MQgpQoyRxFGPPnjGOWUsehpCJJForb13ox60Ho7H/X7/2LV129bj0Nf1MUb/xZtfXFyeD7obxyFTEn523joYrBFC2rbNsgKtwbZri6IwepwA+3QAOOfb7RbmXWAOwA7xtJQUVxICIfSEp4NHklszTgjeNeI+7ixyjEx2Pl988cV0y2A/hQeLwYW3b9+C3kCJL9KWBGx+nBRxKGKAWQEikTAwpgrLCbx6xBnQzovFAptogbZxnECodF33r/+b3/+66eEvv/8jRrGmjW02G2OwCNozdqq8ttstZjRCCNa5LM8IIZnKCCHWWKWUFGLq/rdt23UtY2y5XHAh0HvAhcTNB9BG+EAxhXYCGA9r3cTDAgLjP0FWmNq5E1REMgDGRIjE4QC7jUeJOzPlGEQEfAZ0CMAeAiECyeLjYSoOp+RwOKB+mvpIIc3QseTch98JIhjzLPjlIvnI67RwHF98TObD+IMDilTEGGvQHBYC2FYld6OiODk+4jsiuyDAIbxOBn84djguUCHjPU698U8//XSC88glKJ8h8jkej+DKkRHxEjnnKsuMOS27BkSFjHUCATc3N0mKNqzXa9guIbLAhwCzgYySSVwwaXhwaXHWUSCiMhBcCC58CFyITCnGOeg+PG1YXZLU9x7HsW6O3lvGaFkW2B4jBDdGG+1wMFCrIV6gxsIFDsllC4Fgs9ngDKBTNdWj+ISwtuZp9SyWwCyXSzjH7Pd7VIE2bQ3K8yzG2HVtURSb7UYIBeNrMMsQOrdthw3MRVEQSlSmqtms6zrBhRQCPR6kB5RKqInBeQohLi4upjo4xqgSq0kIsdpYZ0lafLtYLNAc2u/3u93+7OxitdrqUWutOePzxbxpWiH44fjsvdVm9M6Peuj77ng4DsNQ182JYrJ2uol5kYe0VxzkLa6nHo1ztmnrYeh3u6dj/dx1rdadNkPT1vvDzhgdo6eMxBgOh+erm0tG6XKxKPIipOn6LMtCiD4EzjmhNMtzCN44O02P4k6hsG7bFqY+k2s3Ej9CB/iG3W4HUheNPby4SQUO/3lo/PDecXNBnwJ64mzf3Nx8//vfV0pdXFxM/UIgjMVisV6v8ffixE7nFpcOE+kxjQcCWcLTGnvroPpdLpe4tuBv8zx/enpCDgO4hAGUSavaAa3+4F/+9tdND//pr3/IOUaZFLSk4TRVRBBPWdq+LaWE+bo1lhDinQ+pqYBLnmEtBAx58sw55+1J1f6+T9HEL4Now79MfQIPQAq1D2p5kqxegUaRWsGGT2U12om45wiObdtmWXZ1dYXgizBNCAERj99wujbW4rX5tJMZB0hrg1EpaJzLssR/O7E6KBhR0E2UekxTwSQ5yeDnccRhizbxmAiyaPxilBcxF1j1+voaTwOgAAF3Kqo++eQTm6w4CCHIChODRwhBvQKAQ8hpdxgyHz7k2dkZAiJyg0lOL1NHRAj+5s0bqCbwAwAQXd+/z4DHNAyIxwI7NpCNoGuQdH1agyqEYIwF71BYoFGM3xaSrhx9Arx01BzOOkqpVJJhj5OzIJRFWlOBRgKukA8GxRCKdJOMoTgXSmXoHMJdB9QBsOFk9AQTJ7zHJNj3IJ3x6JDLYUY7SWXKPH+4v7+8vIwhMMYQmkVCo1mWheAx3zOO42w+c9bP5/N3795xzl+9egUV5sPDA26ByjJG2f3jw8sPPuj73jv/6tWrSeIRTgs5/ETYCiE2m83HH3+MFIJ/H/xpvxghBAPnXMn1Zs0454Kv1uuiLLXRXT+sFmtK+Xq9ffPmCyFEVZaE0NV6KSU9HPZaj1qfVs2PY9/VTXNsu7Yd+n7sh7ZuurZtjvXueGjaFkKvw/EI5Ne0bX041vXxcHg+HJ+7vm7b+njcd32t+37o+uZYU0JQZikhDs0xL7KPP/lY63G92uR5HrzP85xzxjmnjHMusiyPhGQqY5x1bYOoAuSHxpX3/vn5GV2iifoH9OFpCMkYo5S6vLwEcyulxNoYUBGY+ceMy1TUmmQDDh1KUwBV5QAAIABJREFUTC4ss9kMpqWEkAnSoSy+vLwkSQyJP7iA4LQJIdgUiwgAnIqo2HXd5eWlS8vBsEIRgA96kIeHB5QswGeAO3laul5V1X/5nV//uunhL7//I0Jo3/eE0Nms2u/3QsBW89ThweXBJXHWFVUZQxBCEBKxWiDGuNls4LXpvO+6nnGmpGKUktMKcg5JGSEkAXNikhPqpHxYrVYhRPx7mtQ1CH94E4jmLq2UQizABBAuLRAEYwzVOksKUbxyyqh1TnAulUQRg0yOMEoImUZt+cn+OsP7AEMCSTt+zKcNYjQ5CbM0ThFTfxhbO3BR8ahRQIAynhKb1hqcMppj8L0wxuR5fnV1hVOL4glVCCoJ7/3r16+7rqubBqtDD4dDiFGlRjqqBLSdMQM8m82enp5QGYA5AXkaQuCcGqMPh31VlUJw52xR5JxT+EWHENC6kFKCXNrt9zbtwQYuw0tEWkKbwSZTsxhjUze73a5vYelBpRQxePAYiOwiOR3xky9QzLJMShVjIIRgs2Oe5YRRY0yGMs47KKSDD2M/iJNLaxW879p21D0El4j7JETOeaYyYxwKEWP0frfbbrfW2qHvgT9wSjFytV6vtDboxgOyTchg0giR0/rC0xHC7h3Ujvf39whSQAn9afMauMex67osy41z/dATSg7Hw3yxkJmkjCkpm6Yex0EIedI7OX95dn7YH7TW19fX6CQBtOGJTY1WVKuvXr367LPPYoyMUKQNGK9mShlnoXRCUZhlWVkUxlozjlzI1XoVYuCCH+t6uVpV1UzrfrNdFUWpteZCCCGt803TDP1gjGmaQ9vX2gzH477t2rZtuqFv27ptmrat+74Nwe33z01zHIe2aY/j2Nf1oe87Y7U2o3PGe6eU7Lq2bZv5fFZV5XK5jDS2XRtDvDi/4JzP5/O+68u8OrXExjHL8khiUVSURKPNYjFH9AS4QXBHPQp4HtKcJpqFkxYGWR+PAiEC65imu4k/YBTBKam0OA9/FyggxDdKKXYST4tUY5pF/fjjj4FQY9okP6kJoKHCL8SbQj8SeQhsR9/38/lcJyMQnKVJMejSIA5AKrhWhMTf+e2vPffw/b/9O+SfEECAjJSSrmuXq4XKlDEa+l9jNIR0jBDBuZSSczaMQyTRB88Y9yEIECZaz6oZZzyGIBiDisNZyxkr8jx4zxkTaQ8M8gSeNefcGG2MrqoStIBzlnPGGK3ro1JS61EIHmPA/3LOGTstkkTYnUTNjLFJc4aXRBmTSvoYEMSddfP53GnDGVN5MY46RjKtpnLOcy7eJ8HR8QY3Xdc1uiDI5JN0xKcZCxCIk/QIKQSpC0ETRwGwfeKvJ+wckigFUAgfGOgA7x4JQAhRzWb9OAQSh3GkjEUS86LwIXDGoOzmaSb8xCanragAvIfDYbFYeO+MGUL0bVuH4Jy33ruqKvIsH4YRkHxCqUqpLMuHccRBRB6aqq4+LW4D7QtyXLGsPtZ921MaKaHeGSm4lNz7gFSH5IpElWXZMOgkJPfz+aLreq1NCFHlmffeeZ8XuSyyum2scxmX5OQQRwXnRutZVWH/D2VMMOGsY5FSwgThnLBxhEU+GYfeaJ1nSgpunSU0Nm3tg7PWEBpjDIyz5WrRdT3qCeQDfbILjGVVlFURSQjRExKtM9qMnAuVZ03XCikpY03b5mVhrA3eq9NGB8UYJ4QplYdIZ4uZda6sqq7vDKjwrmOcMhbbtnXWlGVxdXlxf3u/XCyNMZQxJCToaFWar0TGyvP8s88+A9/tvbPjwBiVgmNykxJurONceWes0VhlO6+qqiwYIYf60OtOKu6CLcq86ZoQw2K5bNt2uVgVxYxRprU9O7tYLpd61P3QH+rnQMy+frB+6Mdm1EPXt32zG/rj0B/Hvu7bw9jXzg59d9S68240uvNOGz14a8ahJyHaYF1wWZEZZ6p5NV/OXXCzatY1vWDCaDObzXElIyXL1cr50LYdSn9K4nw+c9agVWbSmNukeUHf2CdVMagwiB5BN1FKAeohtEPvCv85T9O1eOPgUZBCyrKEJeikiAtp5TBuN9pIE+2Dcv/8/BzHG2gPDAdL8/PgYENam8iTrwwoI8YYiM1JUANWdpouur+/B9tsjIGJKuqkX0L18L3/9AM0QGDpBVu0EMJ8NnfWDf2QZbnWxllX5AXWmDAKboR1XZ9neVVWUNeg+4cniJgikgqwbhoC02PnaCqyELih95hoegTfidQCHYH+W0jeBon4JkLIKaSiVgA6nszLACUQIhlnPgQKaSBjs9nMGlOWJaEMXR0kAHwFBHqEdbz7KfkDNSBibrdbSI+m0gHnD0PaOHDeg1Kw2D0Q3ttLgQYXehLstMPrpNtBBxuPFEpt/H4I7UFAvXr1an844JTgHJ+CrFJQo/rkcABJX4puFGUv2FUpxTCMMURrbPChKitKKUwXYMePogqH0lrrvB/S9wJxD4w8qcXQrQEbczgcDrtjCJ4QwjjVWsNsHF44SFer1aqua1isSylDOE0/SCmxpx6qKnRTGGMhBi6F834cx/lsNnn2omEAFk4qUVZVWZRSShKJc05yQSk11j4/P6O0Qjskz/NRa7yL6Z5PDFtRlDiEeL+U0uPxqJR0/rSTw6UdRM65rusRdKBphmcOYwzZGnQozth8Pg/eO+9gW/3ixYt3795Bc4JldrCPXswXPgQYUxZl8e72dspS6J1M2NMYs1qt0MMD4dAcj/HkvuwJIRh6LYoymRedCiAcsLpp6qahSeg5seez2ez+7u6TTz49Ho6M8fl8cXV5PV/Mf+93f+83fuO/eHi4Px73wzAEH6z1hNBh6J13ehy1MajGfAigbkKM4zgSSvl08bOsKHNCyNnZGUZfMVcE6hfHFUBq0qdMc4hAISj9b29v0V1jjOGxTEplzGzhsuAioEBfLpcgY30a1QScQtkKqgdXHtATyzshNwohoMt4PB5h/YB2NKUU9c3UmpogFE32wyEZ6yKmee9Ryr/PqeKoIKIiAuCrPTw8QLU40SFnZ2cgzaYZJqQfHLOiKL7zW9/+5aQH/AWEEHi2gHbQWk/e9ITQLFOcC0rJNJ2AXgveCh7QdPpRnTnvszxXWfbw+AhkHkLgiXoDzYLvD1IVIQZc29QpxWQ51r3iTePECCGh8YAYgCR0jwyMx40EDqhezSrGOWaLnLVCCGesc86HmGUZdu8gAvL3lo/i5uOXI9vjb0QWwRVFlJ9Us4ineZ4jgkxxH83A6cSgigQnAByNwIrXjL5Cnud1XeNw07SWA43Qoii2Z2dDPzR1TSLp2o5EUpUlS54NPMm3kCmnQgefZxollVIdj3ulJD62lIqQOI66qkoAT2RZhGxjjMyUSywfTcbpSil4JkOJi0oOmksSSAieUqrN6JyjFC2Zf1oxj3CGKKyUms8XeJ40eeX65NEtpXTehxDyvGDo2PPT0Dh4ZBzCEILKlFSSMxZCKItyGAZGaZZlPgaYMmETCcDjqEdkSnzTqQfovV8uVjhv4zhOjZmu60LwiCYINxA7GG0hjbPW4pLv9/vVapWrE3vg08Ldqqo22+27u7fQX15cXODsEUKapqaUbDbbUetIiLXu6uoaD+c3fvM33759i0SIbuoUE0MIL1++DCE8PT1BCnVIK7Og+RFcSinLslqtFwDCU81hjNHGHOt66pDhF3rvZ9VMSdk27c3Ni9vbO+f8zYvrssxHbf75P/9nf/In//3r168eHh7Kqjo/v/zoo48/eP3B0/MjoTQSwjj3IfgQaAKFeVFUs1mW51KpLM9Vpsry5LwNQh/dfmNsUZQ4qzaZueJyYQU3GJWJ1y3Lcr/fn5+fTyAJpRV80tDzDyEcDgd0kgE0p6kdPAEgS3Q90U0EOQmhGpLuOI4PDw/gbHFJ7+7ueDJXBwKDGGnCsqfUGEKMEdYpCK247IixKpnNIOhjgAOEFSIYAmOMEX5QuIlYSIVrjjw6gUtQu977X8ZY3Pf+OssyfA3EEfRMhFDO+WEYjbGUMiHEZrNljMWI3QnMe3gwkLKsEHRokgxO/mKjHp13Qoiu7QglSipCSaaySeoQ0x5TYDcMuDvnQLohVUIuiZQArIpfTgjx/lQ0TEpTwA1gPTzlE+lEopAyEuKsAxu+XC4zqbqua7se6gJke/w8QhXqQZtmtoFZMCKOhgFeD4TYyK+4XTjuiJhTFRLTplWS5NghNbSnYUucVAB/vAtCCBgeHCDEKeBfzhhgZoCgi1KMC3DGkKHxBObz+URf4iuwZMKIqqVpGiGk1iZGMpvNCKHOufl8Me0vO/11nA/DIITETC9yBh44eFsEx4lmRQuORuq9y7Ks7zvvfXLWshBBAAFg4y5jrCgKrIwHUIjJpwhEXJbnQoj1ZuNjwIeJIZRFgXp3KlsJIYwzQilmtZy1T09PFPPthAgh8DRWqxUq+qIoXPAkiSamOQzGWPCxqqqp2wRN9m73NJ/PkchxjcnJMdciqU+NRNSIgvOpDEW5bK29vLhoutMoyZdffvmtb30LZe44DM7Z1WrNOW+7rsgLTId0XXd5dYVgBLYwhPDRRx9BOlVVVd/33/rWt7D5ebFcWq0xIcy5UEoJLpE+YYoMmNWnwfWmbfuhV2lyHkw9/qHMi64bsBDCWqOUXK9XnPOLi4vlcv7xxy9DIEpmH3/86b/4zm//d//mD7/66t3hcMjyvCyKLM+kkLCfk0oWZSGFnM3nSskszznnWMoNlQ4+FfR42CiHqIeLiYCI/m1MG1yUUlAoAdWhzYADALz/7t07mcYkkeNJUnKDdkadgR9AeAE+NmkxCcIxUhfONuraSZFR13VVVdP8MybmkF0AznzaIoX/RXCYpIw0aeTQ5cWtD2kYAN90GrzF0cIpyrIMEkfcPiHE4XBAvkSg0Fr/q9/7F183Pfyf//dfwUt5ko7gllLKtR6NsYREKZW17ubmenKZV0pha67WIyHUGE2SyymCIOecEBIpAVVyOB4AH5z3lDFGT26deFt4XogpyJZAZDgfgA+4xpOAD/+tUhnGa1maVQbK4EkdC1mYMaZru2pehRixDNUac3FxUeR5jLFuW5pWROz3e4gCMZYx9RVARwJxoCswyZZmsxkmbKfmAUvCJMQIZA70PLO0oBQBF/oENO2RG0IaYMbPuyS6jclkdJIVAZLocfTOxSSYy/O8KssJ4DDG+r4HRZPnOTQPUw00eUbmeYZW/DAM2OEeQlQqM0YDTU9CSa01Y5wlWgzga7PZYOcBgs5E2gL9rZfrtm1AKxljYgyU0uPxAJyB9gzeHS5SjBjTRfr308vlp44XJ1gYSQml9O7tbZHnqA5xwZBa+qGHixbn3GE4Bn3atLMIWsCiKJ6enqSSXPCJ7aRpSUCMRAqJtDpl7uVyWTc1fhJ2pIQQHABGMfnvx9Pu4gCQu5wvJqZiEmIQSosyhyZtHMdvfvOb6ExywZq24UKcnZ8bY4UUznmpZFEUz8/Pv/M7v/PmzRu0viYl6+PjI8xCtNbr9Rp+DLOqxMdGuuVMGGMY44SG+XwGihLUeZ7nh+NRSInWEU4X7s52s3m4e1yvV5MJJud0Nqs4l8aYDz64ooyv15u+H66ubjbr7WK5ePX6g3/87OdlWVJGy6qKhBRluVguhRCL5VJhE4NSQgip5KmIdw4jimjbKpUZY5fLJQgDnbwSgJqRqnFDYTjx9PQECRlgu00jZjiWU12O60DSRhaEWgSr4/EI1dNpEMQYxBmochhj2GSMMznZxyGdT0Iyn+yucTCwjQ5hEAd7UtCgQOFp5zFgFlIXSuSENk4bvWKad6uqCvIH5AwgYETsqqqWy+W7d+9CkqVIKX8J1cP/9effV2kFFW4FPkcItO/hQsxhS7dYLMdRhxCtdTCn8943TQuwKQT/f1h70yfJsuO6865viz0iM6uyuqvRLaAbaHEgDkSQ4kiU2ZjJbP7o0b6Q4nBoGkIUbSgRBLvRteUa+1vvMh9+cR+KsvkEoI1GawBZWRHv3et+/Pjx4zLNmtJ1iTFqdolU1f6w74feGBNitJm12hAU6MWDkYfkckqTk7sEqyAu2yYKasMxYS4WSw4WcZMfjskZmB8b5eo56020ubu700qv12slpXOubtthGGjsUNnxBGTacsrXUWmtEAHIGINKerFYMJ8Z02yHEIKDBdKheuAFg1ngQ00aU/g4FI7/AjXcpmXU40wNBCg/qZVqm6Yoyr7rjDZCxKqsyqKwWQanyfHikca0cocHNaqGUkS+bNNr21ZrI4Qsy8uQnUkrj3hl57qW8pJ1+C9fvnzJE2B7F8oQoJlSKriotToej2VVGHNxrMsyW5YVV265XPL8U0uQj3TpQHDMQvJECyE4N2hr6bicDkejNRiQ0uGSWmRUWgcfYoysbj4dj9ZaeVmhE0lIt7e3/GabZ6N0ikzvnNNKoV7jgoxzJFVVPj090emhRXnZS6ptjHHczsTIXlmWVhv4Q+AhPUzvnMkM5qBQKF999ZUQYjabPj0/N22z3myyvMDecbffa2N22+3Lly+Lothut5zAt2/ffvHFF4fDAaoB1qjve2vNi+srpfRut9PaWGuDv2xAsZkJwfPvQIQsy4qyaNqWeDSdTpEmH4/H6WT29LgtiqIsL1ajWuvpbCKi3G73MYqXt1fLZZllk+PxPJ/Pszx/eTvv+/jm7Zssz1nTqbSazWbGGu8vdIoxRkhZ5DmbiSEw2eecZdlqtQoh0mPnE1KV9mm1IqeL2hpmiRKKP04pH2Ncr9dD+gchE0GZggliAIofAR6RHVWLEIJfFWN8fn6mGzRK4dGzAdRGkpy8Qn3JWTLGjPOVvGXaXeQnsOwkrapV6WTywWCPxwAIfwCSRry0Wq1wKSXxjN1vZoBIUb+F1vQf/+l/QVETkjI9tYPsKPQkmnDtnXNlWbK4kTiulMJkg0hNv4XApI051WfKZCFENZmEEJbLpRIypDW/Sil2483n8xEgw18PaWCE8WNrLfZYPp0qpOvj4GuMEXbIJCUicW3EJk3bTKrqcDg0dTObzSZlFUI4ns5jKY2WnAAh0pwXDBUXm3hK/nh6eiI9IGeUaR7CpJkGc1ntGce9coBTBpj5eY4IxwJag4KDb4r+Cj6k7/v5fE5H64JThKjPtVaKa+adz6zNbAZLMlKZhF1eFu/6Y6RMjcx7n06nH9VwTog4tt3GXO6c74eBp42H3dgZxhqEFMijyLKsrdvVakW7WCnl3BBjXK2W1mYUUiqtbOT+wGNAi40vFKhlL20wkRW5D+Hh4SE675LvW5vMcM7nszZaKokB+Ga9KYridDh67zmBY7cQFvh4Pum0d5a/9FLaSimiqOt6s9nASvMHWQNJeOITghtCiGMy4/djOpLbbDyflJ51XbPzSypJ2Pruu+9ub2+rqiqKvPfDw+OjkGo2mzdNU1aTLM+GoQ/ev3nz5sc//jFdyizLSE6vX7/+7rvveA7wZrvdbjqpiqJs2/Z4PFVV1fcDl8hY3bZN+GhBiDFmMp1iwwfCm8/nQLGm7l5c3+622/liNptN+qGPIUgllLJ5lv/1X/+/Dw+Pk8ni9evrpsGsW9g8++KLV//jb/6Oe2SzDIDIeS6rUhvjuVwh5NZShY9XVV+2AW6IjHBxY5cL6AZ1w2Sfc24UFo7nmQgeQri5uWmTIS7HDKcjXgGJn98D2gCSW2vfvn1LUqQWYQiJzMHdeXh4WK/Xz8/PFKwiGYPjzsIUPbkHyDKipSyZSXNDAQ1kZQogPhjVQNd1UA6kpY91H2/fvv3BD36g054ifoDKb7/fk6j+2R/+5DdND//hT/5MSOG9A/5KJYSIQz9kRdkPg/NDFLGcVIMbprNZUWTiMvZZdF3XNHWLfs4a751PNvqkuGEYbGa3u10MQQhRn+vlbNHUzWwytcbwpol00HBCCBYad13fde3xdJRCZLnJ8wJIjmpoDBbOOUzVXZp3Z4xOpi2nmtFTIeguKqWac02E6rveWpsXedd15/rcNHWeZ8NwGZrXWmeZjUr64Lu+i0IYa0OMIUSjTds2QohPP/2UFlmRrIGgR0RaccHfeHNzA9vA3SPCPj8/98mRAo0TyQODVRow1B/IP/izo9oEyvKywqzrhmG4sHZSllWljWHbKPDkwvCsVnw1ZEIw6SYZygoh+G0xDdmJJGY9n+vr6xsGIDjcUkgRRVUAbUxmbdO2i8Vi9B0J4VcNoTzPu6Yj7Q0Dnwqfq0zri0eNSF4I4mJmEPPkCUHsuJxaa7Q2eZ73Qz+ZVHmeu344n04heDc4OkBlVWE/ZwvrnJtNp4+Pj2yaV1KeTqcQI3rfUYKy2Wy2z9vZbOa9y2xWloWQ0mYZBXFu8q7tRBRlUcYYo2ckOBZlAZ2NVAxsK4Xs+y4v8tP5VBbFfr8vq3K32xIZj6dj23bGmrIs94dD13XKSMYkAaGJ0+hms9nxcHh+fJpOpt754MJsOjsc9k1b1+ezMWa1Xj0+PEymlVLyF3/7t//LP/yd6P1uu+37Ps+yoeuqdKLGtJTnNkZhjN7vtyH4pqm7DopDKSWHfrBZ3rVdDKIoyuhDVU1ElMEFbK67rs+yfDKZ1HXjQ+z7IbOZlPKbb775b3/5V3/3d2+c8955Y1V9bkSU15vrn/0/f2GNjTEqKWMIWqk8y7uuyxlolcooJaIQQkJ4hhA5om3bMYTLMYB/owIAsMukmqMVjNkJ63piclsoy/Lh4W6zWSsltVYheAy4rq6u8ryAqPRJkO29R7pKBAO2cp5ZRp1gQTZK2qj72WKLdJXczKUTQrx48eJwOND+QY1JxqU4IIE555q6LrJcChFDmE6mfd9nxp7Pp83VlUqKW/4KKaWSylqjlFqtVjaz+8Nhs9mE4BnqlVJMJtVkUvV93zT1+Xz6P/7Fb2yq8e//+E/xlg4xQBAJEX2M0/k8xCCkyIu8rEoffFHmxtqh6yAchqGnWxtCKMscYtp+NPsWY7SZPddnPzghxNB2y+UyOlfm+WQ65SHy6IG0ZVlmWT4M7nQ6nU7HYWAewpRVoaQuimK1WhF/SeAxxvl8gVyMyo7+Upas2UAlXLzJZBJDwIIG3vbSS4jxeDwmk/B52zakSRdcWeWoL4bBGWuNsVFIpVXfdX3fE3BHLQoNLshxk7w/KZVCmm+gqHr79i2qJ5AO/SXyhEzboWk/0rTok/kaXwc/L0IbnYzT+TyZTow1NrPMPYS0L1OmUSnaawxdUx+QbmWS04HUyNl02Ijap9O5LCt9mSEKMQot1WwysRfTMVHkxeAGGiQAWAi6sSHft33XtUKIpq2NMc5RpzosrcDyeVpuSmcOAoEPI4To+955n+X5ub749YcQGKzp+j4KEaUoytIF3/VdXhaDd7vd1lrLnLAbhizLbFqYihETtjk+Dce2TZNneTWpqqrK8jwKEYVQUlplGKfI83zo+xhC33UhhBADHxgMDhVWN/XpfOK/HNyglDyfz8vlQkqVF3kUom4a59xsPmcHqhCX/nBMvvR93x+Ph77rXty8uL+/b5tWay2FvL6+ns2nfd/V9fnp+en5+SmKKEQsikI4//79ux/+8Kv3798F7601eZ7FGIh01N+M5WsttVYPD/c8Uu893D3N1aF3MUQRhdUms5k1NoYoRGzbBvnDYrGMUXgfrMm4tlmWnU5nY7K3b9/97Gc/G9zw/PS43++/++Wb+7v7+7v78+lkja3KMs+y+lzX57OMom3aaTXBY9EYhgCGYXDL5Sp8ZG1EjObychLgCYgb0PoUx/Qap9MpOkCYQCGEUnK7fZ7N51KIGEOW2fl8FmOcTKbv3r2jfwOlSV2C9Agx2MeYFYOcx8dHMNPYmGFBMrUjIAwejMRzPB5vb2/5FovF4v379wi0MD2DJXPOaalQLABZMptlWXY+nbqhpyVMLX75VFprZGBSzmbz3X4XlZjMpuvlYrt9JhEi6/feaa3+93/+T37T9PAf//OfS8mWDy2lMsYqxYBkHpNlG7HeGJNZ6wY3suS8Kswp+W03NzdgJdqJzru267q2k1KKGCeTSfAeZcWoDRjb0VmWGWPTFHQYBhaN5ZPJxPvLioX04hVxpChK5gxFMtQ0SSU9DMNkMiFp8+6puYClKMmWyyVwACqQKvXCehkzmVZWWymliFIplWe5lDI4fzjsKRupK7mHY6d67DGgHXp6euJpnM9niJrRWT5+ZDdC9If2IZQLIagqqIeWyyWHuKoqVvixUIwwym2hQxXSjChM/cgsUb36ZKwNNiEDcQ0YN+EvHRXWyGox/CDc9N1lmRJezavVygffJ/swOjQ8ZL6j1dnxeJhOp1ojMtbT6dS5QamL7paTQC3Vtq2Ul0WEEMrc9rquhZSYv/IVuq6jYiO9cZ8fHh4ov5qm9t7HEKuqapumLMvz6SSlxDWSgXxW+FGHbbfb+XzuvCc9SKWWy6V33igD36KSb4eU0rnhVJ+rtCuXjDifz3F2GXVEvAtrbZGXJk1BDsNAjXg4HEiO/FqARV3XoC4hZFGU+/1BSlXkxXQ6Xa2Wy+UCo4jj8cinEkIEH46HA7+W9VYkPzAB2ZfDySXt+77remMum8lB5U9PT4RfxAsjPOdi0r4m2cNC62RfmGXZ3d0dTh6n02m9XsGuwM22bUsz8uXLl/QPiOa0ani/gBgE9HzC0aGrLEt8JmiYQSYD+TmoNJCheemX0NVj6OTFixullIhCSrVcLs/nmknnyWTqnLu7uyPBsz725cuX7DeUUoJ6wZ37/X50PsYs5Pn5mVoB+SzPZ7lcQjZymKk8uq77+uuv4YfbtkU6MYr9yBZ919vMKqWKqtTG2Cyz1g7eCSlVGrUj/AohNNuIlWy7TlvTdd1uvzfGZKncyfOcTh4TGP/bH/w6ht7q4//gnOf/cF9lxZhOPubysvcqUjo1TQtDB6vufzWSPoyUrk+GE5AqKlmZc1sI2fIjix6a+7xXHvcIb8cmM5wSrAVKJJH28BCagWB9cmwf9Sdk9Y+IlZRZAAAgAElEQVR7O0KI0cxZJis9RIpjlubFbJ+3MUbvnIix7/rgg3fufFkkJ+lL2zQfZ5JvF98Lsh7wiyU1IYyWJomED4/oAt6ce1WkbT8+WYjzEDBue3p64gPQ5RNCUDzB/FCGjxkXPE6U5we22+3YfxuBGIkHtoRrPBYWQgis5Zrkpx9iCCH4ZEswDEN1ESMO49snEDD+g3JxVDqNwrCQNHIj3Bvbg4QeuAUeiEk+URwDVDSEPP7h8BDdtNbsCOGN8DTsRzu/OGBEPYSMLPxBGCPS9vLpZIIujpERXhmqbtJ5TFZdjL+gECM1sgR0uVzyp2hi8Vdj4Mz/ZJIPBM22tm3v7u7evHnzy1/+Ek/M/X7Ps23b9uXLl59//nld17it9H2/3++bphZCdF13dXV1c3PT9/0puR7d39+PvVBq68PhYIwNITrnrc2EUN7HLMtjFHhHcx3IKITykFRDY3CH+u+6jr2NFJ23t7fH43G73e52u4eHh1EFLoSo6/rt27effvppn/x6iZigAeDLKNQhR2JN8fj4SEgF2fR9/+HDB7AUrsCMRk8mEwic+Xx+fX0NPM+y7O3bd1rbEIT3oa6b2WzufdjvD4fDARTfdd3z8/PhcHjx4sW3337LNCvUCHF5vV5fXV0hS2M6QV1UnfLp6YmpCBzpWcngvd9sNvSl9/v9drv95S9/+fr166ZpXrx4QfKAZXFprbTUyocwW8ylUkrrKETvhrwoyEzDMOBbRX7iALMEabfdTadTKO5f/vKX3IL3798/PDwADtgz9pumByEuXrsiSq3NpJpKqUIQOi1XGT8ZkgxwEOgSrUuWDG8JuJPJhLY7V506g5vs034CHtCo4+TyAHzGkUXCvUp+qBAOIvV5dNraxt1gMw//8DGIwnyGIrlmElNg+ZumeXh4wL/MJY/JMV3FGPfb3aSayCi8czGGGEKR5SJEPufz8zPhlYFzGF6T3DVEWosGGKG6p60thBgJyizL2LhA1BvbxYzdEtn5ZzRRaJMnJX3FkaIlDnJzSC2wPZemtXPQoJxOnhLrrkZWh8/GjA/zHDz/mFrrIakDY4xKScRz5HUhBJCWH2B8ySeVzmQyQevJxaMqH58Y8l/COu07Xi59wiFtrKRK438di8i2beklQLJjc+TSvgGCNcKH5XJJUgR9hxDO5/PV1RW5mVs9DMPj4+PxdCI/bTYbkCMaa/KNcy6KCK3Hk4HTcM69ePFCpuk2/hQhDMTDwZjP59iZhGRRNZlMVqsV0XkymSyXy/1+//79+/v7+9VqpZTi3vVd3zTNj370I1gRFjYYY/Isp/54enrabDb4BZxOJypjniT/QqCv66bvh6ZpGWbKsmy320+nM97gaIXinNtut+OWvXHbDEf0Y7fXFy9ecPtub28fHx+hhlRa1c4bJ20w3kSIGAd9gdIUBCg+yc3L5fLDhw8UJdRbpDq4MkIT1L8x5sWLF+PLHdUuxljA0DC4um4gSPp+2G63FAqcNzZHXV1dcUGAL8YYxug+/fRT7z0taBajccghYGmJs/KAb02Z/vr1ay7Fhw8f3r179+LFC8ROKI+5sMMwPD09VZNqu9tpY/Ki8MGHGLI8G9xl8ndsv7O0MYp4Op2kEMvlcj6f8Xy+/PJLa+2bN2+m0+l0On337t3DwwOP5ddLD3+PXPo3/+5Pgw8xyqIojDbz+XzoL/NHJvWB++RmVVYlMgz+LK6oGIgSiJGCgI77vj/XtdIqhui9Xy2XdV1PqkoIUU0mFAFkC5G2m11f33z33Xfb7da5QRsdQsgyW1WVNdnYlx41Qs55mBy0XMQ7TiS5jd41sY/f7y7TEheDIAAdPiI8UNpTSRkVVqt10zRCyL7tnLuUSk1bA0WphIhE5BVy2yhcubm5ORwOhA+TbM/HQMMPw3oxai+E+Prrrym5+P8EoL7v1+v12AIFjxM9iVzANGKNtRaKTAih0yAhDCbsHFVCSErtcYb24wf1sUgUNDSbzXitfnDOuWoyZfJACFGUJfCWKDnSSiQh1182PYQAym6EEMZoKS8rV+H0yM1CCOcuM9LkWpl27dXNr2xi+cAkcpQ/MJyg+KqqlJJZlt19+DCbzS5d5RCOx2MQESrZGHM8HlG2EKaNMVGIGONkOtGsfc1ymt4sYwHEcLGF+pXewSdfQtIzq85ZL3MBK0KN0pQxT3jvtbk08AmOvKa+78qygF+Cjt/t9ovFYr1Za60wumCzE2ppdZGVX1zLYMPBbVQnIxDx3q/XayXNdrtTSp+Op7Ios7yQUnnnBzfItI1ybEHFNDgGDcUpgrYakR9dt+vr6+fn5/P5RF4vigLLh+fn55hkbEyQ5HnOCQSMq+QPSMqnjGCuELZkbKGN2AKSQCXF4/X1tUhWVBwVwnTfOyEkhJIQMiE/RVOT4+3SlPLLly/xTQBnAMVOp1Nd1zc3N0xHc0MZkuAvyrIMuo/6EuThvZ9Op8yTIlpZrVYMXVtrqW+Y6B6GYbffL5aL5+fn73//+2TZyzTS4HA5AnY8Pj7Cb5+OR6nVbDabzmbOucenx7quywQQqfi7tNfrD37vx79Gevh71UOMgmJ5Op0VRVkU5TA4EUWbjDMpzCEB6I5CNYzUMLPdkPg+reLhGI11PbTJ6MlDSOKycURonLIpmu/JmYak41GO0JX4G9PkJ695SOthydvcQJHMMKSUo3ECoBg6tU+GgMMw4Gal0oaJGEJbN5mxl/aRECJEJeVYz7KqKcsyTpVLQ8jz+RzpBfQOeQuDUn5ybBVwIObzOcSI9x6g7dKKU9oJhEhSIOfSOVfXNeLrj9vCQoiyLJnZ5luP4ksODRX0bDYjr5DjcW4h9wgheA4AZGMMslSUf1JKIcWoPBZYlLddTHtUID3Y7jcigEvwTb7fInGJVHuUWSLtbvJpZSYHfVTQUk6BMYUQI50IoBtddBAXcqgArSA7Li2/cLm8jMv85V/+Jb7xAB2Cwrmul8slS5mQwYCa+WzGGB98SOPuo86ScMlvfn5+3ifLo1GqAGtBfB/TtkwKNPqikEJlWVJk7Pf72XRqjLm/v89sxvddrVbUE0ABY81Y3tHJhKDnsPGRRlmdc246na3Xm+12p7U5nc7BhyIvptMZ94sWF2eMXyilZO9In3ac8MH41giHkPB98cUXlMWkhKqqfvKTnyDt5RMyfs8vJKTw2z6mQ8k98Pt8cj6YSFbNxhhaGs45lOXQKXmef/bZZ6PsG8k1x7ttu6Zprc1OpzPX6ng8jub2zL5888038/mcrp5M3abJZMIH2Gw2DMwWRQHxZYxhUKPv+++++w6IxrVFjSaE+Af/4B9QKLAxdCxPQwivXr0SQiwWi6IsnndbH8N//W9/KaS8f3xou04rPSq1OOHooIL3s9nsw/sPp9PJDW61Wm02mzdv3nB6kSliLsLz+TVyg/ifqod/+W/+eOyn0YTcbrcmzZ0TlXCE7vs+s3Yxm8NIcIE5xyPjRPXAD5zP567vhJRucLvdbo6XDrNg5zOJDkhCxg4hrFbr4/F4Pp+LIhcSnYC01ip5MZKDE+CTD8PQdb8a9h7h6sibd1332WefsWO5T9MoY7+diIZwRabxsZH+FkIMXZvZLMvyGGJ9rpXSZVlIpb786kt6m7iGEX9H4oUDwe8h1oNqnXN4jYyOBazYI0v55PwMY/j4+Di2LqBoQfREEzqQeZ5/8sknICa+HU+AgThOv3MO9TQZIs9zVCiEKg4fqqoxPYw7DEi6sIJMeJHprbExBKmUNpoIKKQcaT3u+YgbhmEI7iKWVZolLdMYo/fOGEsVNQzDbDZ7eHiA4MarlXEKAOOlNtWapzq2mmiT8PpGc2NSXTUph2EI3r9//365WAghlJDe+yjF3d0dXwfK+Obmho06Ukqb2ZcvXz4/P0chnHNN3VytNzy9UV6stc7y7FxffJb4b4ZhILtnyauHcgE8PpvNY3IMVEmbb4yR8jIQK9OyGu+9lKJp6izLY4zgNiHkw8ODUnK5WnLvmqa5u7u7XNjlimx9Pp+ZHGKMH8CBKmy/3yd2N8zny/l8HkKk5rbWam201nVzhrQZbzf8cAiBYUmfnCUpf3lZKi2wOh6PX3/99bfffkN/ous6ZpKur69HJwJo0nEWDIzPsadC/bjpeDqdxpQP1OCDUSnCWAgh6H4tFgsSNnUq7YosywEKY03MZ5hMKiFEmxzajTEE4tlshryVD8DNHduWY5DJsowS4fr6Grrs4eGBD8lL994jBiMeuuR4AWnB4lKELYfTcbFaUhKFELTRXdd1fbdYLurTeYTjdDgw/C/L0odwPB6LqpRSamO01ueksSYJUep57//ZH/7j3zQ9/Kt/+yd52krRJTu8+nyeTEqjdZHnx8PhcNjnNnN9H0Ow1vR9F2MYhkGIsNmsvb8UcW3blWWxuVrv97uyLJ6en7z3UsS+79q2qaoqCiW1UcZ0TYuKru/6Mi+11kqqrunmi4W4rIIQWZ4ZY7RW0+lUG+29896dzkcpRRT0R5FaRmhQkbZ7CoE/zCCSKI3rQQZSSbXm044glWSyZAWSfJZlz9uDsbaoyh47BA1PIH784x/T493v91w/KaX3DuEgnv5KybIs+r5jbzvMKfiU+6OScy/9OnIGJ5gbJZKbJiGYspSRBQyyKKtVMoz8+G7ENJhNGgD7N01DF65pmnfv3gGRzuczMhJAE2QC4EspJWV0blBKVVVhzKUzFkWczmaDG8b2TwyhqevM2Mxao7SI0WjjByejaNs2Rh9FcN5lGeb4DMNnkDbcAY4+4Ndq3TWN0VrEaJSsT+dJWfgYTbKK4oQQHYxRfdMIEaMPi9n0sN9lxhz2u/VytZwvFvM5l3w6nRZVKaTEjZmx867rqDxCCC76q+uru/v73/v9n759+/Z4PH7y6tXhsD+ejrP5bH/cSSWVliF65522RgjRta0QMnivotg+P0/LiRJytVhOplMtlfM+CDGbzXrngndCxhB82zZSiPliZu1ldBlwkyWbFiFEjMI59pC7xWJRFGWM4VyfTqfzcrWGp1JKPz48sRXHaKPkZdQ8ywprs6IohZCn5lxNKmX0YrmUWj1tn7QxLvjBucl0+snrV0/bp67vfPTaaqlllOFUn1xwi/m8bmsfvBRiMqnoXkEegouHZFvt04rDrutoof/oRz/85ptvOG+0wXhfgHHMymhsmDQlCgYF7qQq3JflpK5rKZXWZhhc27Zt2/V9RzUG7CDyws5vt1u4rNVqpZNGvKrKpqmN0Vqrqqr2+12MF8NUAj36GgCZUool4XQouaHYroikzsiSEQ7DawDKzWZzOp3Ghbve+5ubG1rKbJ2h+qdCGobhw4cP3KPVaiViPJ/Pm/XaOz+bTru2y6y9/3CXZ3mVdIBEhrZt5/O5VOrc1NfX17vd7vnpabPeiBiXs0WM4eHhYT6fg0eZAej7/g9//zc21fhX//ZPLpgoy3hbvPvgvZLKGnM6Hpu6Zti4Pp/BLHACzjnA5oVldm4Y+rIsd7uttfp8PmkhqLr3+32W5yFEbS/pzg9OSaWkKopCRBF8cM6t1msq8aoqm7aRUpblBaT0fZcMSAKLy72P+FCykrtOGzfBwkR83i64bJRS0DYE4+uPbE3BSkVy2VNKS6WWy9XIks3ms9lshhvM8/MzQsa0Qu5XC46q6uKH0TTNZnNFJ5bzMc5hymT9Jj5aFg0U+ulPf/rdd98RNdq2ZRfuq1evqLVlmnMWaYpylJwCGUYmSl/2GxspJY7oIu2NCEn8CodAhqDYJ4nayyC9p8ELUu77TimZ53mW5WOjmO/FPyMhJmJk4FMlCRkYHxyX54X34XQ6EiZ44GP/IKZ9sYA74tF8sRBKjXHqV6aw5zO3xVqDbInHK4QY6X7OD1HDJ58c9gBqrc/nc1GWg3dKqSgi6qMYwsV0q6mLtLTSZtZm1nmPLLiu6xiC1eZ8Ppd5cdzvb1+9simH7Q/7tutgpZ6fniC7jTHaaPrnFEajInwYBvrkNzc3k8mEvhqQ0Ga267qnx6fT6bxarcBwuAmVZbmcLyBmvffESs5513dBxOVqpZS6mA9eZjWEMXq1Wr1+/Xpcg1yWpRfBh9APg3ODNVZJpS4NMyOTpJgDBq7qkgs6NYH3fr/fbzZrWDK6KTQVxiGYvu8///xzYDhkNeEe91zETjy97Xab50Wf3EmJvDaZ+BZpxcXV1dV+vx8lGNZaFERQiMA+qEi+Qghhs9m8evUKrRQrIKkAkMUjAEE3gViZL06B6JODPZUBn2qxWBADCZvUmre3t2RBSjd4hRDCarUyxtzd3XVdd3d31/d93/VN3SwXi8xm3rnnp2eYDz7VMAyM4H348IHEA4yYz+dUMC9fvJRSSilmsxndafAihcs//t3f2ND73/3HP4NlY0yD+HU4HIiZoF1q53F+B/Ecp4T/KKUMISItcEPPJHDfdwNGldp0XTeZTK3NoAh2z1uppM3z3g02y7Q17LRhCvp0OllrbGb7vl8s5koprS1TEdCIzvlhcG5wk8ksxnh1ddX3PU4V9iO3xTHKwJsPyfdNKYUFJozhGFJHzQazFEppkn9MQilkQlVVvXjx4m/+5m9GRoVWRF2fx3J1SL6M1mYMo/VJjAvE8Jf9iBdbLtAK2eXLL7/c7XaI1TiI4w2E4miSlRgHlIJ3bLxfMH76hw7q09PTWNWOVqOHwwFJFXd4Op2uVqvn5+fb21vSf9PUHFYMztJXyPgwKhmV00c5HA7QaOKjTe58Lw4GnQ+0LocDy597mH2cmlAcOue7oe+6XijZ9b225lzXZVkGIQCYJAByrRQCjg41p0jOVyRFiqoh+biBwsjE19fXnIeu6/Iix5hvtVr99V//9R/8wR/84he/4GkgXyEqkd64rkKI8+kshej7ocjzzJimaQrGMrCA9X6335Pjh753H5nj4uhJvAMVAsNH/RUcb5M87elj+RBOpzMpgS9y8ZGuJtDNBEpU2lrrIMLheIC0nE6n9ACyLGuayxgX0qa7uzvoO201haBWSkklpRBROOfn8wU16Nh9FEIQBCHr6UUT6+/v7xaLBYkNjoWGuUzTlwic+CQo1oZhePXq1c3NzW63wy5pPl8YYwlwAHyevJSSRZaoQig6adGLtHaNeEWYJrv0H3nleu+PxyM9khDCcrm8v7/nFcBwMv4CjUZ9MErMx6E80DDHYOzxcPFhdAk1DLXRZ/JJd84drNM/6O6QCGdZxugGQIqV7GR9shQyX2aVyOgPDw9a608//TTLLKPjFGHGGODyP/+nvwXH1v87y6zWimHLGINzw+l0pIFOziCvjru56Zb4NK51OByYnIBBy3LrvT+djm3bdG07m818iKfTydpMaxNijDEe9wcfAhJYyswR3SNsKIqcHTJVVYYQY7zs9jmnXZvOuRhFnhcu2aaivRvFxRQ3MUYKLsCFTJtwgBKA8bF4ijHudjsCpZSy6y5sT5Z2fzrn2rbJ8+yTT149Pj4cj4fz+ZRl9uXLF7ghjWeRNICDkHOO1hmAmqPMi1wsFsAWlEjOOQxzZrPZ09MTYLBIm+mklHBBYzM8pOXmfDY+QEweSqQHji86QoaJxplkQBb8AL9wnIljSuhw2BNVifUcGKUUj10l717QAyAuS8uZge38KcQhdBq4ckopIS7iBXqt3EOVrD66viuKIsRLqjs3tVQaEQFZvOs6Y0xZFM45SulRNDUMw8PDA13W+XwOyOUhoy9EOYMWi89cVtXxeKyq6uHh4euvvybUcu3bZNrDN22So7K1ZugHeml922VZ1rthOp32Q59lmVQyxDhgVKUUvBnFFrT12MYc29rgjJAm3kOaGyeat22XZZed2EBLPphWwlqzXC7q+ixEDMHPZtPFYh7EZRcFZcr3v/99kNNhf+SvmM/n6/WaoDwMg1QX21F0hlqb2XSqlMzz4ng8/u7v/u73vve90+n08PAAQUQQ5NijytNaCxFHqTF9XWYpxlcP9gJDAOYAKDatfhqGYRj69XrlvcvzTErB/8X4K/U5r4McAF7uk2dalixUXfpnfIzwwIgm4H+o4RhD8UnYzTHgjAH8OfMizeUgKBqShTAJiQg2ugcCBKk5uFMjdOP8dF2HFQfk5/F4HOf+uDjOudvbW2IXKJauGxpCOv/W2u12O5vN9vvdzc0N/i58fi7RP/0nv7Hn0p/82X+x1hhryrJo28YHPwz9bredTWejsup8Pn/yySfgmlGKI9PaS+dcCBeg6tyQ5Xa/34fg2rbNMDDJclJIWVY4f51OJ+d9WVVFWbLEbXCDD+Fqs9Fa7/d7rRVE4e3ty8Nh74agpBoG17WdUnroB600WwYPh8NsNsMei3fAJ+nTqhBKxbFE4IQRwTmOdMNGeQkMZpZlUl6WIuT5Zel0SEPIn3322TAMb968ISpRonJoeCz0rq21mBhZa4Gi9JrAkkVRjFeCjwpgYYq1KIq7uzuq1Pl8jqa7TbaaRVo4TrIki3fJ9M0kR1u+I9qkYRiQ0yHLg/vi+6KkIvJCeW02m77vGVwfu4WEgGFweV4QPbmNI1UCrNNaoxHiS1En0eIj/af4OKBECiFgUwwyqKYT573SKsuzajKJIuZFsd8fMKYGpeZ5jnpExMAiB2rELMtQAEPCIO5iduH+/l4leoqeJ4B9t9sFEWeLOUA7z/O//du//elPf/r27VtYY1gR7tvoP2itDSGGEIq8cM5BIu8Ph+l0Wjd1CGFwg80yyiYWgHN9eLzUhUghyd/Q2RxX5xxDhbzQ5+dnpdTF0SQRhnme13VdFMVqsSB5T6fTu7s7FDtZnhtjhJS0c4nj6O7bpkN6hDgixoigsyjZspdJKbRS1hijjbX2/fsPP/nJT96/f391dbVcLv/7f//vPMA2GV1QAqbB0kBYpEM+nU5RK4GNwAoc1xGLCCEeHh4gbE2aLSflY2qk05jLWBRmaS9FTGOzVPngGO74SCFkWcaXJXBxy6SUzPGBGB4fHyksKNDJ2UQ5JGF0yGEC8492QsDuMgJJCqResckRmbiBkvCc9lGrNBG8Xq+RWl1fX9/f32PEe3t7Sxzw3rMGA6MOVAMMkKM+ANXVdT2fz2jXowxi0rvv+19vavrvk0v/6f+SabO2SLsyTqcTUIVOAzX1brd7enoa81hMuyl4Rkpp733b1l3fh+DBo1oqKaUL4XQ6eR/W601ZlcaYpq6bZEU3AuFhGG5f3sKraq200ev1Wil5f/8Yo3LO9/3AFuimaXHy4qktFovVavXw8KCS7HVUDZq0475Ly/bGv4uij5svUokghCDKCCFCiMA6SHnOonOuLLPb25dFkX/77Tc4YbH0CjAY0tBDMpnQoELMrkdNDrUC840hTZtTIbFoXkqJn/BisYB+5bIlmCa4DHyqsXVBwqbI4IaQTiBY+DFYghAC+kulFHBDpBW13L3pdHo47McmpBCC0J/nRZblWmuuGbwqpT3HerPZjCUmdxuJIU8eOFyWZds28/n8cDiQ/+AKmK2hlByLrUtuTjstRJp89t5n9jLgzc8QYa215BtkdZ9++in6NNgGvsso9LpQ6sGPmePDhw9oP0waTfLJYWIEjEVRyCiappHiAvmFEFKr8/m83mweHx+ns1kUQkr59PTUJ/Q3lrNaaz4nbBIXYbfbcTaoeHhlIB5OqVaaP7Lf79frNRplN/RFkWWZxS6s61ohopAxCtm0F69cot6Pf/zjruu0MkTD8/kMFf7w8OC9z4s8yzIlpRBCCum8jxdP5Wwymdzd3f3gBz+4ubkZiVPsYbTWX3zxBePTVVVpfYnI6JrGYnRssYxKLeIMRxGxIjk+z/Pt9tla03XtbDb13pVlYa2JMZxOF+USGYi7eWmcJCu2Lu1NA3mME/tjTuIn+75HXAvBez6f37x5I4TY7/eTyeTh4QHihLgRk4MvP0z+5hgDjHos3ZJxE9U5kY0eBvmG8p0+EIlEJm9pkQZTGH/j8zvnrq+vYWhPpxNeI8yUDMmTqrss9LSvXr368OEDbdfRlef3//FvbOj9H/74T7VWxmhqur7voBS0uvjKgfuACfhDDGk9HkyotZb1om3bdl0bopdSChGllMH7oihshq96XlWTvCiKojieTkLK+WKRF3kUMcsz5/35fF4vV0VRPD09xRicpw7N+n4Y+kudOP5/zjd/e5ZlTGyGtKQTUNB1HVdRKXU+n6EXiI8gZdIDYBAADqbjP7KeV2tt0uxFnudChLIs8CdgWIZrUBRVWZbAc34z3ZoQojFmrEmh6YiSxDLictLhGIpNwq5zbhzdAkqgwqJ7HJOadvxh+HeoZxhSRDsj7gBAYedArYPyL0szUOQYGIM8z8/nE2kVToCPvd8fhsFdXV1x+qWUhHKUb3VdIwXhfyWpjAGC+0DFE2Mwyfj++vqaq+Wc671TWgsphRRSyShiFLEfBilVm0xHaMjXdW20miR3Nq01SqTVavXu3btxy9iQdjHSuemTOxbweRiGwbu266jN0Rb//Oc/H2kBch4dSxIwj67Ic6XU6XicTCa5tfP5PIjYdR3Tgk3b2uQG2qcx9TFB5mk/I2GLNMA6lxcvXnDjCD0ILvu+x/AGXhFaj4EJJQTU88uXL/s0vqqUOtXn8StkWfb09BRC+Oqrr4beQRGjR+Dwv3//3gVHL0drbbTx3gfvu66fzWYfPnz4oz/6o+++++6LL764urp68+YN+Ga0eFoulw8PD9PpNISLTc4oLSN2A9F8GnXkwPC6iXEjV0ZqP58vvajRIAB6dhicTLtyj8cjRHyRNr3zy1FScDJBSJvNBpD++vVr+BwpJWtZv/nmm88++wzrDmIr+IyMxX0ZW48wsSJtIuIZUq/TX4QoBlqNbY+6rkF4ZFYIZI6iUgq8IqW8urqiMG3b9sWLF0ArcMBwMbJtUcTwSIknRLbdbpvn+dXV1f39/ZCGt2OMv4X08H/+639fVRNSFl0v6MKyqGhJNWkRzel04jnmyQyHtJiTDiQAACAASURBVCGldM6Tmdu20VpxV4uiCN5ba22We++x/NLpFcYYRRQ2s2PDYLvbrperxWJxd3eHTzjg4nA4Dv3FJoEbRVDjhMHqrFYrGgku2Ut8TBP7j3xegRuQ1FJKAsoYL7IsY9ZRax2jIEXz1ZIWwmujTqfjw+ND2zZVVSoltdFlUVFgEUGytC5bCAkmohuBIBr+BzURFDkFB1hgNpsBizjTzN9CUMIUwVYjiBYf7RGiyAATybTPABwxmhUDTlFb0d7kUWitYfMg6/qLX9uFfIO+SNN8GjTAxUbC8T9ZnpTJCDPGOH5BmLERgBujx1w4DMNyuazrum4bqWRZlE3TrFcrrU3f97PZLMusSIuvuWbGmPP57BBrKaWkxM9qRP0qyXyhvKiWiCYxKWQ4Hkor5y8NRkpSThct4pjsbPkXTksIQUSRXXYmCqON0dpm2TAM/TAYY7q+N9ZSULZNA6rN02roIu1Xl8n8jjzBFeNVEq2IHVVVxSDQj2VZXlXVZrPu+yFGIWWo67M2Gt1nXZ/LsowxCKm7/rLznKx8f39/dXUlhTydzpSedV2/fv367u5uNpu1fYvQiM15RmslFQUbxB05eLVaFUXx/v17ChqiLchASqm1GinEkOb2CaxUhCYt3OZCjdKgsezmjwsZooiz+SyKSAfCWiOlcmlfLNgfBhXwPlLfoDFCJLUXffjdbgfJNmr2mC68u7v74osvHh4eiEIm7R0aa/Gx40WU8GmKgvdyOBzIDTFNQfKucT3hhlIWo51BwSE/MtFBDQSv4NIKL/BoTPMf9IdAe9AbkEifffaZMSbGgE0IEZJE27btr9d7+PueS14KL5Q0SlohdTmZBal8FAhjCMRcSCllZm1mTGYMNZdQsm6b3eng/BCjlzJmWWa0VVI3dRe8aPqh6Z13YVLNyrxyvT88H4Zm0EK5tm/rRvhopPa9M1IZobBDKcsyRtk2vVa2PrciXuYSRNq3TOgXiRA/Ho8/+9nP3r17hw8S6gJmSi9VfxrWVWl2FETDOxZCANZGbMgPdL5zYdBWuTCE4DKrC2tUlCpIP4TghTF5iFKZrBu8Dxf6hYskUjurLHMhQtc1QgRrdVUV+/1WyshS6NG4m4TBnwKS7HY7AMjxeKyqiZLGDaFte+9933dChGFovR/ats4yo5TwfiiKLATn/UW9AL9Eb4MAR9mE1VeWZS9evODsWiuNEVmmu64uS+t9Z608HJ4Xi9l8Pi2KTMrY922e2xAclpD8ZpHqHh7guDRipPVijGxvt0nxGdOGCSGUlNraXGsrhBoGH6MUQaooXd+7bjjsD5m1VV7KEK9X60mRFVYH13vXixgurzgv6rbb7Q+H09lkuc2Kw/F8bro8L5UyxmTex6bpdruD1nY6nTsXsqwoiqrrhslktt8fp9P5YrYs0kZxksTt7S1hi1zOEeLAoIaqqipKYfPs5Se3vR+mq/m5b4tJtbm5HrxTRh8OB9f30fuqKEY2j6sEO7FarYzUwof2XA9t57r+tD9uVuvt0/Pjh0fX+cV0qaO20govZVDBi+Oh9l6IKLtusKawJm/qTig7mS32x7of4mK5CUK3nfNRGWv2+/3d3d2IErqu+9nPfrbdPRorovA//NGXTXuWyemyzHMRgpJSKmkyO5nPJvNpVhY+Rqn14XTq3PA/fv7zf/mv/9X/+PnP+74bhk5rmee279vjcX9zc4V7JzQLw2WjWtT7qLXN89L7mOel1na5XJ9O9WF7iC7OJ/NpORVe+N6HIRz3h+lkbnSmpCmLydB7JU1myzwvlssltdfpdMJnaeQ2IevqugYJjYKXvu8fHx+NMZvN5v7+HuwLKUJQ5oi+fv0aA74RREKR6TTffnt7S60jhGAklnro+vqaFjGIjROikgEU9B1yNeccFdt2u+WvkFKez0chQlUVSom2rfPcOtdz3bwftJbn8/Hh4c5a3TTn9+/fQeJ1yUfnw4cPNzc3r169aprm22+/HWkAbsevkRvE/1Q9/Nt//59jjMZmbdvO5nMp5OPTk3OuzAuqB74wr7zrOmrqfhiyPOv7HufOqixFFOBr2hXUInVTW5NNJhMhsG6+sBneDSTwxWIRQzifTkqqcfMB0ClGcXPzwtpsNLyLaaXlSGvotKkY4h7MQmzKkw8J/yLSegOwbVmWh8NhvV4/PT3xM5988gkWFwhJpVbaaGsza62S0mgTQpQfbaSKQvaXBWpGaZXbnEDM1uUxEHD9KIf52JSfeV6AvgkcJhmk00MeLW5AE2VZHg/suzdti5maxh8bAodEOBZM1l4i2vl8Xq/X3AGuELJCIPzv/M7v3N/fu4uvibY2M8bSgC/Lqmla74OUSkrV94OUqihKJmy77uLIL9MoytjbABDxD6+Y7tnT0xOz9yONRp4IIdB+5IUqpdg2GkIwWk8nU/ZFl2Xhko7A++D9xTKWErAfhhhjXhTOe5aUZOYySU5V7tMGeQg9ehKAwaIo+mHw0fP2x9ZIWZZw60C2kOTC4ETuzmWsYbuVUoYYcO4fkjPE4XBguqJPc79AAZFWCufGdl2npIoxTqfT8+k0n833u13bdlVZMg0wDIN3LrO2KIr9YS9E1EZfXV3tD/vVann/cG+tZTqnaVqltdbm+XlrjFVaUjBVVfX9739fSvn4+FjX52FA7tl5H773ve/93S+++eKLL56fn21mSpa/hiClnE6nNsuCD1GIKKLSKs/zfujrut7v90PfUVNy8IBZsOQi+SILIbI0gz2dzkLy+FIXN34/nU794MZHShv/8sMxaG26ri+KQkoFOYHeAd5fKfX69eshOeyKNJJNJ584ToE+pDXJyL3u7+9fvnwJOwS/x3QR0rILfFaKGu7ly5foXFUSwtZ1zSY4ghK6Kbr0lLkjWwhbBdasqmo2m00mk91uN51O8XGg4lFKgsKhPYuigMKZTqdjZxdES3FzOl3caEzy+q3rer1erdfrh4cHqFHsA7z3v4XqgZYIt0UnvzPzkT0n6QHSjdPALkBcrrgMLPWFYeR2cW+5CSJV8RDBxHSyPTenTetxeEYXgV2MEHBjX4veY0jeBgBk6BHwMk1p2v1d8oPUaTcR6RpCiegJswS4g2/JsgzHJIK70irGKNNGVfkrMzUzmUyKIo8xGksqGvAxNsmTI1yekpBJqj9qjQhYWAGDsDj3hHtMbAAv5WXfbAbbKISYzWaz2cVGdL3eiIuDf8dPUnqLtJZ2rIjX6zV/3WKx4AuCjz7//HPowT4tHRqbolxymZY4JssHCXk6sm3A7bEYZ9pjSJaxJG9eE7yBTCNC1BZcJJ9MAwniULe4dVKPsyWNBzuSadT+EHcuWZJxUS/HMpmBQ21zz0fNftM00+l0t9ut1+sXL14QjjmlZVleXV1VVfX4+GithXngOfDJdVLQ8kgZXaTly2Nfr9cmeRZxw0FzI60HtZjnObzcdrs1xozef6ArAhm/cLFYvHz5kjTDsAKzStaapPOO3vvRGMOlHbrffvstCmme8PF4ojwFiNCA4dshoOA+cm1X6xVyIyLsZDJZLBZj/yMmGQgSfpHmNInOtPSJVqPRKeU+kjDnXDv0Lob96ZhX5WQ+C1IMwd893vMbRvQjhGAymdMO6Xc4HF69esX0IsySMQb5E6bi1lr0VDTepJT8/N3dXdu2VAOYkEMH4a3AFXv58qVSarfbXV1dsdJ5tEhgHJrGyePjI5cL3WDXdZvNhs+M0SeqM7q2ZVl+9dVXWZbd3NxgYZCnmhXSeOyLYFJ5c3OTGIgyxcahS+7UvGhuwfv374dh+PTTTzFHoGT5NRID//z/VA+eptxsaq3d7rZSytxm0NBUTNzSEIKS0jBKpmTf9zQPyrJUCUgKIWhU5HkeY4hRlGWplLbWNk0Ly8wiLcKZUoqvyoFjNgSqJ89zNJ04nlPBEO6hmEJapz6qR+A9x5YpmjMi1xjcQ7La5tTScvjkk084YQgKy7JgRMM7F0OQQkoli+wya+29V1oD3PI8i0JopaS4eFcAkEXaJ8r5EEIwYJhCpKbqHPX1vFHuAE5K6HZCCFmWS6G222cppdKK7k7T1Dc3N0izgZ9AIaVUUZSEOZvmsUm3/K/IhEgkt7e333zzjRByGFxVTcqyGga3WCwnk2nbdlKqECLGBiHELMu1NlRpxCOiP4+XZECnhLH+McrATKLlHdKUNRGWgonPzx3GZYtbzWcOITjvrNFucE3bhhirqgqJAciybJIMgMu0+EGESzsR/RhnyaUNDbxlxKNSyihEXuYquayjzMGGc7fb0TvRWqM9AyjwPMmUY3CUqelKR5Q+1qiLH4MjpaFzzmrjk2cfiDJctuh0JrkO8yFjjEVZ2MzEGJSS3rsYw/v377788gdEKHAeqjNQpDGatwz4+Pzzz51z3rsueVp4H5RSmc2999/73vfu7+/KslwsFnS5LpDZB6amedGMQ5dl2bddCP7jWAYpSjU23kSwYIyRTdeAibGZjAwvhIDY9KPBl8sF55WRPnl6nDSwLPN3s9kM9E3XLU8bjcZhXt4Fvw0E8P79e/I31QMv7uHh4erqivfLaaGGmM1mm81mfIOkZ/BB+5F1JoFFCFEUBS7RkN5jyKIwAnKdTqfPPvvsw4cPdV07N9CPgeKDAAAa0pknRiFU7bpuGC6DqCPVeT6fT6cjmCaEgPkxceO3sA4INoD0S4YEFX482cC3LYqiLAqVtrW0bcfXsNZm1vJ6YprgFUKM4HFsAPDDY4dKJx9WCiXuJxAeCDkMw7t372DzOQpgEP44v5OikoMlEsxBKgAW4PMPyXOfN52lTTUxLUAHbIKatdYqTVryp5xzXdvyCmOM/dDTKAaceu+FkEzDAnn4bSAplRR4Ie3toSClTgINcbxijGNfeix4IV6KIh8VgWTKsiydG0ZjZxSiKKa6y+JuHZOpAH8LAZoL7JJfDVMj9Md08lZBmHQheYz5eHyECTsKMgC+S7PT4wxHjJFRNTIBr5IjodODJQqPba1RkkjnbT6fI8VpP9pvYay9/C1uqFJNyblaLBY67TIbzc9JzFdXV5S5H0dwY8zr168pdzBdB2wSwvb7PSK0yWTy9u3bm5ubsXPw9PQ0Vlq8PtItyRhKOqa9rdAmTPMS8Sn8gc+j6y2XYrfbjZ2bYRh2ux12LKANQjB/r0p2EX/1V381YvNRKs3tY/Qvz/PNZsPWgeVySbXBb3POoYJ7eHhAZU7/M8b46aef0is+nc9a6/V6/cknn6Cx2W63TdMsV0seJucNUmU80twCkVTyvD7mxXRaD0cDYHAuCnGua7a3VpOJ0lrpix6EQw7U67qOgD4GwTdv3nAHgeHee0wGy7KkRUHJOyoRmHjdbDbf+973jsfjhw8fvvrqKwIIH/Wbb76heCLgYL4LVP/666/FxQO0k1Ji3UGBCCbjl0gpaX9yVUddz7t376hguo+mzf/RP/pHzjnqJH54hNfUmtfX1+RjqkNcwW0ybCZHqjS88v79e7QMfd+jTPu11wH9verhP/3JnyulnPdwKcfj8Xg65nnu+oETT9EKW9d2HVqRQMGRZ1meKaXyLHfDhZKiM0ObXkgRQqyqSmsDCiZYKCVxuSqKgmkvkVgXhphIEkRnohIRiunzj8O6TqZJI7ggkfL4IJHHHyZoUiWMyAIsWaUVj3weqRRaeBFFDNEaE4Uoi4JgXdd11/daG2utMTbE0LWXqR/GamCrodr4Fhy7MTMNgyPJ8+1GwdUoNyQdwr+JKNDweO8o15hv59nCaMH8iMtSoxjTILpPWwp4yKvVinHNGKO19vr6umma0xGLULFZX8UovAub9VXw8XQ8S6mU1EVeIliyJhuDCK8GwpAB+D7NbFPhcm+RJwG1xh7MiBIAg0IIpfBN2wsR4Wo5DxgJWKP7vmvaVil9Op+kVFn+qwkGClyCqTGm67rMWO4VPCFpgwdCginSXvgQgjbG+cvA3TAMI2gl5j4+PhZp1whBnOpzFP8UaUHs2IHnEI59wiwNlo99MmgfGcUYMXFNUErhHBNCQHQLc1XXtVTS5kYqobUKwUcR5ovZN9/83fF0UkoxBqiSLGq322ECNkaux8fHm5uboii6blz4yh6OQQhxOBxuX718eHgYKX5+idbGOc/X11o3aZlakefWGP6bpmlev37tnCNrCiHoPvI8gQKsHqKi4l7wV/gQOEI67RNs21bEWJYFnBXBjl97Pp9vbm7SkKDnT+Hn8fDwAO4GoVNtwLBxJIwx44Z2Y8xkMvnw4cMwDD/4wQ/+4i/+YjKZ0AmAQUXLWxTFZrN5+/YtUOP29hbQ5tN47CjJ4yVy7KGR6c+DsIe0I8SmNSoyCSyVUvv9TmtNBcbQNRmXYR2TPPY5M+v15nyux8Ai0oxhll2cY0IIo2WOUuq34Ln0p3/2FzHGfhiUUoMbhmFg1XX0gczPxSjLEjKxqWsudoxBpeGOIi9w+aH0Hi0WpBR916fOpyU9aK29v0QrlVaDjbFsuVyiWuMfMif1B4x53/dARfTd3HCOL9Gf0pXKd7PZwC+JJFvi9BCax+lEWDwWB6JCE1L64J1zWqlhGPIsk0LkWU5d5b2PIpblRU2f5VlbX2xTEaeCl2PypCQ2hTSOl+d523Z8GPKB+EgCNCSPIMhliCwpVYwxxKC1stZorXH9LMuKIyWT8UuWZeOwxdj6Dmn3J4K5cabJGPODH3z/F7/4ZsTpDD2wro4GnUsGllVVFUXhnftY3k7khQDRyWMZKokVMVQkI61vkwkPjPzI+RBnpZRNU3NhUDDzxpfLRWRHcYzOB6nUYrGklOGX8wQIMXVd/+irH97d3eXpH7gvThRJJaZVDQoPdn/ZScfkKl1BjrEQYr/f813IFv3FWjVS03wMnGWyHRx5RZs2b/Nl+SMCjz9tRszI62bPT5bljBly78AQUUQc1MfYyuk6Hk9t2+JUimcRLMRiMQ9pipDP8Pj4uFqtYgw83izL+75vmpZjgNG6vBhrXpgKKQQyhCzLlsslbqxZlpVFQQ6o65ry+vPPP3/z5g1NKVAtIxcpwF0GPijxeSNN237y6SdP22chpVQyy7O6abI8P56OMQTIKKwmxqqd2Mfv5z26tA+RWwAry1/E+eevwyZnXA3JbDk5YD6fPz4+whP2yTVZCPH8/Hx1dUXvhL4CxaIQYpxD4ilRLIa0gpuVJPw2MB8jUE1ar8vrY4JBiMg2DvAWji/W2qurq88//5xqm7/aez+ZTEd/gaurqyzL2Eg6nU601vDSKs2oCiF+73/9h79pevizP/+vIYRzXYcQlFZCCAQ5fnCkJh4BhH6b1AVCyr7rpVRFUUghjTbBe2IckIHCOcuzpsFt8TJkqDUNxkidVZYlct02WU8XRXF1dQURSZThcIyJhCiGRpjzYdJu5LHkf/XqFUMDy+VyvV5D8XPNeItJL2SbpsF2P4Tw1Vdf3d/fQ+cpJUOM3nkQTWYzrVRmM6VU07ZCSud9lufW2uPxMJlOQ1rtwoFgyI7oAMVBTsWRuyzLrusRNrRty/WjAgABcfNXq1VytdPD4C5htDkbY2azudZaSgXRRJVKRMuyLM8vTwMsSflCpqGlxncE1Gttzueav/R8rlerFTxAWVYh+OPxZC1LTLvZbKq1yTIr1cXTSVzWrVx01khfEIP3ff/DH/7w+fmZOCLTKs0xpDIFOqoAuHJFUXjvjscjpbEdl5XGOKmquq7Z2Xc61WVZ5MnykzRPBuKyaa2LPB9BRp7M1KazqVKq73ohL4KcqioXy0UU8enpabVaEeLzPEcWwdNr6kZrXZXl9vm5mkzwtqImGC826W2MEWMdXCT7E5I3rClPbOj7oii00VGIGMPxdCqKvK6bLC9ijM57Ghtct8VyoY1s2ybGIKXoOhzo7PZpG3wYhh7Ku0/6rq5v2YLA306ajDGWZSGEwAxcCFEWlZRyvV7v97vNZiOEgDnp+14r7UOQUkFWxBivrq7evn3bd/18MS/yHMh4OByYEcPp4WMiji87m826ro8xbjabkMaJq6qqm8ZYyxgsB8NcXGyLoe+CD2PlB/fLy2Xbtkmj0Tx8Iq9IDIS++DgoSCdy1XK5vLu74+2MkXrcKkF5QZalc4aYgixFTcAcA6iInt9ms7HWNk1D4iHE+STGgcMISUAB5wzBRVuibdv5fCaT/zlRaLPZYNrx5Zdfvnhx/fy8BSX3Q388nMqyfPHiBXuQVqsVedGYCz/2/Py8Xq8BClmW/RbG4v7sz/9r3/fO+8PhMMV2TYrJZDJ0PfygSAacWuun52cpxDAMWik3DFZr4mYMYUhEP09HCHE8HgubxRCMtXmeaauN1Sa3Qokis1qrYeilFPP5zLnhYgYgJLpDYvpIPnDzOXNj0oJr4/3lyVeOf7m+vmaFUwjh+voalQjT1s65PC8nk2nfeylU3/vu4pHiy7I6Hk95XgyDa5rWu0EJaY0p8jyEYLMsLwpptTCqaRtltA8+LwoBbxhFlroCY2O873utLbQMVkXb7c7aLMvyEaLGtFQZsouiR6VNjSxKdM513SClappWCGV0NpvN3RCD94fjUUp1Op2ratJ1fd8PQkiVtgzBF1Njee/7vtVaTaeTDx/eCxEPh/10Onl8fGzaZhj6PM8Oh521JkTfde1kWmW53W6fne9D9Fluy7IYXD+ZTrgkQOaxdzIkO3GdTAntpVXQSimUkkLIoij6zmVZfj7V1hhjDX4DhACqKCS2fT/M54s8Ly4PLcvyosRPJcbI2Nt8Pj8eD3mROTd47wbXa62cH5wb7u7ub1+98jEIJcuq6t1Q13XTdcqIosqD8FEEbZTNjLaqrIr5fNZ1bQh+Op2E4K01xujT8dyeu1k1rU91W7evP/3s/sN9kRXOO5eMRkb9Dy2ZkcwRadBafST1hnIcxV3d0GtjqulkGAYfQ4ihd4Myuh+6oiyO52Ne5k3X7A47H1w1mXgfz+ezc0PTNGVZSWHybKql6ptGStnUTVmWQz/s9ntgBFiBS4FESghhTN73TkkjhYpBkK3n83lzPhmtZZTRh9PxGAYvYowiCC2lklqrw36//P9Ye7Ney5LzOjDm2NMZ75QjsyiRlAFKDYtGty1LkFtt9wD/YxvURMl0a4ANNNyA9EJRxcrhTmfeQ8x+WDuCV350VUIolIqZN8/ZO+Ib1re+tVZLzug0DpTyxWpNKKubdnc4cC5P5/Nqsz2fjkUlSeTdtOVyuVyuoN0JvibQGO/94Xi4uroCZA8pkRkuZtwYhx7dWsNoYgSSKgSKwoDyMNxGUYV5EtYY0ckBmgN4hZqMEAJtBQCPKHyhsYYxGxiA6HsYY9AQQu+FugHa3aCoQjF37oTGEcoiBYSEmKbJUsHl+aMSurm5AfR0ufRdt6CUwTInJSKEXCyW3gfG+O/8zo9DCInES392znLOHx+fgExYaw+Hw93dnTEmpbmAJoRgoI1v/R0otv713/5/3nsDbQmlUkoYAIL7gcwZX/gSp8w9xYsUQmDUCAS5QAeY72mlnHMCxoGCU0qFFEJwLZXPCumoYe3sO50w6SoVlvceRkskt/P4JDJbdZb9TEJIGQIDFUXfB6rZNE3ZuiZWVV1VFTaxKaXWGlxpsAVQbY3jGPNyTWGJMcYIoYwzl/VK0StYa2lKjDL4zGCglLsHheOC4WdZUitAXBk2loII2Q7/UWbv3FIiVZUGehZCEFKczyekSZ2FfGGqjieA54ZuFDQJYwzGFegt8ljV++zQgpkE3giwAtQyAG3xfgvc5LLpPMv+S4WbBI8wOEaI2XRslmzq+54xTgip6gqmmzFvFeBZocEvFSUhxFrTdbONtn1pJx688x61Ap4VpIUJoUIKdPEiyyvFGPv+0jQtIdT7IKRknIcQq6q+urrG/bTWMcYZ43XdMMaGy68dytqmAdW1aZpEEt4UwAQgBuCopH+6FQiAhWRyy3a7vVwukF1D6bdcLrngPkso4pjVTU0ImfsAY3DbpZTjOOx2z5yjUOCciavtNjhLcunNOVdaXS6XRKLLuynln23bdu0CDGkMRYHeGGN+9IMfPNw/AL+FcIBzrmlbH4KU0ju3Wi7/7u/+rmvaYRj6ftheX4GZgrfQD0NK6fb6+vn5ebvdosAnWVj+9es3aK3KiwYpqx/nIfCXL1/quobKOkIBY4ykFKNPMZKUGKOcC8oYxlGIsyTvYy4WC1ST6IRiJuUDMsJphwIQ9GVR8mJYhasNQvz9/X05hOhdgP1i4wF+QWgo8XZwNVarFcYkQHWur6+RYLByNE1ToXQXZAUHA9UbTikgr5BdEnBtldLXN1f4OofDPoYUY3p+fsbIEJz4t2/fjuOAjh8dCTaczufzv/3f//W3TQ9/+fO/NcZMxvR933YthdYNpcF5l/VZEccJIRn8DQKThKqq6ppQyjjHCcZAD3/qdDopKZ1zDGAObCahwZdm53Gb1Ubp/Is1TdN1HZSuwaXDW0xZew4zW1TcNgtoIzjilaAODXn3HR36OI4xBuzlcy601iRRSmmMyTkrlUSqMy9kNTHE+/UgkVJKqcBvmyYpRAwxhmCNEZxrpf+H2pnOqsIBoR/dKEAzHAuExSrrZ9EXNstVdn/D8/Tep0RAL5HZBxizHmAmiPUya2bgPgASTFkaK8ao9UyU1lmjBnAQ0kNhJcq8GurzMhHOPecc6QSlCk42HnUx6oA9X8yEt7qun5+f4H4MIT8ltTGGMe68RRFH8mIdii8ppcg+9SU7ooOmlOK/Y8THOQ8x9kOPeAS8EZiPNc5au16vwezC93IOTSopZNOUdauEEFdXV845NOzArJVWjDAQUYwxsAr49OlTSpGLeTaIEI+PjSk9yXpTAMFRqyIc48sCNry+vgZjynu/3W4LCRhDiIJK8ezgTRK52mxe3d3u988heEpoCoQQulosqkpLrXiWnKobmKm1c7ZQChQMfNOuXWDjzxiz3+9hLlvX9Xa9RkkB3g4wT0qpR0CPkc9G3InVvQAAIABJREFUPTtK6DRNXAjoDYNQBOj/+moL4QDk73KQxnG6urqCUTEuZtu2ddMYY3/1q19973vfM8Ycj8erqyvsVTw/P9d1PfS91jqGICWs5ThGmEDtsf9cMGc4nqJvxmIB6tSmaW5vby+XC4hhGAIDYvXZRGhGI6uqrut/+Id/QOwuVDo0Qzjh5/P5+voae3AAjadsF+G9x+LCcrkEAQTtI5IW9BZJFnrY7XYg4Nze3mIzFw4/EGTDJ5kv4GoJZ5qU4jQZIRTKhYLsWWs3mzUhZMpeJggITdP8we/9i2+bHv7jT382jiPj/Hw+CymFECGGqqoKc7xQYgDzoQSA2ECJ2iXYUTrzO8HMa5vGew/xmRBjVVW6qgghKcztCDDckNcpvZ932QE9T9MEpRSSt8xAcse8V2Vt53LxMIMqBQhSMTia4zhyzqqqBpRf17Wz8+G49BdQrVmWssKbM2bCMGr2VwE9N4voIqwXGkz0AX0AKuWU+SoxQn5u3kKqqgrcasQL4PUkL2OTrM0phMAoFdcsxsgY9cFprbRWhCZdqZhCzBQ3pC7UzmBh9X1vszIMPgAhJASPZwi8XmR5mZfpAX1DVVXjOAxDv1wuQvCHw54xermct9tNoeqCtYlf2+32w4cPv/zlL1E143yfz2f4+opZWtwppUiimUoUUgLNfyqnCDQBwOiY7yGZOedB7sZrrfKeJmVst98ja5bVeufc8/NOa/369WvwcfGyFovF0I/DMG63VyFE1GKEUMxXrq9u6roZh0lwmRIZ+pEkIjjHrIvkKZdS6nn3nEii2VrcZYdzhB75QpwVbC5AiC6rV+GVvXnzBqg3SksEF5Z38kk2NkD+HseRkLjouh/+6AfjOByPB0roOJr1enM5HVerVVXXlFJANxRSg97iJOAtI6E657wPKKqwpQyQ+ng8no/HH/3oR1++fGGM7ff79Xp9PB6Fki74p6enuqrMNHHO7+/vl8ullGq33wNjIVmibhxHRghoxLhKYANO03S59Cj7UMEAaaiqqmmbz58/H49HpdTT0xO2TPAAj6djpTUyU0qRUQK3GExBQCFBYMUjrbNjFQ4nUCY0yldXV9BDK24xCCCA+PDb0K9j6erz589ydkuMJU+UeoIxBmPzwkSCfkZK6fHxEWueNzc3ANmQI7GoW/TDscoH7EsptVgsnp6eKKU3NzcpJcSH6+trQsgw9PcP9zc3N09Pj5fL2Rhb1w2ilsi7X33fo+jhWcsHF3O5XP5v/+J3voP0wBjzIYzjWNVVjJELUVVVk1erkJRQ2yZCYoqM89V63XYtodR5P07TS6hns9mAht/3PV6w0ppz7kO4vr523jHGxn5gWYEd0QpH1jkPtBQM9MPhABwfMD3G1AW4BIYbY0QVprOiBoAd5Hy8A9hCaa0AK6VEuq5jjOO4JDKzejAGQOcoBCy1Z4kI8N5SSlKIFJPgIvigleLA1Mm8CVyCBeLI8Xj0PuBU4XKWIAgWYOFdIDfkuBnxENC1sCwGhwZOvrAimSFsY/BD0ErDdxeHzL+QWaSUSilE1kQDIwLtRYwJfm0oXvBYYkzn86mQiFAW4O5BdB7UJpY3md+/f48aDXx/hG9CkhAshNg0jfehruvT6ayUMsaG6GH+BRUTnMa6rrXWoEihysOnJSRBH9BlbQxkjm6x2O13Lturiewc3l+GAgwi16LBKlhQzLrrKJOFEE9PT3d3dwWnvlwuzjqWdXPRx6AjOV/OjDOff6GSxcdG7sQhxDUunTHJ2gF4U/iyBcmEimf+snO9RbLOvLXWGnM+Hd6+ee3sFHyYJhND4lQIzrTWi9WyACxKa0LIOA44zHVdo1wFEJ8iMcag6gJTHJ+TpIjzcDqd4OtZVVUipGrqy+XSXy7OWimk4MJa23VLIWdJFcAaaLmCc4CwSg5AVW6MBaKF6g0cre12K6TAAjmoJc/Pz8458Lm9c5SxELyUIoaQUrLWxWxnQikFmQXzHtSIhQUHmTxM5pVSu90OIvOfP3/GE4aK6svyDrE+pfTmzZuU0j/+4z/SrL1YCND4vsfjES/Oe7/ZbEAvxGeYpgmKDChf0LpVVXU8HtfrNRIMDjaiGT4GJkPPz8+IAIUdh27YmKnv+6ZpHh4erPPGWCRdkXe/YozwZcHBw18Kbud3wFz66Z/8ZQjBQaVOa+8942y5XNI0BzJA+fPsMUbrLOAdki10cC3L6ceBhqQJvLSkUkqpYRyvr6+tc977sR9QJbGsJp0paLOiOmIxGEeAttpsHy+EgF9VRqUtXgkGgDarLYJRgNMGErdztmlaKWXfD3VdL5cr0JakFNYaoDQoM3MROkvOmaxWTbJYUHksCMF4KzKLU6FaB9hirZNZcwIdZcxCkjhSIi+Iirw6VzA6mvcDpJSMkRgDIUlr5b0jJFWVHscxJVLkLmhmiwKaQ3YX2diEcw4NbZc1M/BJnHNArhD67+7u7u/vQwic02G4EJK22w3nrO8vjNFxHJbL5eXSx2zmjtAMfJZz/vHjx2EY1us1KusYIyEpxljX9TCMjLFhGPO6qaKUAPwBdQREJvzm4rUrZsFzhl4b4JLINiEhhqZtnp+fUZEhZWqt+8uA/RgcJFyk3W5HCEuJgCYARSljrFLaGDsMoxCyrhtjrBByHKdhHLRSAOhBO6lnq3Azmillc8OXoZAxBq6hyoJuwIgQWfBCMXWArFsR1kb/MWSDe2QyNB+z0U3w5/OxqlRdV31/qaq6qprT+bTounEcl6sVKtmU0mQmrbXz89ObSSXPz5hXkTQToE0WuQIZlBEyjiNgT4Bvm81mMpOua8bY0PeM0uBn5iHnArYcyEm4jNM0mWkkmSU1TRNgq6ZphJDPz89A4VDzwQmnWyzAAhJCrFYrINIk+zuZcez7XikZgm/qmjE+DLMrbcqsYqQlcK5S7i141vNn2TwYV3K73T49PRFCEE/gYAoTTJ2VfUMId3d30PDAZUfqcs4VlRrcJnyjwpIqA9HD4VAKRJwBQHmY5BUyPaIZpta4uRhgIP9BoqJpGogyWTuhyAx+np+32UXVOaeUpNkgEjETIfp/+fEPv216+I8//Rml1Fg7jqNUSkrJOKOU1nqWkCuuL9M0RZKY4LrSQgpCqa40oWQYh0rrGCJuJtAStB3O2hhjVddt247TuFgsQoyY+Imsfr5arcBxBDMBfxfQG8wtgTOUASOw2pBZ5M456M5X2fgejGP01GVqMk0TIUkpLYQwxhpjpFB4u5QSaw0OBzp6Mcs+R5OF/4C5xxhJSilEmghJSQoRvJdCVErDn4tm3Y6Q5UuHYSzzA5QPZS6NS1X4jmh6SvdQEgMhhJAEXIhkvTPOeV3XWlfL5fKbb75BGxuyBVURwMF3L0q/m80a5w+FcLk85/MFhxVCQ8hwbduM4xBCePXqFay5X9yWGQoDGCXzZsyrV68gHeOce/v2bV43mb16rHUhhLaZYV8sjkCgOGbjijLfXq/XGPKjm8EaOyIO8GXMnMdp1FXFsyA24qm1VgrlnEMTA7Tk+fkZmgRop0gG2Vxep8fMEyEAJ+rx8VEwBi9rNHyANGMMGE277OKAu4r3hSOHiFPA8VKKItxDQxRQOP5XtL/oGFCrIuXjh6xWq+Fy3q5W1oz/7J/9cBpHmsiiWzDC8ZGkVoCSOOcxQQnNgdPBs3AFWKfTaF52n+jI+76vlMKIFacI7/f27o4JjgHA2A8xRotDqzUCH5TmAJefTiczjvh3fGzUZ8657fYqhADXQiklYBaUU7qqkFDBXKKUIoKr2ZsIwmtsGIa7u1chxCKVyvNOHDL3ZrNBP1SWrna7HeYunHMsneEJI/99+fIFAGnJ1iTj56gpUVigXqHZvwgfG9LxXdc9PT0hUimlQKMCpIxWCZ+NMdZ13ddff805/+qrr1JKpUxs2xbrwy/vLwAS2Dx8/vy5buZx43K5ZLPUnHh+fgYGG2ZTkBHb8qhEIXFWVdV3QGz987/4K0II/NM5dOgYk1xQQlE4I+ejcaOEcikA/m43W61Vf+n7S79erTiX1jnnvK4qrXVV1YfDgVHKOW/atq7rmCIgJkKInX69bYt0gkqqrhtcXXTEh8MBZYLznmfMh1IKbFpKOZnJ+3DY7zE+YZSGEDDWx8vGe8JokTEOfflxnIwxbTvLmDBOMWwoiCQ6Te9nsBgvGEcnpRTDrBWKZ4LyVkiBqbvMeuAQKIVULQrPYRhKFMCoAMgyQgzLlrmEEJx+1CAYHVMaq0obM6UUheCXy5kxGkJq2/ZXv/oVbiOCHQahbbMghJrJMMopYykmIaSQvHDAUGv7rIdBKX14ePje975nrV0ulyg2Oafee13ppmmstXDxk1KBfYTAhA+J2Tto44ibOL7GGM5ZRrHIOI5t2+VSIHDOcU8KYQMEcIzE1+v1brcrUbiqNAp5XAlsSAkpTuczmHy4G7hLwUfnHEbT5aKaycSUBBfWOikEKFecc2sddA4wp00pPT8/N017Ph6Hofd5vaaAnN47IbhgnFIWQ3DWdd3CWWcno5SutCYk0UQY4xTasbl2KfUsiCUo5wEvAFlCTVPQxRDC+XxGy1Ip7YxB1/X9r35TCBVCbNt2mkxVVZe+B0XHGCO4mKZxvVmlbEQKPPZ4PC4Wi7pu0FQBUZmn1slvt1vG2Ifvf/X49MQ4u/QXSknbtSnGuqqGvpdSNXUtlZrMBC44vBPwxtfrdVVV+8MupsgFd84pKQvQPwyjzL62AKullEJK6xzcXgH7gFeitT4dT2M/W41VuqKECM4pYzGTVqusvozQXHYq4aqGHgKZ5vr6GitQuH2Iv4vFAuY56JZQfQLwIYRgLIdNDoQC3BSfzRiEEA8PD1gkcs6N47jdbsfsFFLWY8vYHGAG1rbv7u6stfgj1lrK2fFwmPN6jDbrDiCAYE1d6yqlNE3mq+9/fxwmvCnESYxqTqfjdrul2aQS6gN93/8f/+b3vm16+NO/+CuTiQ11VUkhSCIplw82+6ZixYNzTmLSUtW66tpWcEFSGoehaRdC6mGcQkyL5WqxWvsQvzw8CiFIilIpQkiIIcaI0S5oUUgzpbhWSkErQinpnEUoDMFzySc3Uk69d1yIFFNdNSkkJng/jf04XIa+HwfKGf5vkRV+Yoz4zBBBS4mu15usJWAZp5QReNuJ7BgK5lJW+U9jdp1F0AFQTgnFQUEHjaELE0KpCozJqqrBRRqGEZBUGTmWQQ4SEs6Bz4oXCAQYXOO80vkXaIgK9tqEUOjlEUK6rt3vd+fzSWslpfDeSSlWyy0k/0LAVhQNIWqtE4ld16LxAk6N45tScM4Rkh4eHquqqnRzf/9AKda1aPBJ6zpGYq2nlHsfkS1CllkNL+SzkEdBASRYkeECUuGIjFWlnbfTNCI7ghCCyhrjClRPqBtQQ9jslUYpS4l4HwihVVXHmChjkxm6rkXWZIwqJatK+0Senp+5EFKIaRitMTQlQsg0jTGFtm24YJxTKUVVaWsNCaFS6njYC872u+cUwuV0lEJAtwd9TMgaHt4Ha5ySOvrIKK+rmlHWNq0UwoxT27TTOFW68tZJDvt4KqWglOz3h5SSkhVnwjlPEh2GEXrp3ocQopKaM+FdiDEJIWNMlLJpMlKolBJj/NJPz7sTZdKHJKSarFFaJ0pjXg9y1vaXS13VlJL0YtyFK3Y59+9evWvrVik9jSNlrGoryqlQPCXSdC0XYrPdjNPIBbcQxaLcTNNquTLWcClVXXEpnfOccZ5lXEmWtOFaH05nXTcpEa2qGNMwjJwLrWXbNpSScRxgI09I8sGdzz24T6i+JRjwnJ+PR855ikkpvV5vUqKJMEDw+OvAfZ/HgVLi2MgsNsWKD6BzaAVA1Q1Z6xfIJ5gsZeJY0F0hBNpHsGBSSgCvStNZirnVagV0C6NgeLSgNsJUHOVXgQ0AlaMf9d5f395QzmJK1lnkIa3UcL6MwzhOE2OsaVpAoEIo7+Nut//+979/Op3u7u7Q5eSmdp5H8uzQh/v4R3/4r75tevjzv/xrNJh4xwAQgErzvKwIUA+tIuYwmAQgcxwOh/V6E2OyxiaSgG8AgyMpaiUhzY+JWUyREMLInIQhjYkvicYNIKzL2tfeey74+XJOKcUQGONDPzDGpBA+ehe89z6GADRQScUFjyH2fV+YRZgTUkq9D5h6gQKI8jnmRXl8DDSqGP4QMqumCCHyCDfCQosQCkzMOTcMQyKkquq6aegLf036wj4ac4gqa+GBJoF/gteLfAMOLppEOvNivdaaUgJBKZbXUF7ymmTeD8LcT0qpdY2yXSnV9xdKKdC2rpuV9XBMw7y/bQhJ0PlJiVhrORN4LJQSkpX0USvFvKsBjvZ+vwfoj6bnfD4Xr1PcLnzUmPfe0c9BjglPoxhTg7+ERgQ/XGel8cVi8fz8DKI3DqTLUqCUkcnY7faKMY7M4X1omjbENDsNNO35dAY3ZrVaTWYqmDsyED4bJQQ5D6MgDPPhG4wtXPVCbBGIK/oeKHqB8YJbiu81ZelvKSWhREi5WCxixMokA+aDcqT4FWJ+jroBqaj0EEKIFCJJCYu7j4+P5/MZknBCSCy+gM2Mgc04jpSRtpsfrOA8EcI5Px2OJLK7u7vVes05I5RCn5Mycj6eQbAp/BlA2LWuDscjZbRuaq1107aLxaK/9MHPi2aox7FgLKRENm3q2llLsgYiIQl4II7H3CFJJaRkjIFc9PHjx9VqtVgsBONN08C2ARX91dUVGvcvX768efMGHgwYsIH7x7MaI269z34PaBpwTco4Bx9Da43pNMI6YwxqmD4riaE0LFt+BXkTeQsqZLVdsHJR7ENyEQ2K1hr/Hc0irsDj4yMK1vP5vFqvl6uV9/50PHrvU4xt281XPaUYI04gykpk0NPpdH19DZwTDALGmLUzt0pkDXkMMP7nFFv/SXr4+V/9V5+Z7GBwFuijsGVclsYt/xGkGlwYY8xiuSSUmsmAY7BerlJM0zimYBljhFKlVNt1jDHoMJOYUl43k9nTDc+aEIJxGcp25xwXwlgjOOdMVFUVQzidzoLzbtHFGC7n8zAMl/O5a1slpeD8dDginwEsAsRfVdU4TmCSxRgxBSJZIAhRGFQBLJUsl0t0D3jciB2UUqVVPwyJpOVqNVljrI0kVXXdtC3nArg2EGSkE9R04Bd++PABDHGV9+CQI/EZEKfwB7G6iaEfndkIM2ZFCMHGNVIaOIuQ2sVxBzJDCJ2mMSt7Qw6IrNdLcOqB0SMVIa5RCht3KYSYJjP3LGTWAkKTh7iJFIvyB4kcXA6SVa0QVhD1EPLwrvHtkMiL8mDXddDDwbVEmAOqC+QXEqG47WU3PuY9TSFEJOnLly+vXr3SWm/Wm8kYSmiICRNjrZSz7vn5eRwGrXW36I6nU9d1GKSzrJeQYgSYHmPcbDaAI4BXgI2GjIWWGkfoeDy+evUKMQg2A5yzwq1y2YmMC55IZJx3Xdc07eVyGcd5MR5PoAwtgPjhFuAxAmmh81ibCM5L+HNZzpJzAcCtaZrHx8cyTwrRM8oW3SL4EEMkKVW6MsYO/Xh9fV3VVdPUIcZEMQabUkgAZnFrcDz6vr/eXhkzHY/HrmubrFCQYjodTzioJOv8OOeub2+QGru2rasZOO37/ubmGh05viy+XUrp6upmHMbgffB+GqeH+3vOWEpJZNMR/LaU0qtXr3a7XV1X33zzze3tLWYMBZTj2f2FZT4kWHyFRgVoC/tP6BVQ3LRt+/79e8iDxhivr69R2rtMocZQDUwT3G4ASgjNhBDgVBi1/uZv/ibJBGjG2N3dHSJA4Zdj3oDhttY6EXJ1fVVVlXeuH/pK6+BD2zQhxqZtMO3Hxtw4jrASenh4IIRgb45kLxmIvyGnYoSDwPIdzB7+5r/8t5iFUEDOK5M0lLQYHmqt0T3gsCL7lfXLrlsoJfvLZRj7ELzWKgQ/jgMDjxPEDGxgeW+MEYzjbiN7I4KkvPiGNy2yKzUX3DrLoPvIZYyxv/SMsuViMUzj+XwmifR9v1qvcOFZ3kP+H6CPy6WHIu6YTdmQnELmh6Ws0q6UWi6XWDrD3QPugZufCDHGoC2dc6S1XbtgjBXRQLw5klXNeXZAopTCRAw3H0glCnMgeEj7qEwRO/y8TzuvU7is9IeJCOoa4Dl4d977aTK6UpOZhORSir6/CMkhp4FwE7MvNEKb967rQEGBiNhsU5PSTKOqZrG8ESAvMFk4DeAL4jgCiUZpA3laFLOYUuBhogws04KqqjAMRx2NpW7kLQCDbrYOnCXlEfTxBFJKMRHKuNaVs36/P9zc3BJCrXVCzWUpTYQSinpQKcWF2B/2JSHhl7U2hYgaAmgJqId43cgWSIeEEBwGWMcguSLBr1YrlH14RCwv1dd1nUiEJsysXjWa8vbL0AtVLc1ctRgjqMmgMHLOGaFID2hJ0aItl0v0o1i6xN0BDbcf5o2KuqpJBgZiiNb40+l4fXOTUhyn0XrrvR/GftEtgNQX1GWapsfHx65ppZKzVnbT4OORSNB/8yxkjffivMeeV1PVq+UKg4eqqp6eHiFPBMopPo+UcrVa42ei3IYqNcr51WqFJQzQkJqmubu7O59PdV2DiwFsBymkFKzFKd1nkVSQPpBTQVtAqLm6utrtdkXd6JtvvsFEBKMLwAB414VaiUYZJTLeL4IYviM6lZ/85Cdl5AmIBQADwAk8KEBS5/PZeR/AiNHVNE3eufV6NQ5D3TQkRw/MX3EsMQZ+enpyzl1dXeGkSSn7/oLo4b3Hw8Fz+w7Sw9/+1/8f3RDyT2F24lECA3l6erq5ubFZECJkqVF86N1ud7W9kkL2/aW/XDBKYpQ466yZMFiOMUqlqqo6Xy7TNFVK49DXdY0NRhRr+MJ4kYBNxnEcxoEySggtrZyZjBQCk8/+cjHTJDivq0pw4awtrUO5wHHWpZhrgdPphHj08mQDdCqFPDyLio8pdgKw1kgoSYSEGIQUWHd23mHITAgptafK4tJ4pLjwd3d3iKqIDri0qHyxHA46P8Z9+CSZpcNx/hDRkKvAZUR1gysHrjelDEAHpSQlYoxRCnZDqQyTXRZKIoS0bXN398oYcz5fKKVN3RJClsulDzMFFvwNIE54bk3T4F8WiwWYfCp7qBFCwGQHKQhR/iV8j4wCDo8QYr1eI6q67BUBuCBm54AyjkYsJhlojjE2bUsoJYQSMtskXF1d9/2FC0EIQSGC+Oisdd43bXP36tXj4yNsIFPmwsUQUAzh+KGdx/mHdR3EUFEO41x574Hp4fNst1vvncv6cdfX18AZqrqqKg1rKc4FIST42YvCZ6dYvOXb21s0HCnvn6LxzeQZAuYFnjyqBEJIVdVCSLwXlFNgH5CUOGOMUiFEiimGEHxQSuG7ni+Xq6vtYrmc7DQMQyKRM45eTc3rnHGz2Qguhn5Yrzf90E9miimBzh99xDksORvEtnGarq6uhBD7551War1en47Huq6HocfBw0gg5S3Rvh/evX2LbhLy76fjEd0nmicQinCPgOBBuQTrDsCQwZQFtaFApmWkB2wHxRy+HSYcfd9/+PBhv9/HGPf7PeQ4wcQtS5cF4sYbwdsvjh0oevAfQVSbpun5+fnDhw+I/iQ3+uiVgd0hIWEO0Q89RK+1UjGEy/lSdsViSrA3xmW32UkFRTP2qHH1pJRVpdGUQ18Wx9IY8x1oLv3Nf/lv7IXKLr4zMidO7fv378EmRMLn2XsH/BPEta5r16tl3/e752fBWVXpbtFN4wAiSlXXIQSp5GKxuPT96XSqdYViTc3mBw6l90szTkxxQwjWWUjJpkQSIeMwmslQQgAu7Z6fjbVt0wrBhZTImwWYqrPZA6VUCIlaoKTDgqqhUiulHJg2JT1wztGy9X0P/hUEQpSUjDOltQ+BJqJ1hcQJZh7uP2INDlnf98XwD1AG3je4K2B84zMURgAuSVVVoIfi3iIu+2x3AXUjn0mxxhhKoU0brDWLRdf38MCaN9LRnYBoiPzkvdtstsMwHA5HxhhnAn+prtQ333yD+wnJOZSugMuAz6LBJ9kms3ykMhRFpwLGMA4SInJ50ahLWP6FtMcyEQtIi8yC7SCExCyqM45T03Z9P3AuYkyn0/n29jYlcr6cAdMPl15Jib/IWsuF+K3f+i0od+KQoNHhdIZ6hBBICaghpmlq2/bq6mrKlnN4dGU4ibw1jmPXdR8+fO/Lly9YWccepXMOvExCSQhhuVzFGK1xpXqYOwPGrLWvXr168+YN2gKkZAT9uVWKUSmJxgtdI4JmCJEQcnV1Bd9T/Oa6rilJgnOlFEmEUqK0NsakmKqqnqZJKumc5UIYZ5xzjNPz6QzuE1xrgCwJzsdhSClSxiZj4DjNOXfG4WzjY1xfX+OsRpJijLe3t/3lctgfrq+vzTQZY5bLRSZ9eYRRXPDggxRivd6kFM00HQ+H/tJbY4AIgU2OmQohpO8vq9UScCWgy8PhAEZWjBH5CZ0HpbQ0oMU97ebmBsRogKKYyrx79+7jx4+vX79+fHxEERDymhfqAJt1UEjWw0AyAHEL7T6SCjIZ/v3Vq1fPz89o46psMt80zatXrzCsnXOetafzSSnVtW3TtjHE8/kMUSNC6TRNb9++xQIEjLNQ5mK4hRywXq8vl8tqtVRKAZDHEwAa9h2khz//y7/2WVoOX55lLypc/t/+7d/+9OlTwd1SVgqKWU3eGCOF6No2eH84HDhjbdOslsuHhwcUwlKpEILSarFYJEKOxyOncx/Hso0zqmnkQMQCdFIppUvfW2eFkJQQkmhKSXBBCZVCeOeGfkiZcyW5sMbWTYM617+wKFHZBQFVTBkMlqkjyZ6uoJ055+C86GaDw3lBYZwmwiilNJFEGeOc4+YP/WznC1njPAz3JNNqaZZUAn6Fa4b/F0Ur2jVgYvhTQGD87L1Oy6gNWRzlPOR/RV65KNnIex8jOBXRexjn/VrtDp8EnpFID5wYpn00AAAgAElEQVQLzjk217p2oWdNQIrpC8nKuLg8bdvCjw88wpdQL4Bv5Ga8XJapDeg4rbVV9rOLmVGOfABcLuRCHkBt8VVXWfMc7YWUcr1eD+NIKNNaoxABXnxzc/Orj7/CfZ6GUXARs2ijkDLEuFqt4OuAr08pTVkHAkmrqOgg3W42G8QaQGHoumL2QFRZG+7Nm9eMsf1+j7cAcR5jDGHzjoWUSmsdwmw6D2Da5eU17/2Pf/xjlJaLxQJuMMDW0D1IIfA3QjrFzz4ZI34a1K7KrKLROoYoOE8xMkppIpSQ4+GodF3X9WQMIUlIGRP6Niu5hB8yWA+YslJCBefIUpfLhdD5jozjKIVEp4LeDkY9Ps5Ez65t97u9c2672QzDgN5XKbXb7eS8BGONMTSSTx8/3Vxf11VNEjHT5FDoZJV+fM0p24WhbMeLQBcCtyj0uCilUbkCitBaw74NyDncFPBD8D8555bL5f39fW67Kdz3xuweWjiQyFiMMQh8DcOAKcU0Tbe3txj4YWKHowWqtHMO0zXkclyc8rGXq9U4TYfDQUlpjJnGSSt1OZ8Xi4XI9dDt7S3clBETcAvAjHp8fMTNMmYC+/b+/l4IgZeYUvoOmEt/8fO/LaBYuS2ICIhNX331FVbwC8pctlsLVYBzVlWVD/54PODC1009DINUijMmlbLWKqW7ruOMH3Y7Sig6PjFreTZ435jeIFcBbFFKOetGY2lMyA3OOUpZCJ4xChKRkIIQSii9ub1x3kMalmWaOcI6viy+F5rHrGfnEqUhBBd8IkRIIZWKKQ5DDyg5BA8tMO8dpWwaJ601JRQWckrqFKPgMvq03VwJIdumOxyO2MJllHvvgFcwRvFhqqo6nw9SQp0pCcHBZIgxDENPCJQ2vFKKcxYCtrIJIamAIWVohuePgIX8jYFeVc0BaxxHrYFxTQAMVbZkATjjZmEZzrlAn05I0no2blJappSGodda13UFAzIUJtvtlhDy8PCAMA2VtCor30FKCHQvyI6gwIkxjOMQI87bzBZDRAb9wzmXYuJcKK2tMYvF8nA8ci76yyC48M4zyrDbTChr284aa6ZxvVqdjsdpHJWU/eVyfXVd1/rp4VEKPvQDp4yLeeEjppRCqOv6/v5eK+W9iyEoIdFylcqRzWqAAdAcmAKIm5xz/E6llPfFzZtOkxnH4ebm5suXLyQzwlNK4zQZazgXGJtpXXE+r8sCecApXa1W0zRtt9ubm5uYvZpVliwkhAQfaaIhJkqYs361XBPCtK6CD6fTSUi52axBlANizjl31sYUUaIZa6RU1vn98SSkgGOjsYYy0l8uh8OeEkoIcd6v12szma7rhr6/nM5t02LYlmIUgldVlUJ0zp2PR84YSWm1XFGS6qryzo/TGGIYh55kvaO266y1uH6YX+JgHI9HjDdQ6fdDj+kUIJSYolKqaZurqyvCaD/0MUWttbMWtEkkAJLVD30Wp0FRj7pQZ4Wbh4eHkJ16q6yGCVrd4XDARKHIQuNHFVIsY2wyU93UfOYos3fv3xtjEa+BWVlrQcZrmoZRir3l7XZb1fXxdJqsEUJcX13hA6eUsMEXQnh1d8eFeHp8NJMJIQjOI1bWp/Hm5qYADGVuh/ZdStl1HWImCLvTNBIyi40Pw7heb9BWfgd7D3/8Zz9H4Qm0N2ajRPSAxpjf+I3fGMcR85DCXEIBGIsDRoxciJiS854yJqSMKXEhEiE+BOSARddVWmulHx4eOOeoj6pseyDzihZuCwBQkRUsvA1YF0QbTilJJDVtUzeNVBJdTCKpqmubt44BGeGn4ZtSSmGlgn8OQ09IMtYkmnz0hCYf/Ol8quqKC269SXFW7HLOU8piJFVVe+8ZZSQRzniKyVlnJrvoFpdLL7ioqjqEKKW6XM7TZBhj1k5ZZjUQkmJ0bVsP40VwKbPqX9s2dV0hFkzTCHU+rdUwDH1/SSlqPbsIYDgBQK8wgpAh6rrGBHWaJqU0mEjGWAjHw3deZJtfPPAQAiAUIWYBbSwuee+E5IlErTXnrO/7xaJTCuaFoO0SQgigauccHISKmiwhBPgSmSnYrOtaQlLT1IQQUBuF4ITQMvQL2YjbORcj5VxUuuZcgnk1TdbamTtHCEN45UxQQimJ09Q7Z1bLRQjOWaOUeHp6uNpszqdTU9dKCmOmrmuVkpQSRkmK8Wq7TTHsnp+0UkPfU0pgPSIEXyw6qH10Xau1wjdFRizjdM4ZY8RaQ0gSgksJnZIJGoIotAG8dF03DsM4uZQopYwkClVzDJMLjgxCbVVVu93uRz/60Zg9MFB7AQcPLkihldJ9PxBCU6JKaSlk2y4ulwvnQkrx9u1b58xut6trvVytI0kw/YVTUiJEVZUPqR96IYWuNGNMCC6FsGYajd1st1IIKeVw6W+urhbtokArnPFxHMdhpIm0dRO9H/oeOtuM0fVqtVwspBTH4wE7DcZMulLDOPTDRWlFYkJ3iFOqlLq6uvr8+fNiuSCUPO+eCSUhhkSSj34YByGEVFJptb3eUsGMs/3Ye++6bCHlsxknGkfGGML9+XxerVZYjE1Z5xFnzBiDWQXaU5xbwCSAvPA7gSWUGQMT3MXgQ4gkKa0enh4Z5x++/5W37unpiWdfccQc59zt1XWKcbKGUrpcr1ar1e6w98GnAN+LeRtjtVqhtFouFuMwC8qh4JNKbTYbpdT19XXbtrihIFMU2Lmua7BM/bz9EME2bJr2cDgg5w3D8O/+6FsLev/Jn/9n8mLjgWVaG8pASun79++dc6DrApJGr4cE6F/w0IEe4PeIbJ+L5mC5XKJ2gHUfEiCGfmAEYyQLRAjhrwBZCNBAgcQLU4T1eo2XCggIEE3MynpCCMiN4ZRgoEfzr0KUdtaGGJxzjLKYouAixlDpyk5TinEaJ2wtMcaV0qvVCqS0Qndzv5b7t33f1zW8fCkQZ+ccoQl9n3OATdw0Tc67YZhwtgDFYNcRRXd+5QEIGxh1hS+xWq3A7QGOjwiLfxYcBpkDHxXTi9Ix2LyYDbAow2gBKX8cxyIjg/0jUERC9tellIYQCaE4oyCnI3qSzNPHb8PsEfNwwCzgaKHF8d6DRCuzdgiKdLSGKSVrzTTNMlBFgbmAS+M4tm3jvZdKGDPBSBlnOISQtVKoywLdGBvghKMQefPmDWig+MDQ8MAJ11kMQynFuShUNKgI21nUK7hsFQcYsCCH19fXWI8CsCalJJRhJ1RkP0SELZulX0C8wTB8GIY3b948PDykzPwGrg0VSEhb59l1wKMms2O577quaerHx0fnrNYVMCKSEbCUEuO8qlsfvLV2s1kzxlKKdV1rrZ6ed03TtF13Ohy22+1+fygi7Sj2SbaVxYXCCA33UQix3W611v3Q97O6V8Ks+HA43N7ephDLf3zz5g182XBmSoTFD8fFtNZKIXSlr6+vq7o+Xy7L5XK/2wnGcVaRPvGaEJHAFsNdAIUPuRwHnmXvNkwmEDQwWIJAhcgbA6hyyu1IifgYtFLBewCwmFJs1pvT6QQByq7rfvGLX+CeKiG7rmOCT9OE+UG3WFwuF28ddCRXqxXGBlrr/X7PGHv16tXXX3+9WCwKphpjfP36NThUjLFPnz4haiHc4fzc3d3hnrps0ItrjvsCnt4f/eG//Lbp4Wf/6W9InrrgDWHUBrs7KeW7d++cc7Cvc85dX19jCEOyRy7gSJlVg3BhEL4he6KzrSOUTzCDBWa9XC4BraLTBCQNKVD3azk27n0oHLJC0ri7u0NCctkuEQkcf8T9U20s0B5wtprMpQsw74yRIW8LmWKcpqmeV/ajMbl64qyqajAiQB9K2cTGZWOsEHxVVVxwxugwDowzM02UzQcO+kWMUWttllacRcJD3oFKWe0HpwFBE2+aZivpmRmy35d9bMQR8P/AlyB5JowTj8fo84YjIQQPEz+/wBdt2yJCoWXmLzZgUDGgp+RcMMYBduP3Y7ZcHEDx9xbOa4wBvfDlcgFoDsxdZ3oCIQS1Hv6gFJILHmP0wVtrbm5vrLPGGs5Y09SEEjJr2wkphVIC0ljlIZRNPQQ1nIqsSDivX2D6DcoKZuyYOWM9hWVdhBBiSvP+ykv2hM4OVEjGPG8nIO0tl0uZ/cW89zElDDB5FtUpdFig9qDxAHJRSj0/PwNzn0H8rgN44qyzxhCaOGcxRpTbzlrnPDJZ31+cc13XMsZOpyOl8wQFZXuc914JSVRXGulEa23MNI4jIYkyDtVeM00hm5TAWgrhEjNSktVkRd7DqqoKz0Rptb3aHg4H0A7RC6L+e/v6DSZGOG84h+W04HHhseDZGmNCjFVdSSlfvX4N8TslhZlmnwOWSeQhm6zgxOL654JsZlQjty0WCxj+xBeSECBPUkpBiCpfDQUrZo0Qacbd79rWO3c6nu5ub29vbx8fHxGLt9vtL3/5S845p6yu66Zr27YdxoEx5iAnyjhjDK4PjLG///u/RzS/v7+/vr5+8+ZNof+EEB4fHxEbq6rCYX5+fg7Ziq6qKvTrb968QczEacG/4GrjK38H3cOf/uz/xVsHBQWMEcBnmKJgcohTi7cyjuNLlPbl3Qh5dYhSCvIGsg7kGEFqRjWKduH29hbOvVjUQhEBZwyMCvHNh2HEUUNBhB8LYwkERJxFlpUyEV7x7EANXC6XKNVxaFBBI6w4Z2kiMSYIH5JElFKUJEKSMTalCFdeMqsFUtwKmfkwLuv4pxQ5Z1qrqtKn05ExGsLs1y2lRBSLEdMdkiIp81I0xQ8PDya7fBhjEBoIIUX+F8caDsYi26zjfaF8SCnt9/tClCovhWe5bJz+Um4g6KPvRnkFoggo5KWLx7Io0g/SQ1XNap24V+VUkBdaICi+6roGpxYmRfi0YtYtT3kIlHDh86HyJG/Xo03ebNbee2NNCD6lNJkpkdQPvRB8HEcphbUWTtFIh6VEQFhEQqJ5E63UnoWuCvIJWLb4zXiwzllIZiGPIp1zzjGXKsLgmRc3z41NliFKs6oogaIJmjM0K8g3pSyQUp5Op6I2Cl1rtH3wUaiq6nw6UUpgdGrM5L1LKU5mUlL3/fDu3TtjJmstdLyPxwMOLdBa4OOEkLbtKOVCCinlNI2EECH58Xi8XM5N2yHTd227e36WSvaXC6RzP336hH1msCQ452AVjlmnFljZarlardcxRdR/SAZYaqmURhvKMt3cZPkpPDqU+Zzz169fbzab8/kMQW8pJf653+8ZZSor8mIqWw4qThG6w0LxRMzBfAtbe1999ZVS6suXL/MgKsbD4bDZbBAlUlZiRrMeY9xutwSdcSJd00qpvPevXt3tn3dKqVevXi2Xy6+//hpwArBW2OS0XaeU0lVlrW3aVikV/TzWRaNQVllTSpgnSyn7/teLWQDzUVehdEYJhRT4gx/8AKP4d+/eoSjE4ZymCQNd6Ij8X//2D75tevjjP/s5bhHPspflzmMcRyl9+/btp0+fSjBC8472ufRuWIPyWeINeQyF1el02mw2CM14GWCOOudAzAIDNYSAlmW5XGLyDgZ0CMEYWxrAl+oL0BMO2d0a2AvMoTCeQisDB4KXJzJlCRTvvcl+A4ILSmkMIYYgBI8pTuPEOSeEaq3btkMIwM4LchXJ3UNKKQQPwR9QhoBol/LEWlNVVYyBEJJSdPbXnmvoY3CCsZqPuSW+I6hKKByEEHlhLYH7TLKOOr4RPhvJ+wHolgBeFxZmzJocJW2AmYot3N1uh04FuFbGMeY9A8YY5wLzC/SFZSEL4a9sZeNpp5SkFFJKiH4j5OW6QRSOFm4jfr8xU5gNwuIwDPDRhZs0Bt0A/QsEB15WWZ1DKgXA2LYtnj/J8Cm+C4IykgTUKYZhqOu6bVssmuVaOxlj0e/iN+eYnlBqIDGUkhPjHDyN0hpqXQ396Ly7ublRSgHRwpvFC8K3RmzCNBL4ACDc8/n8+PgYQlgsuuA9TOdTSt5bfFMuBOjg2+3m8fHxfD4RQjAqx/UsI1Y8c0JYTLFpmrZtnp+ftQbFzh2OJ5RuWqmu7fpLLzPzWGuNp4TDBi8WkNaQ10UWO9jt9+/evwdwj5OAvZ/gPDS4EJRVllktNHo0ELiwd3d3ANwfHh/Xm7XS2jl3Op1IiJv1GsXffr+HukHpA/D28V9wJeu6xkoNXgrizNu3bymlnz59ws1CKYyfiQ1EVANgJ3vvgYti1A8l0P7SL5fLx8fH7XaLotMYg6VUQogztq5rIWWMkTK22WwceKEhhhAgc/D582fYpkIIua7r5+dnnCUwmgCKPD09ee9RDqLhLlXOfr9HID2dTm/fvkXEQD9HslNAVVX/+l9+a1GNn/7pf8KdCVlAOMZYmCeEEJSNEFlDpMOZwC0CS6zKJmjIojFT3Uung0apBBq8P845lsjwpnEr8BlI9mnJBJu590S0Khho6TRVtgTB1Au9GIIF4jhSDrBp7z1CBoqg4dIzyighJBEl5dAPxkyLrvPBW2MxV6jrBj0Hy7qq+GAlaErJQ5jb/PP5hEwGLaP8VGFpYoQQbdswNrOYy4wH6aoIyAB1AYQaQkB5jmWcKntNF611HGswHGIWx8adQfufgdS5cfbZjgb9QTled3d38FjHZ+CZ6CylRApEb4G16pDVBea+O0NSKB1APBNC9P0lpYTWHoaO+Aqn0/lFyuFYwghIzHH2vYBXhLU2xgBdQu8dIQSaqdZarRX84jebzcePH1VWspoL4a7DG8F7x7+wvANRKnpUJ2BXm2yRhCI3hAiEer/fo37CkIbSGVklec9fZ303UGvKS0kpLpar0/lU1/Xt7S1aQ1QteEeITXi8GBLgV+ktrLWn02mzXl/OJ2B9KUUcb6WkFLqu68+fP79//845F2Ow1jpnx3HCWcWjQNMcYmzaBezK67oax9EH9/btW+/d6TxLXjtjKl3FGFGX4M2iDS2iXph84u7TrCscY1Raf/Px49XVFQ4baoKUUlPV1lqcYSwrIFBi0IomiWaKzs3NjRDicDhQRvuhr6qqqmtr7X73zBnfbrfAZ758+YKkLvOCPT7h3d2dMeZwOKAcAd8a9RAO9ocPH5AtkJUx/cJkW+YtWgigHo/H4MNqufTOS8GtMV3bWmMXXQd1Ka01cB4UK8vlMjjPGIN5H/qeh8eHpmmiD7ibGLt+/vwZruNzd1VVeE2oVnGcKKX39/e4X9CAAabNOcdwa71eY0kFkzNUveV7pZR+/1/97neQHlD++Cy+KIRALYZ7AogGVQBuQmnekSHw39HXM8a6rsMAAxMIjAoBv4LCvF6vTVa+w28AbavImCAevaQeex943hJAOENYubu7C/lXaf/fvXsH01c8L7TDUClBPsPFLhIRHvPw7DFnjSEpaa10XaWIbBQwpcS9xbUhhOAr8HnXmlBKnDNSyuNxjzAnBIdDRF3X8J6D5VnTNoyJgrci25UpAmpqIKokzwzQe8E9rcxOXCZlI9CjpcC35pxD9l1kBTHAU6XTQvRvssEZEKSytpOyVHLKqy2oQK21QkhQvHGvkDUxaMUAGSBVjvUCVT8AU8CPGQ2Yxf7waTNtKTo3NU0zTaOU0rl5P9E5KwTnnGYYLYbgge/h+yJRAZ0rXR3PKl6F3Fy2Rsp0hzEGfUCatSBLnogxrVbr4/GIsUreP6AxhpflUWmeeJZ5p5m5IIRw3itd1U1doIOipFa4CfiDMjsKuKy+ifEpDlh/uXhvpwlSgAG9xXK5YEwAqXDOXl9fgzRszCTlTBPA2BZ/EWXM+1jVlRBCKWmthRN42zaH42kuJrR+uH/AvQOEGzNBkc8bmqyu6x/+8IePj49IxqUmsM764HFES+j33gNyQUw4n8+Y6OJnxhjL6gzLG8jf//73vfcT2GDGzO2vscOlxyDz7u6OUno4HMp4Ej8QrQC+L9xiMBAGIAbGDcAMPFh8yAI54roh68DS6rg/sCzdjxkVNFfaroU4AlTzSPbEbWDqXlfee+edtZYL8fHjx3oW5Z4whQU3F+g9Uj7N5HsE3q7rACTsdrvdbodqO+WJ4GKxwFIFNngKPwitDNAdzvkf/N5Pvm16+OM/+zmCEXofXCScYJWla5HTEIOUQqKb0W0kDMaYdYYxZp1r2jqRVFdVIqmpakbpNFuIqJQi5UxXFRdi7gxSgiROijGFwDjH2LBt28fHR9AEET0ppYwRVG2oJfF5IAmFR4xrdnNzgzmHyC67+oVwCsIQsHsgniTOA3ZU0HTm3rCr6xvnAhjVlFIIoGqtITbetg1UmlEH4VFAzWKa4FCWUiIheOcsYhyg5xCClIoxjnuFYAqhGORjnS0QAEaXcQ5SL6h7GGYifBfW1jAM2OhBsWyz6B4+NiGk0JZKQMffhYIAP/nNmzf39/eMMRQsPO/0pkxeQOdBZyVR3jTN5XLebDbW9NfbzTiNSinrzXKxet7thBCUYv/Zr1brw+FIKfM+aK0Z49ChRB2A6zeOI+fwGkL5xTnnq9U6xqSUMsYhVmqt8ZAJoehdEFlARsAVAqgCHJn+U4tZgJ/4nfinS6HtWiFFP40++JCSdW7oLySl9+/fHQ57KWG6UjNGC4JamHvoObAvQinlXJ7Pl2kywRPnQkwR1CYAMkjhlNJaV1IIlCMxhKur7fl0VlJSmjjwO0pijJxRKfjQ93VVOTcH32kySiml9GazUVoREmGGyrlIiXAutK4YY23bpURCiHXdnE4nxjlhZHu1FpzhNg3j8Pz8fH19raUeL73gTHKhhETwwjCS5VVNvC/OeV1XP/7xj42xX748dN0ihCilYIwsuqV33kyTYLzW1dAPbd1czucUExoOVD9lNy1leT5QVLA4djod2665utr2/YUz4a1PkUomYojojRDuoZfc9z2Y8cjQgPjwIRGOMXdB8pbzTqhLKbVt473DrEhK+Db7MlwchgFxw3u3PxyFkozzy9Bb76qqElLtD4e6rj99+gRkArglY0woSejsmcoYG4dxuVi4yVDKoPMPjIFzaowJ0clKSS26rpVKLhcLzvg4DJEkF8PdqzvKmfc+pjiMI6E0pVBprZRaLjvO2X6/d86uVovT6eicY4wCdDVmGsehqtQf/v7/+m3Tw3/46c9SHmaWNhyDBPbCqYYQAtoJ1AswAwAtPaWkteKcCSnGcba7AT3cW8c5t7iQWhNKuRSMc11V5/6C/jSGaIyhhDjndFWVkvDh4QEIBkK5y9vLBRpWSldVjYiZY/o85oI3Ov4UODYo66CkBpQJ9YK1tlIaIRXQGcgVnPO6aRB3YlZGwykcx2EYes4ZpfMKK8Iu4mDKMj5oNUiWbo2zqDVWtWdCJ8k+M6WEZIxlOXHSZHvIkEUDMawmhJQCH9kO/wQKgRY75fln3/c3Nzfb7RbzT/JC6IJniSQ8Vcyib29vIQqCpJuySAMyBB5RCL5pakISz3aqhJAUQ4xRKhlCEFLUdfu0e16slmaaCKFIwGj5S9OJNFN2ngE/huCtdfim5eNhdIQ/uFwuUyIxxv3+gByAtwCBNjwH9BAu+3GibERWAJrEsvVTjDFRkggZxhEzKuTUuebwAUu55YK4F86g6P/A7RmGARxjzjlmmNNonXPGWsa50qokWpfdxGgiTV0LIcZhWK1WlxkEJ5QkzjkypOCcJMIZG8eRMcxRUIjAK4KuVuvlcgE9jMfHR2Adi8VyHKcQYhkjUUqV0sYYwiKhxJhJKskF55yDZ9lWNWd8HGZ/lPKo0WyVKQIe6TSZpml+93f/+T/+49fW2qapUkohRNQThBDvvFZaSamktFkFb71eA+JD3wDqEfC9mPWGpZRaK8bpZrO2xnEunHXeOsFhjuLKejPuF0b6eKdoAXnWKwUxDLBqKeBijIzRvr9gW4gQgtmVUgqyAviy2IteLBbL1QrEWbiLY8JfyDghBJT2WO9H/FlvNvf3969evULxt3t+3mw2h8MJyc+YSWt9PB5Wq1UkkbBoneVcrDfrFJKU6nI6S61GM4UQbm5vY4qUUWMNoXS1XI7DoPUcWLbb7X6/y10XIZmO4b3D5/x3f/T730F6oFlNGoUkz4qtaCdxrIGo1nXNucAN9Fm6hBBS1xXjs5YOpggIas5YKWWIESzglJLOw7e+7+fBL8YJzk3TpLLlslLq6ekJIAboTCnLf5IsMY+NBJr1G3y2+2CMfe973wNzHOAJcJK+73dzPTvLMc71I2fGWh9D3TZM8BhTiDGkiFWUXCbPhBxgI5j3Yk8YdwkA3csBCZ4bcgDyjZ/NG2iMEUcTLSFKJ0i44D/iZ5ahJX5yzJa2QNJwtWQ2p8MiQpHpJ9mDCPow+JxFRhuPkRACwApZDdMC7FXALQ5ZBNcM5TY+P4jFgJUKG6dtm91ut1qvrXU+hOPxdH1z/fXXX6vM4wwhbDab/X6PIRPueeF94eGsVqvz+WSzAS+udMi+Q/jrsB2KxyiyWRNj7P379zFGAMExRmhTl0ZhGAaoQtEsD4w/yzmHqBcGUbe3txDeoZQmgjXrVGW/aARHkuVA8Nlw2uu6JiRR+mv8bRygdkesNQjEJA+T8PxpRhK89wjrjLFxnFKa+aNTdvVAg4VTxLJy0ZxjKLu9vQUcDxmbwrRO2awNf2lVVcM4Qp2v6zq0MpfL5ebmZrfbMUKLPmbe82coYsYsWAlEC/DF4+PjD3/4wxD8/f0XQgiMlEspSbPFSNu26/X6/v4eng1lT15laUhrbfEVx39cr1fr9UpK1TTNp08fMW1OKXVdg8tSZbvQzWZTZfMG5ADMk/CIQLVClC9KsZxz6AgUmrWYpavCarXG5g3mEMAtKKVg6BVwAqUGybNVEGrwfMCTBoCx3++7rkPjMgyDEBJQWAj4SN3j4+N6u/bBc87vv9yP47hcLKfJKCkPxyNjrO/7zWpdV5XgvK5rM03H/X6xWHz58kVnqxi84vLYRWaKAxv4f/7Pf/Nt08Of/Pl/pnlruqAQBaTG+cAFKHjo5XJyzoLHieKY0llVsRCZ8Fkv52NVFnAAACAASURBVEtVVRLy2lozxqBjjnnjOAzOuaHvIWYbY4QRKS7S119/DZJWWd0qM72YrduRHkrgxoc8nU4fPnz4+PGjzWrYYEfE7PyDy1n2J/AHQwg44gCdgLMhyJZhb8yUx5B9F1AKIUa7LDiKz1mGrjFvEuHGAuWosnA/z2Z2WBJE1YPvG7LrL44pQkO5fvjrSkDk2ZgTgYBmcUdQBmA6iB9bfmDpYPA5QRfByAtku/I/pTy7Y3k9FQkbicc5NwzDatHtdjtCqRAykdT3Q900v/jFL9q6AVAGty+QyspXQ9eCshod/fl8wifEECtlRkp4oZVWkEDEdEIIHinyJaYskMTBL1Ah9SwkNbv12Wz+5bxz3iEa4lJg3E3S7Fv1+vVreAIjhBXFWZ8lV/2sXaqbpg0hOOdh29n3/eXSU0ats3jveK1AVoP3IFZgY+vq6go23THOrAqgHIhBL094aSixLw2UFYUdqMAgg5DsGkuzMFSl9TAOiSSsU2AIBJ7l2A9I/KA/oIbzeVcGIC0YYmUU4Zz9yU/++fPu6eHxXoq5pC1TPWA70KFjjH3zzTdVVWEAg5/JskcvwJkysBRCeue6tlutViHE4/EAFAPJAGJHWA9CDohZKQtPCTgSJqZIJFiWxsHDYB+KYTgApYBIiQCK9HlFNM5Kzx4suPm0ZCuqNBOXKVp2pHkcvLZtD4cDykfUTzBa7/tea3U8Hqdp7Loupng8Ha21jPJxGKfRUMaUkEJKyigv1h15ndvNvgYUvltCiOvr6/v7ewDpaLJJXjwax/Hf/99/9G3Tw3/46c9ili/FJApfG3mPZbkLhLkQQkw+xEAIsc4KKThjiUTgdCarSxacxBkbQiAggwueUsIqPwLT4b+39+7RUZXnHv+773PPJJOQhAAGyuGiBaweFbBFsKj1oIeKgIouOYB4A0R7YKm4REVppQKiHlAUpN7qpVZcRfTX6mkLIoqXloqwuAhikAAhydxn9n3//vjOu09+OlCkKOjv+fw1mUx29t7z7ve5P08y6TgO85jneR78Rbw0CfuaYRiJRAKFNngU8TTiGYNiha/Q4QnUeBhqa2vb2tp855jD8/19k8jmJRqSJDFRkCQxm8uFI2FRklLpNBMEVdNkqSSK/cIxxrv7Orz5Ha4FSbo2r5WzeQs5SA6Bh56wQfs6oJ8aFOQTx+BRRQYeztz3+/k3DaaS74KHfYAdCqF+hOgVPnjZ4ZP4fCcY1A2B19lh6w9xT1o0Gq2vr0d0ztejETSG5Avy4a/YIv1HUVMV0zR1w3AcNxQOZzM5JjDDMGxu4WGNIVYEgQ3Rgm8EeU2qqjqOjePD4hFFEbsMkhRh4MfjcYtPzcO3aRgG2nQrfGCZzIsYEM1CX1uVdwTAhgKNUlFV9PmCE1xVVcT9XNdlrofm5NFo9MCBA4iL+lokYoAWbytrmoYkyYIgpNMZ27aj0RheC5IInR2SDFuVZVmWaUKyQv+Nx+PwgxmGDtGIXca2bX9oJfYa3HyozPl8IZ/PQ6lCd1I/URs2DS4W/1QQBFmVs7lsoYBSCePgwYOmaXbv3t1z3GQy6WcVYk1C28ATh5WP0U8I/ARDWiQa6tKlob29ralpTyQSxYchnl3XLRQKbW1tNTU1GALhR8VQowoZ2alTJyjmSHGMx+OCIEQiYazzTp1qJFmUJalYLIiikEhUZzIZURSrqqr8fqiVlZXYB6AQQL2DixtG6oEDB3DH0DPcskxVVWtqamDGYfkpipLL5bF1wGiDzoEFhgeQleoxFTTdw1J0eF0k4g3YjjRNw8aNon3GmCyXkiMsywyHw4VCPpvNirKYqK5qaWmRRLlQKOSzOeYx0zBFAV0FtVwu79qO67iZdFpTtYpY1F972ENyuVzPnj23b9+OdoQY1eXvOcfAufTnNe8i3ujxJEiPz86G4sy4KY3rRAsKt5Ru6OKBxI3zffqMsVIgyPVc1/Xg6BAFy7K0gIYb6nlesr3d8zzmMdu2XccRRVGUJOQ4ubyiolS7yCW5wCeDQhigEZPSodWPwyfXM8YQBIMWDHXGD9iKPGldVVXXcyVe04TsGjh/BPZ/vTdwQI8PF/LtLair2Ozgw8WODKUVNxYbKz7PeFC3o9vE4QnvKPTD14/LxyeFDrmDEi9WgOnjVxjAZofu7PD6Bmzo2EahMTHuOYX4hxINmwZCV1VVuJuQt4cEMBwZBqLAh2xDSum6jte6Xqyqqszl86IoFYoFRVHzhUJVVVU2ncHDL0kSRkhBlGKzxjYHi16SJKhXWI3FYrG6uhr7XWtra2VlJSo3sfVjc/eNAKQ24rtAfjpcIthl8FQfPHgQRiHEiR/ziFVUWLYFtdePYKVSKUmUcpms7yKHag9lH/oBFpvv8HFdRxCQuecVi0XLtFVVdT3PMA1RKlmcvtfRcRyTW9umaVZVVUHV1TQtl8t6nodvobq6GpUocLRavFMsHj1JkrLZUn8nbKwQwzIfmyzyzED4eWzHkeTSO9iaDxw4gN4zXTo3wIqFZubvsEU+bh3rFp1ybNsxDL26Oh6OaPF4bNeuXbGKeFtbuqqq0p+NisChX4ffpUuX/fv3wxOg8noXJHFiAUNzMgwjGAyIopjJZLSApmpqv34/1DTl0093FAoFWVYTicSBAwf8Jon4RuDG8bepjn456JFYNliBCBnGYjE8ETofUaVpAehY0WgU9wfnKfPNAdYqdEqH9/aAUxenUVlZCZ82Yywej4dCIdQD5HI5USzVVFqWGQqFGPMsy7IcS5TEqqqqZHvKcz2BibpuMM8zDENRlVwuJ0uS3xwed8Y30VBVVywWi8Uissja2tpMPiIpEomoqjrwjAH/qnj441trLT43DTEGyEm4knydFyvb8zzPsx3HRgGnqiqyLHme6yei+Dtg6QiiZNs25txKspTNZmVFhpUqCEI6lbZtW0EXaNR8BQJQmlRVRQI+cqixS0Iwyh3qiZA7K/MRMRKv7EWcSpZl9P/wN2iFj+Ky+aQnz/PyelGUREVVbcfRAgGPMcM0VEVVSxa01dEbw3iXGIQrEQJV+JQIKE2+SuLLFZHPDvOdAzavqIJoQUMLmG5I8/XPU+DALyxwV7vJGwFZloXaQGw9sLTwrfleGuScYOkgaoKdvbq6Gjs+bhfKITFWJZfLxWIxJLwi8oQdHMq4x/NWfdOzkM+JopRMpSRJth07ny/kC/lwOIxMcDzPJp9qguXEM1Uy0CokSWptbZUkEafUURJ4ngfXE8LvkGHpdBpnAr8NrAp/DaMyS+U9PGpraw8ePGjxESCO44RCIfRu6dK1a1t7Gxa5yNtBG4ZhmIbIBLSasHn1uMczuCBlsea5FSvoOrIwBV3XMbjbdT1BEiRZ8q8X1S2O4yh8Dmg+n4dCWl9f39LSgvrnRCLhhw1SqRSWum3bRT4sj9uRMka5iaKYSCTcDunOAP5Y6JumZXnMjVVUMD45AzumoiiyWMr6w/lYPO/Zj3nYto0Eh1wu59heUS8y5oTCmsecpqYmvWgpSkDXS5XtNs9sxtaMesNIJFLsMDFQ4knq2JQZYygRF4TSILZCPu+6Tk1NdVVVPBBQt23bbho2Oq0dPHgQ3bl1XUdUCU8EHg1YWujtBkUNSw77OGOlAa7wqiHLBok2WIfY8aFE+nsjTglJgwJPhMHaQ00PZGpNTQ1m3vmaYjqdrq2tte1SlydkLgUCmiRJtmO7nhONRqORmGVZqqzmcjlJEAVByGYyiiyLTFBkWZWVaDjiOk46nRYl0ReB2NxwPsjjam9v921HURTP+vf+/6p4+H/eXON1aMMg8kRyBAM7WlWO43jMReY+452CYG0wJvq5QNglsSNYpsUYQw2hgmCUIgeDQaw/vVgsFAqiIHqeVwrTBQJ+aAgDpPyuig6fLSzzTGdZliVZMo3SyGjHdUN8bDcejABvCu97umVeN4sNV5IkRVX8eDhjzNcudcOQBBG17ALPVIE4wX3zXdi4UTofiivz1D0EqbC1YfPFo+7HnH17yB+lwHgZlJ+qiwdJ4ONEPF6uJfDWEYwx5PbhEvyABI4DGY/Txg6CihY//COXSrc86EpQ/aqqqizLgiDBnYQk9oNPEBIOj9jjDjt2aX6ZIEqFYlGWFVXTXNcNaBrcCFgqyBHwXbe4WMaDHPC8+5cAV4avByD8KPFyENj+qKKC80QUxerq6iIfVynyID9uFLLFcRsVXtWYTCaj0ZggMH84kizLCOfYloVOy5COfoYYlAz/y0W8GrtQPl8QRTEer9R13bYx89nxmMcEhi4DNq9iUVUVoxRMw3Q9F//dsqzu3bvv29eMMB5a2oVCITRcYbyZmK+IsFLPRAsrB4XfuJ++FoKNGEpDOBx2PUfTNDwIkLJQxSRBRFwBO6a/PDwOhJMkScFgsFDUbdvM5VJdutY1NNRXJzpt/PumynjC80p6GEQXZKHruuj6gGSEYrGIVnQo+PJ4YQ0SkHRdz2QyGH+m60YqlWxvb+3cUNepU3U6lUsl0xDSKJOKxWLIdICwx1Mj8VGacG0hDudrvaIoonMBPqzxZz+bzWGwEmOsWCyixhv+W3jMJN7/Di8QKBJ5Ni2OA+ekwDvKwNjChhMMhlzXDQaDuXzONE1NUxVFsRxLURXTNMOhCGPMc9xQMJTP5X1/tceYpqr4L7FYrKgXTctEzjFMbZhBbW1tuEbHcfAUIKR02oC+/6p4+POad30to/RrSRIEwfOEUChcLOrBYEgQREEQLQs+KKbIiiQrcJw4jhcORWzLUtSSiwl30HEcTdNMywxHwq7nup6rqorj2AFNk0SxmCtYhikKgm2VYsu247je/4269ANKWMd+sodfRi9JkhbQdNP0mOcyxgQmiGIgGFBUJZVOu54bCAYEUQhHI7quO57DBEFgpbontcNgGZeXi8ui6DpuMBCMRiL5XF7PF0WROY6dy+WCwZDnMVGUFEXFBlcsFjt37owdCp5HpzRXp1SNjI4UEm+EovCeSB4fN49LYIxBHMJxhJbxwWDw888/x8JCKExVVU8SbNcJhkK2a0uKLMgSEwSJq9WYUOjfPUkq1ZEx5qmq4rqOZZmWZSqK7Di2JImKogYCgVgsXizqsqwwJkiSHI3GCoWiKEqRSDQQCDImWKYbDATx7cuyEghotm1JsqRoKhOEQrGI6IplGKZuFApFWVYCoXAymQyGgqqmSLKoaaplWNgWYTWikrFYLAaDmqoqul5UFNm2LV0vVlTEisUCfHS+do+vDNMmJKlUBmyatm07+Xwhl8si1In6IOSkYkXBLkFWmM0bOBuGgc9YlgUHQiaTMW0rUd2ppeWgwMRUKl1REfdclklnRUH0XBfiAT5PHoUSJEmxbQdJ3pIko7xAVQPFop7L5XmJeFJRZc9zdaPoCa6iyoGAZlqGbVtVVZUVsThzBM9j6AdsmbZhmOFwNJVKMcGDCl9dXQ3XBLYziTcIgb3ox2MZ89CCyTQNTVNRouF5Lm4seterqmKahm1bqqpZhhEJR7KZjOd4kUgkHquw9NIMvvb29traWpGXCra1tXk8J0VRFCQRMMZCIS1fyOq6LonKSd16yLJSUVHxedPuRFU1HD5eh/pW7CdtbW2MMaTnYWNFSjouCg1roZnhWkKlFsviwdZWw7C7de3etUvX/fv353JZNFzBelZVJZvNofkgxAZ62CFqWCgUmpubYVRhA41Go7KshELhdDrjup4oSrKsoFgklc15TAiGwwVdz+ULoUgkncl6PGEJaV0KH6XsOA6WFuowkEvtOI4a1NB4ynYcSRRTqVSPkxrb2ttcwZM1ORQJqQHNtC1ZVQOhUDgUsgw7Foo5thOviOfyWce1q6orZVUWZFFSJcd1mMg8kQXDQSYyP6SEUKXruuhsjUQDwzD8+XoQRWee3u9fFQ9/WfseY8wPpdq2jdp6wzDhiMezihshljI1BddxHbuUjA8FDeLB4+mnHm+R76vJjHcfKxQKjl0KYNi8xooxBgUTxzQMAyob3L5wWGOpiXx2myTLRcMQRBGjRh3HCUcisqIYpmlalqqpTGCxigrdMBzH8ZgnCSLi3i7vTW3zvj2O7SiKwjzPMk3XcTzX1fWCFlAlSfbjjQjHIR6r8PHLjGup0Gr9MAlMAahvMm9b7QcwcJdsXvEABdl3CyC9x+Wpk6ZpasFAKBwSRREubJg+tmO7tus6pflc0KC5O8vzrSWFd3W2bVtWJMd2HMdxXS8SidTUlHo8MG6NIYUGbqhCociY4DiuLCuiKGCLQWGjYVmaquVyuYAWYIwZRR1X19bWFoqEc7lcIBgQRdGyrUAgoMolVw/iK4yVGnKgYyhELBx0hmFUVVUhcg45inTyUChk26YgCnj+uRqDJDEdTwIWBvJ8JD6TPBKJtLe3w0fvui6CzHAsuDzB1zTNfL6QSCRCoVBrayvK4KGFGYZx7rChO3bsgOYLV7Ioivl8QeowNQh5box3E4KXD6U2bW1toXBYNwp53rUN8WTGWKyiwnM8xEux+xQKBdd1gsGALJf66BUKhW7dujU1NXXt2hUGhKZpsFR8tQM/4r/D9RGPx2HLwkaBIYLVruu6LEmqonquGwwG21pbPdetqqpyXRfSCHqof5Lw0mC9MZ56J0mSZRuGURQEoVAoRKOxaDRWLBZd13FdLxqNMcYQY8CHGY/oQmF3HKdr167Nzc0QDFhv8DshtJ7L5XS9iCxt27Zdx00lU8FAsL6+HvXh6I0B0zMSiZimhZIdxlg6ncaFIG8VmyZe4ytDNhHMSoi9UkhPkTUtgGWsKIppmYVCQQtouq6jmhnlUzCnEJVBvAe2O4o2YrGY7dgeY4qsYD8xDMPUDVXTMvmsn7wHTRHLkjkMtTeBoBaPV5imIYqCrCiBUEiSZUVVJFkWJUmURFVTmccCmoYQEbrc45JzuRzi8zB0PM9Dg+1jMC3uf/+63uXpgwLPbJNlOZPJovQcmx1K2CSpFJbwlwt0Z4zN8o1Qv2segj8Oz9PHY2+apmGYML7woGp8bIPGm/LDiMbw7mQyKfBsDZXPuxcEQRBEy7IkURKYIAoC87xYNBbQtHwubxqmpmqSKCWqqrKZrKHrsiShOF7jHXX8SCyEv8un1NmlOljdcR1ZVhzHYUwQRcx1F7DmVN441+TDSn2bDl57xmMh8GsZvItnPp+HTxP3AeejKAoy/Px2HVDBkCPkeR4mLHkdmk+U9N98HoF9bFUwtjzPE0XBT6jA3oHYnSRLzMMQQKeioqKqKoFNEIENPBgI2ITDYdu2LctRFFlRFdu2EBlTFFmURMd1kd+J6fYiHyCRTqfVAMbMqZqmFbF/KZrN24cgLwuZl5IkSh2APQcg2PAwY/dUFFkURVEQdV1PpdKmaQYCQUmSPM/FHiGVkk9ySHmCxYBn2Pu/9h6Ow7MGRF7qbFlWPpfTtEB1orp5b3NA00zDlESRecw0DdctpeRiI0Y0m7FSSRRMCqhT0OUVnuLseR46/cGwxjuapiEXK51OR6ORQrZUktLe3q7xTouCyGS5NPMAK6qurm7Pnj1YpR33fT8u5YffDMOIRqNI9MIirKiogO8IdwABJ/gM8a1BHsAZ4jvWcDSLtz33HXqqqnInGKIRjmGYruuFw5FkMpVKpZPJdCJRDSvN92sJvKk+hBBjDJGVdDqNdm0oVvA8D1lJgUDAtq1sNltTUwOzmDFWKBQQdjL5KJoO+cRBP/wT4iPQYZogoai9vR2SCd8alFR0vsP5hEIhQzccx1UVVRQEURCikYheLLqO69iWH8vxPM/vzYOxDYZhNDY2IscUVouiKK7nKrLsuq6CG14saoGAYRqGbmiqViwUJVGqiFWIgijLcrFQdEqdulkoHBJFCZ7sUDAsiaKM6ciBgCxJzPMCmgbvLnYeZBvjOyoUCuhwhVsB390xSGyFcwnPCbYtuAh1vQhpBlvVcWxU60g870Lg3QuCwaAglpLosRrwOOHhwfcBYYCwu2EYnusZfCqcy/PfRVFE6g72XHwHjDE4amBnQGB6pZESpYQzj49ewG6I/H2swl69erW3t8OG1VQNWkMgEEAbcCiSAT7S1uPjpWRZFgSG6UCYthYOh3k/gyBOw+864Kd4u66Lr03lrYpkWYY9bvGSb0gLPH5ohgNnCASzX+EBKYW7J0mSbhiqpuL2+nfMsqxcNivyDskIKmLL8LxSmz9fPRdL7bJFSZQgHgKBQDxe6fEJEwov8lD42D5Zlg3D8jzEaZjnuYVCPhIJY/SYwFiAj3Ly3JJh5LpuvlCIx+NMEDAI2rZtkYm+LontAJYl8nMkSUqn036KERY9vmLcIjzD8AKiG2BbW7uulxq+KorCmMddaiWtX+DtCGEYwTWMtYftACapLMtw3FmWZdsOHu8in60tiqJt2V98sef888/fvn07ohqQnX7VBdQXiDeNj66EkwEngJwTURQkuTTvGvMsBUHI5XJ1nWqTyfZoLCIrUrFYkGRRkjEn0cQXB2MUKqo/psnPd8CGJfJBTzDRIAvRmwtmLvQhxHuwFeJhxPmgPBgVr7A8cItkWfb7Fzm8LzquNJ/Py7IEk903xZD+wJiYyWTi8TjixnhO/a8GGh6cqNAIw+FwIpGA6gnfrGVZsiz5VpHCW6kjxovBDxAMjLd9Qxk5IklIooNGiy8CKRX4TvGgIWcP2xH2Vs/zotGIy1zTNCRJlGTRti1FkYt6IRQKInsVDiVfzhmGUVFRkc1mTdOsr6/fu3evIAihcKi1vd127NIT4TFJklRZYcyL8NI/LELXdVEwKEtSoVAQBGbZlqapoVAwGAzohlEolEoR/c3EdV3GvGgkitwEeKSxzeICoe4gqAYr7RiUxa1Z9z6eH9YhOi1JUqGYD4WCrue6ruN5rmVbrueKksi8Ui9rbGQi79SIXqS4DFh/SMLN5XK9evXat2+fyzvvy7KMicqMNxmFWxCX6rtl/P0UUhqalN9WBRu9KIiiIBq6YVu2IiuyJAU0LZPOSKKkF4uhYLC2U63rOO1tbVKHDmiwb3C2cDf5KT1Y8ZFIBPNKVVUzTcN1vXA47DhuoVAIh0OMlcpoE4kEGgvjorAWbT7UzD8UnlKBz6jCE4WnF48f4+U2/tYPO8a3dURJMkz0cXL8rdN1XVmUYHAIgtDQ0IAsFFmWMWcCWZvIR8JTWigWImF0g5FSqZTjuAju4d5CtvmakaZplmmLooDiR0kS8/mc5zFJlizbzuXyAU3LZrKmaYqCgFVhWZYnsFAopGpqKBRSIHSdUp0gUsVhF9bW1hYKeUh0xpN0Ld4a3XdIwmlgGIaiyLIiqaqGzhCCIBSLusQbKGHHR6jTsiysTD/wznjXQkR0sS/AdOOKm+B5LJfNFotF5nmSJJXSrEUhk0k3NDQwxuB0hd6KFo2macLlrfCBd3CW+ra1L4QEURB5SBn2WTweT6VSlRWVHu+iCCle8ppKIjZ9fN2IOki8CB8aInwUfpzPdd0uXbpks1l0D8xkMp07d25tbfV4q02ILt8iwX9kPKcOmx1UbwS3edZsFikJfnYGd5C6SM/XVC2dzsRiFeFwRFVU07RQt19RUQHXPHZwmTcdgLyHocCjSkH4kVw+U8B1XVkuxfOQ1w5PrOu6yP31PA+rQiqNeCrJEkmSEN+GqIOi09bW1rNnTz+XHXFNhJ14HCWEfx0JRzBzSVNV+Bs0VcU5tLa2irxIzc9F9sMPeGQOHDggCEI0Fk2l04V8gTFmmZYoimgr4rhuZTzuOo7ruJ7rZtKZQr5QyOc7d+6M2+7YtmHonscCAU2RlWwmZxpGKBiURBHublmSDN0QBAF5t/BhwAg2eY2a71lBRP38n/7kXxUP77z3N+g+Ep92Aq3fsktpkQofrOa6rud6iJ06vEwGBxFFQTd0NE1EKq7vCDJNE/2QsbLV0ggBOJE9uMsdPnIOm0XJt873Wb+6nTHmJwbgt0KHFkmCIPhWre9BsiwLGZDwyjHGFEWBMIAXAg9esDTGUpD4HE1VVSRJRD8oUUQXGg+7DISZn73gj1iAEQZLAttHIBBIJBJoViHxUVa+C87lLUxs20YPffzK4mWZEBuSJKma5nILiXWY6iN4gsBHISGBB/upZZmMd3OCVotdxnbsutq6ZDLJGDqRlYw5PKV+gAue9Hwu7zgevDowHxVFMU1TVmTHdZnnMY+Joghj2U+GSaZTwWCwsqrS8zxZkXVdh4cdX5DL2/8piqJpKhRM/Ds/Mmnx5gqCINTU1PB2/Mg5QZspN5vN81vkCYKABFzYCnC1wcfi8Mp/P+/Lz5THUvfvtqqWdH+HZySbpuk4bjQa+eKLL84777x9+/bhLmmaZpo242MqEokESlgTiQSWruv+X2d/7L+6aWAz8jwPfg9oAO2trZFI+ODBFrT7dV1HlqVIJIy0GdwKiWc0+KIOVwQ7zOJDfC3LQr993/Eiy3IkEmlqakK4UuYV5rCu4BNDkwysTD+OAvUF6hc+IPLphoyPEBYEkTFBlpVAICgIYj5fcBw3nc7AvYNqSrQyhW2NEkjY3JIkoaUHFpvEq1WQ8GpZlmka0WjE4x3JMDsSyjvj8wVkWcaXLgiCKEq+7lWytnUdRlIwGDx48CBCONC3oIxC/cfVmaZZWVkpSZIsi4GA1tp60LLMUCiYz+dUVVFV1XU9JD7gcnjMXICvD348+MQOtBwIhUKKqiDCp8hyKpUSmaCqatHQkQMNn4HjONgSIcYCAU3Xi4LADENnzAsGg5Io+56JAG+JLfEKUDRzhRaOR9tPhPG9O4FA4Bi05HvrL+8gXINrxpMsyZJhFFDWgLCk49iY6xIMhkSeIoYn0HVdLaD6mzIKYSRJ8ivRsZtDy8aTz7zSrDSYSMi89lVs4G+OkDH4Q6TKYekEtIDoCXqh6Fi24DFREH7Qo4dpGLZpiUzQi0VJEJnndaqpSbUni/mCFgggCNxxq4J7BFqJVWo3pkej0UgkzJgnMAhINRwOlAe9ZwAAHitJREFUIxbKmIfnE3tQsViELwsHhCiVeJm+oijV1dXIvYOcF3jaFSS/y1vM+5sX9h0o0YgVu64rK4rIqyhww+Gnhk6BSBRqjOEaDgQ00zThRBZ5GzjHcVRV6dSp1nEcy0JGgIg4k8W7WLt8aqZSKnYRXdfRAprj2JZlKYpsmoaqabbjGLrBGJNESdd11A+WMnrzOUEQtECAMVYRrzBN09RNuCIRvsN+19raylipWYvFS8Tx7GFTYHwqCy82lFVNRaoJbkIoFBZFMRgMYGdE5BMJ7J7noQAC7wRKk0EdVVWTySRUWugWoVColALkCa7tiEyQJUmVFc91HdtmzHM9FysTs6l9pQSSFaoSSm1aW1vRhQ0GBJIji8ViMBQq6gVBFPzgBOrsVFUt5PIQIdjm4HaDdxdrHvIGehUeUsYb3kBTFnijZdgQ/fv3b2lpYbzvS3V1NborYrOAYo4bhTADvm6ICiwSmDtwG0Kk+d+F2KEHgc27vKBWw286hE9CA3AcJ8SHszq8KCcWi+F9fN2apqVSKRj0sP+KxaKuG5hIgckQsPMQwcZuAI9CZWUlqo6SyRSeu0KhYHUoOYIvXpblPXv2yLJcXV2Ngh6ZDwLCnUHtSyQSMYoFTVU9183ncpZpaqpayOej0VgoVBoFAVvTtm2MJMDdwPLzAy35YiEUDmmqahhGNpMRRRHtgmRJMg3DtixNVWGgMM9jnmfxTnFIKZRl2bLMQr4gCZKmqrZlmYbhuW4sGlVk2eYJ+kjigM/WDxRB/YXLAf7DoT85FrOmscphu3F3HjPNUiI/VJ5SPFZRREFSeO0JtkJFURRFtviEH56dZmN3cxwH/lwoL/gM4yN8/WiEyCuiYaH7ixLbFpJc8X24fKy8JJYCrS6veuvcubMfB0Z9Y4BPQGOMMX58mVcX+zcB9h0eD4wmZ8wTRAFzUQRBDIXClmWbpsm7Oou4Il9jdXkrNMYYjGv8CraL7/SE6cPT5EXsy0KHchNsskhdh7fNtm2Jrwx/a2CMmaaJ8fSCIMBUr6urgwsyECi5Cv1vFucWCAYUWYnFYgcPtjqOk88XoND5ybW+ug0jzzRt13UCWkm1MU1TliXdMLRAoFDIu65rm5ZlWa5davEbjUZlRXEcx0OyiqpYlmUZpf5R+DahVei6XijkIQz8WHQ2m/Xj+f5pYKvSNFVWZKSQCTxwjUQdXCDGPeEPYU5hFLbF+xyIvNoZEhQuDti7mUxGEiTcdrnDuCTLslRNDQQCu3fv9pudyLIcDkcURYUEtSyrvr5+37598Xi8paWlqqrK4XX1EDyu63rMtezS8AkkquL2OpaZzWaQ3eu6TigUFAQmyxKsVaj5vr8Xt4LxFgaQbQrvF4t1ghCUP1jJNE10GHR5EQMkEExzVVXh0iytjUAAmoH/AuWHhUIB/0LlHUq4aVWKosGOQYshkddG2LytFuwYODnxT/ExX1vCC6wfVnIzMpQxQn1WeZdclCt2NK/hkY5GY+3t7VAIJN43EE+WwhPK9+7di5ImzPnBIoQr2LbtYrEoioIkCqlUEkJX14uGoTPGkslUOByBce/yzDTkkji8SwoUStu2YxUVuVzWsq1IOMIYs9AGXxAlnr7lm6eO48D/Y1gmTgx5yRg2bFt2saD7vgQo1n5CgcjHZ6GtFnINbNsu8IHqSPjM5/M/HXo0s6YF3ylEEARBED7i8T4BgiAI4kSExANBEARRBhIPBEEQRBlIPBAEQRBlIPFAEARBlIHEA0EQBFEGEg8EQRBEGUg8EARBEGU4nHjYvHkzira/tbMBhmFMmTKlpqbm5JNPfvnll/33lyxZ0qNHj4aGhtmzZ/tNh95+++3TTjutqqrq8ssvT6VSeHPv3r0XXnhhPB7/8Y9//Mknn3Q8Qt++fROJxK233nr4w/r07dv3nXfe+aYulSAI4oTFOzTYWP3RmN8a99xzzxlnnLF79+5Vq1ZpmrZz507P89asWROPx999990tW7Y0NjauWLHC87xsNltRUbFs2bK9e/dedNFF48ePxxGGDRt27bXX7tu37+677+7evTs6bfz2t7/t0qXL3/72t3fffbempua555471GE7snLlygMHDnx7F08QBHFicCKKh4aGhtWrV+P16NGj77jjDs/zrrzyypkzZ+LN//mf/zn77LM9z3viiSfOOOMM/2wxMWb79u2KoqTTac/zXNft1KnTm2++6XnekCFD/N1/+fLljzzyyKEOSxAEQRxp7GHVqlUDBgwIhUKNjY2LFy9mjKXT6Xg8PmfOnOrq6k6dOk2ZMgUNsNLp9IQJE+rq6qLR6Pnnn9/c3MwYe/HFF0ePHj1u3Lh4PF5XV/fEE08c6h+l0+m9e/cOHFgafTdw4MAtW7YwxjZv3nz4N0855RRZlj/77LMtW7b06dMHPaoEQTjzzDO3bNmi6/rbb7993nnn4cMTJ06cOnXqoQ7bETiXDnWxpmneeOON1dXVp5566ty5c4cNO5qRTARBECcgRyQestnsmDFjbr755j179ixYsOCmm25Kp9OMsXQ6/b//+78fffTRmjVr1q1bd8899zDG5s2bt3Xr1nXr1n3yySee5y1YsAAHWbly5VlnnXXgwIE777xz6tSpfJL7l9m/f78gCPF4HD8mEokDBw7gfbTRx5vJZNI0zY5vMsaqqqoOHDjwpTdxBDQ3fumllxobG2tqaqZMmYIJumUPW/bEyl7snDlztm/f/vHHHz/++OOPPPLIkdxMgiCI7wRHJB4CgcDHH388YcIE6P6CIPhB4IULF5500kl9+/adN2/eM888wxibNm3a6tWre/bsGQ6Hq6urSw20GauoqJgyZYqmadddd51pmhjh+VX8HsX+O+io/NWgMVpnH+GbqVTK87z169dv2LDh/fffX7t27Zw5cw512EPdh69e7G9+85t77723c+fOZ5555uTJkw/1hwRBEN85yoiHp556Ci883utbkqSVK1f27t371FNPfeSRR9DZHPTs2RMvevfu3dzc7DhOKpUaN25c9+7dR40a9fnnn/uf7N+/P0YgdJx19VVqa2td1/XFTzKZrKurw/u+pEkmk7FYLBQKdXzT/3DZN6uqqhhjCxcurK2t7d69++TJk1etWnWowx7q3L50saZpNjc3n3TSSXizsbHxMNdFEATx3aKMeJg0aRKGYmJunyzLq1ateuihh958881Nmza98MILHff3nTt3+i9qa2slSRo3btygQYO2b9++du3akSNH+p/saBAchqqqqq5du3744Yf48YMPPujfvz9jbMCAAYd/c/v27ZZl9ezZc8CAAVu3bvXnen744Yf9+/evra2NRqO+46iqqgoDlcoe9lB86WJVVa2pqWlqasKbe/bsOZILJAiC+E5QRjyoqrpo0aK2trbFixcPGjRIkqRkMqlpGmYl3nfffcViESPMGGMzZsxobm7euXPnrbfeetVVVzHGkskkBip9/PHHTz31lP/JI2fixIlz587NZDIbNmx4/fXXr776asbYpEmTli1btmPHjgMHDixatOiaa65hjI0ZM2bLli2vvvpqsVicM2fOZZddFo1Gf/CDHwwaNGjOnDmGYSxevDgQCAwbNkxRlEmTJk2fPn3//v3btm27//77L7/88kMdtlAovPjii21tbV86sa9e7IQJE+6+++6DBw9u3LhxxYoVRygCCYIgvgN8NZlp5cqVvXr1CofDw4YN27Fjh+d5uq6PHTs2Fov17t17yZIlV199dWNjI/w/ixcvbmxsrK6uvv766zEFcPXq1T169EgkEv/5n//5yiuvVFdXL1++/IUXXhg6dKj/Lxhjra2th8qmsm37xhtvrKurO/nkk1999VX//UcffbRHjx5du3a9++67/Tffeeed0047rbq6+vLLL8foRM/z9u/f/7Of/ayqqurss8/etm0b3tR1ffLkyTU1NV27dr3rrrswErbsYXfv3s0Ye++99zzP69Onz7p16w51sThmPB4fMmTInDlzLr300q+bOkYQBHFicvTDRJHriemmx0xYnXj07dt32bJlP/zhD8te7MqVK/v164eYxOzZs1Op1MMPP3yczpQgCOJYcrgo8TfKxo0b/ZzXjowaNeqSSy759s+nLK2trc3NzdFo9FAfeOutt5YtW7Zs2bLW1tbf/OY3zz777Ld5egRBEN8cRy8eNE2bMWOGJElH9+ennnoqckNPZE4//fTLLrusf//+uq6Xvdi5c+dOnz590KBBNTU1t99++5AhQ47LeRIEQRxzjt65RBAEQXyPoYbeBEEQRBlIPBAEQRBlIPFAEARBlIHEA0EQBFEGEg8EQRBEGUg8EARBEGUg8UAQBEGUgcQDQRAEUQYSDwRBEEQZSDwQBEEQZSDxQBAEQZSBxANBEARRBhIPBEEQRBlIPBAEQRBlIPFAEARBlIHEA0EQBFEGEg8EQRBEGUg8EARBEGUg8UAQBEGUgcQDQRAEUQYSDwRBEEQZSDwQBEEQZSDxQBAEQZSBxANBEARRBhIPBEEQRBlIPBAEQRBlIPFAEARBlIHEA0EQBFEGEg8EQRBEGUg8EARBEGUg8UAQBEGUgcQDQRAEUQYSDwRBEEQZSDwQBEEQZSDxQBAEQZSBxANBEARRBhIPBEEQRBlIPBAEQRBlIPFAEARBlIHEA0EQBFEGEg8EQRBEGUg8EARBEGUg8UAQBEGUgcQDQRAEUQYSDwRBEEQZSDwQBEEQZSDxQBAEQZSBxANBEARRBhIPBEEQRBlIPBAEQRBlIPFAEARBlIHEA0EQBFEGEg8EQRBEGUg8EARBEGUg8UAQBEGUgcQDQRAEUQYSDwRBEEQZSDwQBEEQZSDxQBAEQZSBxANBEARRBhIPBEEQRBlIPBAEQRBlkL/6VjKZ/PbP4/hSWVl5vE+BIIgyfLe2o447yQl45l93oyPrgSAIgigDiQeCIAiiDCQeCIIgiDKQeCAIgiDKQOKBIAiCKEOZzCWC+BItLS2ffvrpvn37MpkMYywWi9XX1/fs2bNTp07H+9QIgvimOPbWQyaTObcDY8aMeeyxxxzHOeoDXnrppTt37jyGZ/gt89FHHwmC8K/cgaPgqaeeisfjqVQKP+bzeeS0JZPJMWPGVFZWDhkyZMeOHUdyqPXr169duzYajQ4fPnzChAkTJkwYPnx4NBpdu3bt+vXrv8FrIE4MlixZ0qNHj4aGhtmzZ7uue7xP5/+D67rnnnvuwoULO755yy23vP7668frlI6cnTt33nbbbT//+c/HjBkze/bsXbt2He8z+jLflHPp0UcfffXVV3//+9/fcsstK1euXLVq1VEfatq0aaSlHgXpdHrBggVfevOaa65JJBLvvvvuJZdcctFFF9m2ffiD/PGPf2SMjR49ul+/fhUVFXizoqKiX79+o0eP9j9AfF9Zu3btHXfc8dvf/vatt9565plnnn766eN9RmVYvXr1tm3bjvrP9+3bd8MNNxzD8zkS8vn8zJkzu3fv/vDDDz/xxBP9+/efMWPG/v37v+XTODzflHiIRqOxWKyysnLw4MEDBw7829/+1vG3pmke+aGGDh0ajUaP9Ql+/7nggguWLl3a1tbmv5NKpdatW/fII4/06dPnlltuqa2tfeeddw5zhPXr10ej0cGDBx/qA4MHD45Go2RDfI95/PHHJ0+ePHDgwL59+86YMWPZsmXH+4zKMHz48AcffPCoLRvbtr/9fXnr1q2O41x33XXdunWLx+OjR4/u0aPHifYo/aviwfM8+KMPg+9amT9//sqVK++444558+YxxtavX3/NNdeMHDnyrrvu+uKLLyzLGjly5MaNG/FXa9euHTdunOd5Y8eOhXNp06ZNN95444gRIyZPnvz+++/jY5deeunu3bvx+uGHH37yyScZY7t377755psvvPDCyy677F8xXI45ffv2vfPOO2tqaqqqqmbNmvXKK6/07t37pJNO+tWvfoUPrFq1asCAAaFQqLGxcfHixXjz7bffHjBgQHV19dSpUwcPHrxmzRrGWDqdnjBhQk1NzYABA+bPn//V/9W1a9dx48Y98MAD/jtNTU3du3dXFMU/mc8///xQp9rS0tLc3HwY2QAGDx7c3Nzc0tLydW4D8Z1h8+bNAwcOxOuBAwdu2bLl+J5PWcaNG5fJZF577bUvve+67nPPPTdu3LixY8cuXLjQMIz9+/ePHDkSO8aWLVtGjx69d+/e6dOnZzKZsWPHGobh/202m/1GPWldunTJZrOrVq3yLfj58+ePGjWKMfb555/PmDFjxIgRP//5zxcuXOg4zosvvnjzzTf7fztx4sQ//vGPX726Y36SRy8ecrncyy+/PH78+Pfee++rvy0UCvl8PpvNvv/++++9996pp56K93/3u9+NGTNm1qxZu3btuvfee6+//vpnn322T58+s2bNsm377LPPfvvtt/HJv/zlLxdccIEgCPgxmUzedtttF1544QsvvHDllVfOnj173759hzq3Bx98sFevXs8///yUKVMWLVqUTqeP+jKPOWvXrt25c+eLL774q1/96rHHHvvoo48efPDBWbNmtbS0ZLPZMWPG3HzzzXv27FmwYMFNN92UTqfb29tHjhw5e/bsrVu3SpL07rvv4jjXXHNNNpvdtGnTc8899+STTy5duvSr/+v2229/8skn/b27vr5+9+7dfhTks88+a2hoONR5fvrpp7179z6SK+rdu/enn3769e4C8R1h//79iUQCrxOJRDKZ/Fqm/7dDIBCYNm3a8uXL/WAb+MMf/vCnP/3pl7/85WOPPZbJZH7961/X1dVNmjTpgQceME3zgQceuOWWWxoaGh566KFYLPbSSy9pmub/7SeffHLVVVc999xzXzrmsaK2tvbGG298/PHHR48ePWfOnDfeeMO/sfPnz+/SpcszzzyzYMGCdevWbdiwYdiwYZ988gm6dOzevXv//v1Dhgz56tUd85M8msylXbt2vfrqq++///6wYcPuv//+zp07f/Uz1157LV5EIpHzzz8fUpEx9qMf/Qii4o033hg6dOi///u/M8auuOKK119/fePGjT/96U/nzZs3depU0zQ3bNjgH4Qxtm7dup49e1588cWMsaFDh7711lt//vOfr7zyyrJnaFlWOp3WdX3IkCFvvPGGry+fCEyfPj0Wi5133nm1tbU33XRTJBIZNWpUJBJJJpOVlZUff/xxr169HMepq6sTBCGVSr355ptnnHHGpZdeyhiDRGGMJZPJV155Zfv27XV1dXV1dbNmzVq0aNF11133pf9VW1s7ceLE+++//95772WM1dTU9O3b96677po5c+bq1au3bt364x//+FDnuW/fvlNOOeVIrqhbt25vvvnmn//856O/KcQhOPfcc4/vCXxVg7ZtW1XV43Iyh2HQoEE//OEPly5deuutt/pvrl69+oorrmhsbGSMTZky5fLLL//v//7viy+++K9//evUqVN79+599tlnH+aAPXr0+MMf/nDttdeedtppl1xyyRFqS0fO6NGjR4wYsXHjxg8++GDp0qXPPffc3Xff3bNnz9tvv726ulpVVcMwotFoNpvt1KlT3759169fP2LEiDVr1pxzzjnBYLDs1YVCoWN4hl9bPNx2220tLS1XXHHF1KlTD7NKnn322bJio7a2Fi+am5s3bdo0duxY/1e6rp911lmWZW3dunX//v29evWqr6/3f9va2trxgA0NDQcPHjzUf7/jjjuefvrp66+/PhQKnXfeeePHj/etkOOOr4spiuJfoCzLjDFJklauXPnkk0+qqnrKKaeIosgYa2pqOumkk/CxUCiEKP1nn33GGBs6dKh/2LJ3mzE2c+bMk08+2Y+8rVix4rLLLps7d+4PfvCD3/3udx3VpS+RyWT8WPThqaioyGazVVVVR/Jh4rtFbW1te3s7XieTyVgsdmw3oGPItGnTJkyYMGLECP+d5ubmxx9/HA5nxlgikcjlcqFQaOzYsbNmzZoxY8bhD1hbWzt58uTx48evWbNm/vz5iqIsWbLkWJ3ttm3bFEXp0aPHoEGDBg0adN11182aNWvFihVz587dvXv3vffem8vlGhoafNfTsGHD3n777REjRvzlL3+ZPn36Ya7uWJ0hOwrxMGDAgFWrVr333nu1tbX9+/f/un8uSRJeVFVVjR079qqrrsKPW7duPemkk0RRPOecc9auXdvc3HzBBRd0/MPq6up//OMf/o/79+//t3/7N7z2vSW7du3q37+/67qtra0zZsyQZXn79u133XXXKaeccuaZZ37dU/32WbVq1UMPPfTee+9169aNMYZvur6+3g+06Lre2trKGKurq9M0bdeuXTCM9u/f3zEE3ZFEInH99df/8pe/xI+NjY0bNmwoFovBYPBbuCLiu86AAQM+/PDDSy65hDH2wQcfHMUj/61RV1d31VVXLVq0yN8iE4nEzJkz+/XrxxgzDOPTTz/t1KmTZVnLli0788wzH3300YULF/5TxXHPnj2bN28uFArDhg07hme7du3apqYmmPWMMU3TTjvttE2bNmWz2dmzZ8+bN+/0009njN122234wDnnnLN06dKPP/7YMAw4YMpe3TE8Q3YUsYcrrrjimWee+clPfrJ8+fJJkyatWrVK1/Wj+MfDhw9/5ZVX/v73v2ez2Zdeeum2226Dsjxs2LC//vWv//jHPzqqxoyxH//4xzt27HjjjTeKxeLatWvff/99fCAWi/3pT38yTXPjxo2bNm1ijImieN99961cudIwDFmWDcP4p+mbJwjJZFLTtGAwWCwW77vvvmKxWCgURo0a9cEHH6xatSqdTt955522bQuC0Llz57POOmvq1KktLS2bN28ePnz4V+NyPr/4xS9Wr17NGPM8r62tra2trVAo4IVlWYf6q1gsdoQxm3Q6Tall31cmTZq0bNmyHTt2HDhwYNGiRddcc83xPqPDMXbsWNM0N2/ejB9/+tOfPvbYY1988cXBgwcfeOCBZ555hjH29NNPd+3ade7cublczn9qTNP8khvNNM0333xz2rRp8+fPP/nkk1esWHFsr/2CCy74+9//vnz58qampvb29g0bNvz+978fPnx4Pp/3PC8UCrmuu379esgDxlgikTj55JMXLFhw/vnnQ6SVvbpjy9HEHiRJGjp06NChQ3fu3Pnqq68KgnDRRRd93YMMGDDghhtuWLRoUUtLS8+ePefNmwdHR79+/WzbPuuss76k3lZWVt5///1Llix55JFHGhoa7rnnni5dujDGpk6d+vTTT1988cXxePy8887Dh2fNmrV48eLly5dXVFSMGjXqn6bfnCBcccUVb7zxRs+ePevr66dPn3711Vf/x3/8x2efffbKK6/ccMMNyWTy9ttvr6urgxvn+eefnzZtWp8+fVRVHT9+/GEs5YqKiltuueXXv/51S0tLXV1dx1+99tprHY3xjtTX1zc1NUE3OTxNTU2dO3f+rtxk4msxdOjQe+6552c/+5llWZMmTRo/fvzxPqPDoSjK9OnTZ86ciR/HjRtnGMYvfvGLQqFw+umn33rrrdu2bXvttdeefPJJSZJmzJgxc+bMgQMH1tfXNzY2jhw58uWXX/bdre++++6HH344ZcqUPn36fBOn2q1btyVLlixdunTGjBn5fL5z586TJk0aPnw4Y2zixImzZs1SFOWcc86ZOHHi8uXL+/fv39jYeO655z744IPnn3/+oa7umJ+k4Hnel946AadYfNOcyOOA9uzZ884771x++eWMsWKxWFlZuXfvXj+A8c3R0tKydu1a1L4dnpdffnnIkCFUukh8E3y3tiMaB0R8q4ii+F//9V+vvfZaOp2+7777Bg8e/C3IBsZYp06dOnfu/E/rdNavX9+5c2eSDQTx/YPEw4lOQ0PD888/f9ddd/3oRz/avn37ihUrvrV/PXjw4Gw2exgJsX79+mw2S24lgvheQs4lxk5s59JxZ/369c3Nzb179+7WrRtSXdPpdFNT07Zt2yjkQHzTfLe2o++Zc4nEA2MkHv4Z1NCbOF58t7aj75l4oHkPxD+nU6dOJAkI4v9vUOyBIAiCKAOJB4IgCKIMJB4IgiCIMpB4IAiCIMpA4oEgCIIoQ5nEVoIgCIIg64EgCIIoA4kHgiAIogwkHgiCIIgykHggCIIgykDigSAIgigDiQeCIAiiDCQeCIIgiDKQeCAIgiDKQOKBIAiCKAOJB4IgCKIMJB4IgiCIMpB4IAiCIMpA4oEgCIIoA4kHgiAIogwkHgiCIIgykHggCIIgykDigSAIgigDiQeCIAiiDCQeCIIgiDKQeCAIgiDKQOKBIAiCKAOJB4IgCKIMJB4IgiCIMvy/bLom8EKZzjYAAAAASUVORK5CYII=" - } - }, - "cell_type": "markdown", - "metadata": {}, - "source": [ - "![road_damage_one.png](attachment:road_damage_one.png)" - ] - }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ - "if not bb.to_dict():\n", + "from ipyannotator.mltypes import BboxCoordinate\n", + "results = list(bb.to_dict().values())\n", + "if not results or not results[0]['bbox']:\n", " \"\"\"Annotate if not manually selected a bbox\"\"\"\n", " bb.app_state.index = 6\n", - " bb.controller.save_current_annotations({\n", - " 'x': 298,\n", - " 'y': 93,\n", - " 'width': 536,\n", - " 'height': 430\n", - " })\n", - "\n", - "bb.to_dict()" + " bb.controller.save_current_annotations([\n", + " BboxCoordinate(**{\n", + " 'x': 298,\n", + " 'y': 93,\n", + " 'width': 536,\n", + " 'height': 430\n", + " })\n", + " ])" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "Let's now create a inital set of road damage images by using the mouse to draw a reactangle containing\n", + "Let's now create an inital set of road damage images by using the mouse to draw a reactangle containing\n", "the damage on individual images. Below you seed the annotation for a single image." ] }, @@ -203,7 +191,7 @@ "metadata": {}, "outputs": [], "source": [ - "img_path, bbox = list(bb.to_dict().items())[0]\n", + "img_path, bbox = list(bb.to_dict().items())[1]\n", "print(img_path)\n", "print(bbox)" ] @@ -228,12 +216,14 @@ "\n", "def crop_bboxs(bbox_annotations, source_dir, target_dir):\n", " Path(target_dir).mkdir(parents=True, exist_ok=True)\n", - " for img_file, b in bbox_annotations.items():\n", + " for img_file, items in bbox_annotations.items():\n", " img_file = img_file.split('/')[-1]\n", " # box = (left, upper, right, lower)\n", - " box_crop = (b['x'], b['y'], b['x'] + b['width'], b['y'] + b['height'])\n", - " Image.open(Path(source_dir) / img_file).crop(box_crop).save(\n", - " Path(target_dir) / img_file, quality=95)" + " if 'bbox' in items and items['bbox']:\n", + " b = items['bbox'][0]\n", + " box_crop = (b['x'], b['y'], b['x'] + b['width'], b['y'] + b['height'])\n", + " Image.open(Path(source_dir) / img_file).crop(box_crop).save(\n", + " Path(target_dir) / img_file, quality=95)" ] }, { @@ -242,7 +232,8 @@ "metadata": {}, "outputs": [], "source": [ - "crop_bboxs(bbox_annotations=bb.to_dict(), source_dir=path_japan / 'images',\n", + "to_crop = {k: v for k, v in bb.to_dict().items() if v['bbox']}\n", + "crop_bboxs(bbox_annotations=to_crop, source_dir=path_japan / 'images',\n", " target_dir=path_japan / 'images_croped')" ] }, @@ -285,12 +276,12 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "## 2) Group the damage type in groups using classification labels.\n", + "## 2. Group the road damage types in categories using classification labels.\n", "\n", "1. We can now use the `Im2ImAnnotator` to quickly explore the cropped images in order to find some typical damage types we are interested in. \n", " * The competition list the following types {D00: Longitudinal Crack, D10: Transverse Crack, D20: Aligator Crack, D40: Pothole}.\n", " * Hint: Check out https://en.wikipedia.org/wiki/Pavement_cracking to find some typical crack types.\n", - "2. Select a representative example for each damage type your are interested in and move the file to `road_japan/class_images`.\n", + "2. Select a representative example for each damage type you are interested in and move the file to `road_japan/class_images`.\n", " * remove the existing dummy images first\n", " * give the image a nice name illustrative name such as aligator_crack.jpg. The file name is used to create the class labels.\n", "3. Label the images by selecting one or more labels on the right side below \"Damage Types\"." @@ -332,23 +323,11 @@ "im2im" ] }, - { - "attachments": { - "road_damage_three.png": { - "image/png": "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" - } - }, - "cell_type": "markdown", - "metadata": {}, - "source": [ - "![road_damage_three.png](attachment:road_damage_three.png)" - ] - }, { "cell_type": "markdown", "metadata": {}, "source": [ - "We can now check the class labels that we have just created and save them to a json file." + "We can now check the class labels that we have just created and save them to a JSON file." ] }, { @@ -393,7 +372,8 @@ " if not items:\n", " items = {path_japan / 'images' / 'Japan_000060.jpg': ['Japan_000060.jpg']}\n", "\n", - " json.dump(im2im.to_dict(), outfile)" + " json.dump(im2im.to_dict(), outfile)\n", + "im2im.to_dict()" ] }, { @@ -402,7 +382,7 @@ "source": [ "## 3. Refine the inital class labels in a supervision step.\n", "\n", - "When the data have been labeled initially, supervision is a great way to further improve the data quality by reviewing annotations generated by hand or a\n", + "After initial data labeling, supervision is a great way to further improve the data quality by reviewing annotations generated by hand or a\n", "machine learning model." ] }, @@ -430,7 +410,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "We can now use the priviously generated class label to group the images by class." + "We can now use the priviously generated class labels to group the images by class." ] }, { @@ -439,7 +419,7 @@ "metadata": {}, "outputs": [], "source": [ - "with open(path / 'classification_labels.json') as infile:\n", + "with open(path_japan / 'classification_labels.json') as infile:\n", " image_annotations = json.load(infile)" ] }, @@ -527,7 +507,7 @@ "metadata": {}, "source": [ "The annotator now shows us a grid of images annotated as belonging to the same class. You can now quickly click through\n", - "this batches and select the images that belong in a different class." + "this batches and select the images that belong to a different class." ] }, { @@ -558,24 +538,11 @@ "ca" ] }, - { - "attachments": { - "road_damage_four.png": { - "image/png": "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" - } - }, - "cell_type": "markdown", - "metadata": {}, - "source": [ - "![road_damage_four.png](attachment:road_damage_four.png)" - ] - }, { "cell_type": "markdown", "metadata": {}, "source": [ - "You can repeat this process for each class and then reclassify the wrong labels in a later step. The Capture annotator is most useful when you already\n", - "have imperfect image classifications for example form an pretrained model or you have many less experienced annotators and a limited amount of experts to check there work. " + "You can repeat this process for each class and then reclassify the images with wrong labels in a later step. The Capture annotator is very useful when you already have a dataset with imperfect image classifications which could stem from a pretrained model. It is also very usefull if the dataset was annotated by many less experienced annotators which needs to be checked/improved by few experts. " ] }, { @@ -598,22 +565,15 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "This short tutorial has demonstrated how annotation UI's already included in ipyannotator can be used to quickly annotate images.\n", - "Clearly these a very simple examples and the real power of using the ipyannotator concept lays in building project specific UI's.\n", + "This short tutorial has demonstrated how annotation UI's, that are already included in Ipyannotator, can be used to quickly annotate images.\n", + "Clearly this case is a very simple example and the real power of using the Ipyannotator concept lays in building project specific UI's.\n", "Check out the other notebooks to get inspired how this can be done." ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] } ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/09_viola_example.ipynb b/nbs/09_voila_example.ipynb similarity index 63% rename from nbs/09_viola_example.ipynb rename to nbs/09_voila_example.ipynb index f109994..1f9bf39 100644 --- a/nbs/09_viola_example.ipynb +++ b/nbs/09_voila_example.ipynb @@ -9,14 +9,22 @@ "# hide\n", "%load_ext autoreload\n", "%autoreload 2\n", - "! rm -rf ../data/projects/bbox/viola_results" + "! rm -rf ../data/projects/bbox/voila_results" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ - "# Voila example" + "# Voila - Using Ipyannotator as a standalone web application\n", + "\n", + "[Voila](https://github.com/voila-dashboards/voila) is a library that turns jupyter notebooks into standalone web applications.\n", + "\n", + "Voila can be used alongside with Ipyannotator. This allows professional annotators to create annotations without even running a jupyter notebook.\n", + "\n", + "This notebook displays a bounding box annotator to exemplify how an organization can use Voila to allow external professional annotators to create datasets. \n", + "\n", + "To run this example use `voila nbs/09_voila_example.ipynb --enable_nbextensions=True`" ] }, { @@ -38,9 +46,9 @@ "outputs": [], "source": [ "input_item = InputImage(image_dir='pics', image_width=640, image_height=400)\n", - "output_item = OutputImageBbox()\n", + "output_item = OutputImageBbox(classes=['Label 01', 'Label 02'])\n", "project_path = Path('../data/projects/bbox')\n", - "annotation_file_path = construct_annotation_path(project_path, results_dir='viola_results')" + "annotation_file_path = construct_annotation_path(project_path, results_dir='voila_results')" ] }, { @@ -56,20 +64,6 @@ " annotation_file_path=annotation_file_path\n", ")" ] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] - }, - { - "cell_type": "code", - "execution_count": null, - "metadata": {}, - "outputs": [], - "source": [] } ], "metadata": { diff --git a/nbs/11_build_annotator_tutorial.ipynb b/nbs/11_build_annotator_tutorial.ipynb index a262102..3752454 100644 --- a/nbs/11_build_annotator_tutorial.ipynb +++ b/nbs/11_build_annotator_tutorial.ipynb @@ -5,20 +5,20 @@ "id": "ad7b2ebb", "metadata": {}, "source": [ - "# Build Annotator Tutorial\n", + "# Build Annotator - Understanding Ipyannotator design to easily extend and customize\n", "\n", - "The current notebook will demonstrate how to build new annotators. \n", + "Ipyannotator is a framework that allows users to *hack* the inbuilt annotators, thus, extend and customize the framework according to their needs. In the other tutorials Ipyannotator API was used in simple annotation projects to display the easy usage. The current tutorial will demonstrate how to build new annotators that can be part of the Ipyannotator API.\n", "\n", "Ipyannotator architecture uses four main layers:\n", - "- The **View** is the layer responsable for render the visualization. Ipyannotator uses [ipycanvas](https://ipycanvas.readthedocs.io/en/latest/) and [ipywidgets](https://ipywidgets.readthedocs.io/en/latest/) to structure and mount the visualization layer, but we also develop some internal components like the Navi (that help our users to navigate throught the images that will be annotated).\n", - "- The **Storage** is the layer that receive the data and stores. Ipyannotator uses different types of storage like txt, sqlite.\n", - "- The **Controller** is the layer that acts like a mediator between the state, storage and view. This layers tells when the information from the state will be stored. \n", - "- **\"Model/State (in memory)\"** is the central place of the Ipyannotator layer structure is the \"Model/State (in memory)\" that centralize the data and make sure to sync it across the application. If something changes on the Model/State then other layers can listen and sync the information.\n", + "- The **View** is responsible for rendering the visualizations. Ipyannotator uses [ipycanvas](https://ipycanvas.readthedocs.io/en/latest/) and [ipywidgets](https://ipywidgets.readthedocs.io/en/latest/) to structure and mount the visualization layer. Additionally, internal components such as the navigation menue were developed which helps the users to navigate through the images that need to be annotated.\n", "\n", + "- The **Storage** layer is the layer that receives the data and stores it. Ipyannotator uses different types of storage formats like .txt and .SQLite.\n", + "- The **Controller** layer acts as a mediator between state, storage and view. This layer tells when the information from the state will be stored.\n", + "- **\"Model/State (in memory)\"** is the central function of the Ipyannotator layer structure. It is assigned to centralize the data and ensures the syncronization across the applications. If something changes in the Model/State layer, the information is passed on to other layers, ensuring synchronization of information.\n", "\n", - "The image below shows an example of how the layers are structured and how its communications are made. \n", + "The image below exemplifies how the layers are structured and how the communication path is set up.\n", "\n", - "The annotator developed in the current notebook was a minimal one called CircleAnnotator, that draws a circle every time a user clicks on the canvas." + "The annotator developed in the current notebook is a minimal example called CircleAnnotator. It draws a circle every time a user clicks on the canvas." ] }, { @@ -36,14 +36,14 @@ "source": [ "## Model/State (in memory)\n", "\n", - "To develop a model/state ipyannotator uses [Pydantic models](https://pydantic-docs.helpmanual.io/usage/models/) to assert the data type of the output model. Every change made in a state is triggered using [PyPubSub](https://pypubsub.readthedocs.io/en/v4.0.3/) and this events can be listen to other layers to assert the sync between components.\n", + "To develop a model/state layer, Ipyannotator uses [Pydantic models](https://pydantic-docs.helpmanual.io/usage/models/) to determine the data type of the output model. Every change made in a state is monitored using [PyPubSub](https://pypubsub.readthedocs.io/en/v4.0.3/) and the information is passed on to other layers to ensure the synchronization between components.\n", "\n", - "For the `CircleAnnotator` we'll split the data into two states:\n", + "For the `CircleAnnotator` we split the data into two states:\n", "\n", - "- **AppWidgetState** it's a common state for all annotators. The `AppWidgetState` stores the canvas size, navi index and max number of images. You can use him to communicate with the Ipyannotator navigation component (Navi) or on your own custom navigation component.\n", - "- **CircleAnnotatorState** it's the state responsable to store the `CircleAnnotator` data, it will store the circle radius, view layers, current image and circle drawn.\n", + "- **AppWidgetState** is a common state for all annotators. The `AppWidgetState` stores the canvas size, navigation index and maximum number of images. You can use it to communicate with the Ipyannotator navigation component (Navi) or on your own custom navigation component.\n", + "- **CircleAnnotatorState** is the state responsible to store the `CircleAnnotator` data. Is stores the circle radius, view layers, current image, and circle drawn.\n", "\n", - "**Observation:** The model/state doesn't have to be restricted to a single class (as shown in the image above), its data should make sense accordlying to the structure of the annotator. " + "**Observation:** The model/state doesn't have to be restricted to a single class (as shown in the image above). Its data should make sense according to the structure of the annotator. " ] }, { @@ -55,9 +55,9 @@ "source": [ "from pubsub import pub\n", "from typing import Tuple, List, Dict, Optional\n", - "from ipyannotator.base import BaseState, AppWidgetState\n", + "from ipyannotator.base import BaseState, AppWidgetState, Annotator\n", "from abc import ABC, abstractmethod\n", - "from IPython.core.display import display" + "from IPython.display import display" ] }, { @@ -85,7 +85,7 @@ "source": [ "## View\n", "\n", - "The view layer should store all ipywidgets that will be used on the annotator. The next commands will start the GUI development of the CircleAnnotator." + "The view layer should stores all ipywidgets that are used by the annotator. The next commands will start the GUI for the CircleAnnotator." ] }, { @@ -99,7 +99,7 @@ "from ipycanvas import MultiCanvas\n", "from pathlib import Path\n", "from ipyannotator.navi_widget import Navi\n", - "from ipyannotator.bbox_canvas import draw_img, draw_bg\n", + "from ipyannotator.bbox_canvas import ImageRenderer, draw_bg\n", "from ipyannotator.debug_utils import IpyLogger\n", "from ipyannotator.storage import MapeableStorage, get_image_list_from_folder" ] @@ -109,7 +109,7 @@ "id": "78a251f9", "metadata": {}, "source": [ - "The `CircleCanvas` class will be a component of our GUI, allowing to draw circles, backgrounds, images and also clear them." + "The `CircleCanvas` class will be a component of our GUI, allowing to draw circles, backgrounds, images and also clears them." ] }, { @@ -144,7 +144,7 @@ " draw_bg(self._multi_canvas[layer])\n", "\n", " def draw_image(self, layer: int, image_path: str):\n", - " draw_img(self._multi_canvas[layer], Path(image_path), clear=True)" + " ImageRenderer(clear=True).render(self._multi_canvas[layer], image_path)" ] }, { @@ -193,7 +193,7 @@ "id": "6a52f48a", "metadata": {}, "source": [ - "The ` CircleAnnotatorGUI ` is our view. This layer communicates with the states, for example, if the state index changes our view will clear the draw layer, change the image and redraw the circles that were load to the state." + "The ` CircleAnnotatorGUI ` corresponds to the view layer. This layer communicates with the states, for example, if the state index changes the view layer will clear the draw layer, change the image and redraw the circles that were load to the state." ] }, { @@ -298,7 +298,7 @@ "source": [ "## Storage\n", "\n", - "Ipyannotator uses json as a data structure to store the annotation data. The package also allows the users to change the type of storage accordlying to the users needs, for example, you can store your data in files or databases like sqlite. In this example it's developed a `Storage` module that will keep our data in memory (using the `InMemoryStorage` class)." + "Ipyannotator uses JSON as a data structure to store the annotation data. The package also allows the users to change the type of storage according to the users needs. For example, you can store your data in files or databases like SQlite. In this tutorial a `Storage` module is developed that keeps our data in memory (using the `InMemoryStorage` class)." ] }, { @@ -335,7 +335,7 @@ " self.update({str(image): [] for image in self.images})\n", "\n", " def get_image(self, index: int) -> str:\n", - " return str(self.images[index])\n", + " return str(self.images[index]) # type: ignore\n", "\n", " def bulk_annotation(self, index: int, annotations: list):\n", " image_path = self.get_image(index)\n", @@ -353,9 +353,9 @@ "source": [ "## Controller\n", "\n", - "The controller serves as a mediator between the states, the gui and the storage. This layer listens for states changes and stores the data on the storage, it also can load the storage data into the states.\n", + "The controller serves as a mediator between the states, the GUI, and the storage. This layer listens to states changes and stores the data on the storage. It can also load the storage data into the states.\n", "\n", - "To demonstrate how the communication works, the `IpyLogger` class can be used as a decorator to output all the pubsub communication into the logger. The `pub.ALL_TOPICS` parameter will get all the messages." + "To demonstrate how the communication works, the `IpyLogger` class can be used as a [decorator](https://docs.python.org/3/glossary.html#term-decorator) to output all the pubsub communication into the logger. The `pub.ALL_TOPICS` parameter will get all the messages." ] }, { @@ -493,9 +493,9 @@ "source": [ "## Annotator\n", "\n", - "Ipyannotator's design can be described by three properties: input, output, actions. The goal is to develop flexible modules with a common interface.\n", + "The Ipyannotator design can be described by three properties: input, output, actions. The goal is to develop flexible modules with a common interface.\n", "\n", - "With all `CircleAnnotator` layers developed we can now create it's single instance. For the current annotator the properties used are:\n", + "With all `CircleAnnotator` layers developed we can now create a single instance. For the current annotator these are the used properties:\n", "\n", "- input: Image\n", "- output: Circle\n", @@ -530,7 +530,7 @@ "metadata": {}, "outputs": [], "source": [ - "class CircleAnnotator:\n", + "class CircleAnnotator(Annotator):\n", " def __init__(\n", " self,\n", " project_path: Path,\n", @@ -538,21 +538,24 @@ " output_item: Output,\n", " *args, **kwargs\n", " ):\n", - " self._app_widget = AppWidgetState(uuid=str(id(self)), **{\n", + " app_state = AppWidgetState(uuid=str(id(self)), **{\n", " 'size': (input_item.width, input_item.height)\n", " })\n", + "\n", + " super().__init__(app_state)\n", + "\n", " self._circle_state = CircleAnnotatorState(uuid=str(id(self)))\n", "\n", " self._storage = InMemoryStorage(project_path / input_item.dir)\n", "\n", " self._controller = CircleAnnotatorController(\n", - " self._app_widget,\n", + " self.app_state,\n", " self._circle_state,\n", " self._storage\n", " )\n", "\n", " self._view = CircleAnnotatorGUI(\n", - " self._app_widget,\n", + " self.app_state,\n", " self._circle_state\n", " )\n", "\n", @@ -593,9 +596,9 @@ "id": "736be9ab", "metadata": {}, "source": [ - "The following sequence diagram shows how the CircleAnnotator comunicates with its components when a user clicks on the next button navigation.\n", + "The following sequence diagram shows how the CircleAnnotator communicates with its components when a user clicks on the next button navigation.\n", "\n", - "![sequence diagram](https://plantuml.palaimon.io/png/XLFDRjim3BxhARZcqXtw0aOH84w0Oi0s13227Oh2e6NM5QB8dYIdoPv-bAMaTcfi9rlKzqFoizrUcGuj7i3HxvwCf1_a73QqqgenO5NpviMNpc9pGF3K_W5lUn9Y9NrhOUV82b6r9w2JBpnwWaMdp5wmf5TITMWyhBhkbweRIW1qW5qNts42N2ihDQsCQVcojGD4aAc13QBBtTFksnqileUkpgHr-yxtNldpBPTnWn6Vdtfr0Vt4eqet9ZmNEWXLkYUOwgpH7D4bg1oNUfLOZIKo7zt9rdZRwi5YdTw3pTfRFVAvucxwJHHDTkJuGOrcWWFYLhUxlDZb0RSjGDJeiK97KB0CvnRgcro840qyBEEW6HWEgF963CTG3kePX-vBPMewYTWftrp2oQ3lM9pYFJnArBf2kL_2SsbuoZ8KiDBq8b1wTHooJLnnJPW5jqK6ZpodUZqpLzMd5r7J3ELIsHQ21tjOBaTxoA1CtSZUMiwgVELlbiP2J1EZnR7n9i-WwlM-nBXcrHfthz6baR_Em9kmZmElyVxYaw1N6zxj9W_mSNEMV6yD34ptFA5EXcNor7Fkau-fJEUyAip-8tF5pvkVA4xQcXMToIFzJXo6V4FJK5pLRGW9DThHvTV9W4ze8Stq3rnjyJsf_nSMg-ul)" + "![sequence diagram](https://www.plantuml.com/plantuml/png/XLFDRjim3BxpARZsqXts0aPHD3z0CM0R0WJ13aLXq3Rh2f5bJwBIP4y_EfQxE32sKoJo-o7rnOz1o4jiB8IzSHrvQZ3mhyYkvEyS0jMyiAPsw4tz9l2fyrGtXCBjRnGV6M1HIkjn5zW35EqH-IXR8M6yxOpRWqgAAKr7Jd3HTJzDLNC2K43gkk6C4-3A-DBomhbMIDNF461NeHeCBZTFkwytUFkjd-h4rhRlsXSZfskkuiv6Ud-APWJze8D97TV_tjfUgB2HSQgp8dUWaA3bPIcQnAezi_ixNTawyQqzMwpIkRTPYRSNFYFkUjv4iUml79KwCGCDA39kTiljRjdZDbk4YeGA2enRbQ6Q-_fw2T17WryUXaKpT1fG8GxwgvQ7mJ8CBBbn5H_XND3EHpWPnax5UUZZVKdM5bJk7_0vTxfbtXUeiFm2L8evAFI32-D11NNA3EzrJ_DwKgwfZYzGyGmbLHGFcwqI7oxU8SCyJLD6xzb9_kgfuIGqqY0HqYRhPOP5jFkSXcSshGjtba9Q-VCCl6PjDbJpNV8PeQEDec2zLFXaECyIlSCpCpnFg9DbbJpndFtB3wbCznmLPWbmNPn_-OdYPAmP_cmUwNFICCetSZKFJtKTGa8fu_hJoL1lv37jz0zSvUazgVyNDbG3FBAhOcF_0000)" ] }, { @@ -609,7 +612,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/13_datasets_legacy.ipynb b/nbs/13_datasets_legacy.ipynb index cbb6f00..60db275 100644 --- a/nbs/13_datasets_legacy.ipynb +++ b/nbs/13_datasets_legacy.ipynb @@ -784,7 +784,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/14_datasets_factory_legacy.ipynb b/nbs/14_datasets_factory_legacy.ipynb index 7080de2..b5a1947 100644 --- a/nbs/14_datasets_factory_legacy.ipynb +++ b/nbs/14_datasets_factory_legacy.ipynb @@ -37,6 +37,13 @@ "from ipyannotator.datasets.generators import create_object_detection, xyxy_to_xywh" ] }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# Datasets factory" + ] + }, { "cell_type": "code", "execution_count": null, @@ -89,7 +96,7 @@ " project_file = project_path / 'annotations.json'\n", " image_dir = 'images'\n", " label_dir = None\n", - " im_width = 50\n", + " im_width = 200\n", " im_height = im_width\n", "\n", " create_object_detection(path=project_path, n_samples=50, n_objects=1, size=(500, 500))\n", @@ -207,7 +214,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/15_coordinates_input.ipynb b/nbs/15_coordinates_input.ipynb index 483a6b6..444cc1c 100644 --- a/nbs/15_coordinates_input.ipynb +++ b/nbs/15_coordinates_input.ipynb @@ -61,6 +61,14 @@ "from typing import Callable, Optional" ] }, + { + "cell_type": "markdown", + "id": "57dbda92", + "metadata": {}, + "source": [ + "# Coordinates Input" + ] + }, { "cell_type": "code", "execution_count": null, @@ -76,9 +84,11 @@ " uuid: int = None,\n", " bbox_coord: BboxCoordinate = None,\n", " input_max: BboxCoordinate = None,\n", - " coord_changed: Optional[Callable] = None\n", + " coord_changed: Optional[Callable] = None,\n", + " disabled: bool = False\n", " ):\n", " super().__init__()\n", + " self.disabled = disabled\n", " self.uuid = uuid\n", " self._input_max = input_max\n", " self.coord_changed = coord_changed\n", @@ -103,7 +113,8 @@ " min=0,\n", " max=None if self._input_max is None else getattr(self._input_max, in_p),\n", " layout=Layout(width=\"55px\"),\n", - " continuous_update=False\n", + " continuous_update=False,\n", + " disabled=self.disabled\n", " )\n", " widget_inputs.append(widget_input)\n", " widget_input.observe(self._on_coord_change, names=\"value\")\n", @@ -147,8 +158,7 @@ "metadata": {}, "outputs": [], "source": [ - "# hide\n", - "\n", + "#hide\n", "inp_coord = CoordinateInput(\n", " input_max=BboxCoordinate(*[2, 2, 100, 100]),\n", " bbox_coord=BboxCoordinate(*[1, 1, 3, 88])\n", @@ -206,6 +216,25 @@ " assert value == 2" ] }, + { + "cell_type": "code", + "execution_count": null, + "id": "2cad0388", + "metadata": {}, + "outputs": [], + "source": [ + "%%ipytest\n", + "def test_it_disabled_all_input_if_coordinate_input_is_disabled():\n", + " inp_coord = CoordinateInput(\n", + " input_max=BboxCoordinate(*[2, 2, 100, 100]),\n", + " bbox_coord=BboxCoordinate(*[1, 1, 3, 88]),\n", + " disabled=True\n", + " )\n", + " \n", + " for inp in inp_coord.inputs:\n", + " assert inp.disabled is True" + ] + }, { "cell_type": "code", "execution_count": null, diff --git a/nbs/16_custom_buttons.ipynb b/nbs/16_custom_buttons.ipynb index b567852..354bd28 100644 --- a/nbs/16_custom_buttons.ipynb +++ b/nbs/16_custom_buttons.ipynb @@ -45,6 +45,14 @@ "from ipywidgets import Button" ] }, + { + "cell_type": "markdown", + "id": "66e35533", + "metadata": {}, + "source": [ + "# Custom Buttons" + ] + }, { "cell_type": "code", "execution_count": null, @@ -57,7 +65,14 @@ "class ActionButton(Button):\n", " def __init__(self, value=None, **kwargs):\n", " super().__init__(**kwargs)\n", - " self.value = value" + " self.value = value\n", + "\n", + " def reset_callbacks(self):\n", + " self.on_click(None, remove=True)\n", + "\n", + " def update(self, other):\n", + " self.value = other.value\n", + " self.layout = other.layout" ] }, { @@ -125,7 +140,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/17_annotator_explorer.ipynb b/nbs/17_annotator_explorer.ipynb index 55a9eb3..c81efb1 100644 --- a/nbs/17_annotator_explorer.ipynb +++ b/nbs/17_annotator_explorer.ipynb @@ -22,19 +22,6 @@ "# default_exp explore_annotator" ] }, - { - "cell_type": "code", - "execution_count": null, - "id": "857d758a", - "metadata": {}, - "outputs": [], - "source": [ - "#exporti\n", - "from typing import Optional\n", - "\n", - "from ipyannotator.im2im_annotator import ImCanvas" - ] - }, { "cell_type": "code", "execution_count": null, @@ -43,15 +30,24 @@ "outputs": [], "source": [ "#exporti\n", - "from ipyannotator.base import BaseState, AppWidgetState\n", + "from ipyannotator.im2im_annotator import ImCanvas\n", + "from ipyannotator.base import BaseState, AppWidgetState, Annotator\n", "from ipyannotator.navi_widget import Navi\n", "from ipyannotator.storage import MapeableStorage, get_image_list_from_folder\n", - "from ipyannotator.mltypes import Input, Output\n", + "from ipyannotator.mltypes import InputImage, Output\n", "from abc import ABC, abstractmethod\n", "from IPython.display import display\n", "from pathlib import Path\n", "from ipywidgets import AppLayout, HBox, Layout\n", - "from typing import Any, List" + "from typing import Any, List, Optional" + ] + }, + { + "cell_type": "markdown", + "id": "493a4790", + "metadata": {}, + "source": [ + "# Annotator Explorer" ] }, { @@ -78,7 +74,13 @@ "\n", "class ExploreAnnotatorGUI(AppLayout):\n", "\n", - " def __init__(self, app_state: AppWidgetState, explorer_state: ExploreAnnotatorState):\n", + " def __init__(\n", + " self,\n", + " app_state: AppWidgetState,\n", + " explorer_state: ExploreAnnotatorState,\n", + " fit_canvas: bool = False,\n", + " has_border: bool = False\n", + " ):\n", " self._app_state = app_state\n", " self._state = explorer_state\n", "\n", @@ -95,7 +97,9 @@ "\n", " self._image = ImCanvas(\n", " width=self._app_state.size[0],\n", - " height=self._app_state.size[1]\n", + " height=self._app_state.size[1],\n", + " fit_canvas=fit_canvas,\n", + " has_border=has_border\n", " )\n", "\n", " # set the values already instantiated on state\n", @@ -293,32 +297,38 @@ "source": [ "#export\n", "\n", - "class ExploreAnnotator:\n", + "class ExploreAnnotator(Annotator):\n", " def __init__(\n", " self,\n", " project_path: Path,\n", - " input_item: Input,\n", + " input_item: InputImage,\n", " output_item: Output,\n", + " has_border: bool = False,\n", " *args, **kwargs\n", " ):\n", - " self._app_state = AppWidgetState(uuid=str(id(self)), **{\n", + " app_state = AppWidgetState(uuid=str(id(self)), **{\n", " # \"Input\" has no attribute \"width\", \"height\"\n", " 'size': (input_item.width, input_item.height) # type: ignore\n", " })\n", + "\n", + " super().__init__(app_state)\n", + "\n", " self._state = ExploreAnnotatorState(uuid=str(id(self)))\n", "\n", " # \"Input\" has no attribute \"dir\"\n", " self._storage = InMemoryStorage(project_path / input_item.dir) # type: ignore\n", "\n", " self._controller = ExploreAnnotatorController(\n", - " self._app_state,\n", + " self.app_state,\n", " self._state,\n", " self._storage\n", " )\n", "\n", " self._view = ExploreAnnotatorGUI(\n", - " self._app_state,\n", - " self._state\n", + " self.app_state,\n", + " self._state,\n", + " fit_canvas=input_item.fit_canvas,\n", + " has_border=has_border\n", " )\n", "\n", " def __repr__(self):\n", @@ -333,9 +343,9 @@ "metadata": {}, "outputs": [], "source": [ - "from ipyannotator.mltypes import InputImage, NoOutput\n", + "from ipyannotator.mltypes import NoOutput\n", "\n", - "ExploreAnnotator(\n", + "exp = ExploreAnnotator(\n", " project_path=Path('../data/projects/bbox/'),\n", " input_item=InputImage(image_dir='pics', image_width=400, image_height=400),\n", " output_item=NoOutput()\n", @@ -365,7 +375,7 @@ ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/nbs/18_bbox_trajectory.ipynb b/nbs/18_bbox_trajectory.ipynb index 5de0a56..35070f1 100644 --- a/nbs/18_bbox_trajectory.ipynb +++ b/nbs/18_bbox_trajectory.ipynb @@ -44,7 +44,7 @@ "#exporti\n", "from ipycanvas import Canvas\n", "from typing import List\n", - "from collections.abc import MutableMapping\n", + "from ipyannotator.ipytyping.annotations import AnnotationStore\n", "from ipyannotator.mltypes import BboxCoordinate" ] }, @@ -53,7 +53,7 @@ "id": "a907db61", "metadata": {}, "source": [ - "## Bounding Box Trajectory\n", + "# Bounding Box Trajectory\n", "\n", "The current notebook develop the data type and algorithms to store and process trajectories." ] @@ -66,29 +66,20 @@ "outputs": [], "source": [ "#exporti\n", - "class TrajectoryStore(MutableMapping):\n", - " def __init__(self):\n", - " self.track = {}\n", - "\n", + "class TrajectoryStore(AnnotationStore):\n", " def __getitem__(self, key: str):\n", " assert isinstance(key, str)\n", - " return self.track[key]\n", + " return self._annotations[key]\n", "\n", " def __delitem__(self, key: str):\n", " assert isinstance(key, str)\n", " if key in self:\n", - " del self.track[key]\n", + " del self._annotations[key]\n", "\n", " def __setitem__(self, key: str, value: List[BboxCoordinate]):\n", " assert isinstance(key, str)\n", " assert isinstance(value, list)\n", - " self.track[key] = value\n", - "\n", - " def __iter__(self):\n", - " return iter(self.track)\n", - "\n", - " def __len__(self):\n", - " return len(self.track)" + " self._annotations[key] = value" ] }, { diff --git a/nbs/19_bbox_video_annotator.ipynb b/nbs/19_bbox_video_annotator.ipynb index 5048e5d..1fd4345 100644 --- a/nbs/19_bbox_video_annotator.ipynb +++ b/nbs/19_bbox_video_annotator.ipynb @@ -73,6 +73,14 @@ "ipytest.autoconfig(raise_on_error=True)" ] }, + { + "cell_type": "markdown", + "id": "79e9e594", + "metadata": {}, + "source": [ + "# Bbox video annotator" + ] + }, { "cell_type": "code", "execution_count": null, @@ -143,6 +151,7 @@ " on_trajectory_enabled_clicked: Callable,\n", " on_btn_delete_clicked: Callable[[BboxVideoCoordinate], None]\n", " ):\n", + " self.on_label_changed = on_label_changed\n", " super().__init__(\n", " app_state,\n", " bbox_canvas_state,\n", @@ -162,22 +171,26 @@ " if on_trajectory_enabled_clicked:\n", " self.trajectory_enabled_checkbox.observe(on_trajectory_enabled_clicked, names='value')\n", "\n", + " self._bbox_state.unsubscribe('drawing_enabled')\n", " pub.unsubscribe(super()._sync_labels, f'{bbox_canvas_state.root_topic}.bbox_coords')\n", " pub.unsubscribe(super()._refresh_children, f'{app_state.root_topic}.index')\n", "\n", + " self._init_bbox_list(self._bbox_state.drawing_enabled)\n", + "\n", + " bbox_canvas_state.subscribe(self._update_max_coord_input, 'image_scale')\n", + "\n", + " self.children = self._bbox_list.children\n", + "\n", + " def _init_bbox_list(self, drawing_enabled: bool):\n", " self._bbox_list = BBoxVideoList(\n", " btn_delete_enabled=drawing_enabled,\n", - " on_label_changed=on_label_changed,\n", + " on_label_changed=self.on_label_changed,\n", " on_btn_delete_clicked=self._on_btn_delete_clicked,\n", - " on_btn_select_clicked=on_btn_select_clicked,\n", - " classes=bbox_state.classes,\n", + " on_btn_select_clicked=self.on_btn_select_clicked,\n", + " classes=self._bbox_state.classes,\n", " on_checkbox_object_clicked=self._on_checkbox_object_clicked\n", " )\n", "\n", - " bbox_canvas_state.subscribe(self._update_max_coord_input, 'image_scale')\n", - "\n", - " self.children = self._bbox_list.children\n", - "\n", " def _refresh_children(self, index: int):\n", " self._render(\n", " self._bbox_canvas_state.bbox_coords,\n", @@ -275,7 +288,8 @@ " super().__init__(\n", " app_state=app_state,\n", " bbox_state=bbox_state,\n", - " on_save_btn_clicked=on_save_btn_clicked\n", + " on_save_btn_clicked=on_save_btn_clicked,\n", + " fit_canvas=False\n", " )\n", "\n", " self._app_state = app_state\n", @@ -283,6 +297,7 @@ " self.on_bbox_drawn = on_bbox_drawn\n", " self.bbox_trajectory = BBoxTrajectory()\n", " self.history = BboxVideoHistory()\n", + " self.on_label_changed = on_label_changed\n", "\n", " pub.unsubAll(f'{self._image_box.state.root_topic}.bbox_coords')\n", "\n", @@ -299,7 +314,7 @@ " bbox_state=self._bbox_state, # type: ignore\n", " on_btn_select_clicked=self._highlight_bbox,\n", " on_btn_delete_clicked=self._remove_trajectory_history,\n", - " on_label_changed=on_label_changed,\n", + " on_label_changed=self.on_label_changed,\n", " drawing_enabled=drawing_enabled,\n", " on_trajectory_enabled_clicked=self.on_trajectory_enabled_clicked\n", " )\n", @@ -322,7 +337,7 @@ "\n", " self.btn_right_menu_enabled = ToggleButton(\n", " description=\"Menu\",\n", - " tooltip=\"Disable right menu for a better navigation experience.\",\n", + " tooltip=\"Disable right menu for a faster navigation experience.\",\n", " icon=\"eye-slash\",\n", " disabled=False,\n", " # Argument 1 to \"render_right_menu\" of \"BBoxAnnotatorVideoGUI\" has incompatible\n", @@ -578,7 +593,8 @@ " pub.unsubscribe(self.controller._idx_changed, f'{self.app_state.root_topic}.index')\n", " pub.unsubAll(f'{self.app_state.root_topic}.index')\n", " state_params = {**self.bbox_state.dict()}\n", - " state_params.pop('_uuid')\n", + " state_params.pop('_uuid', [])\n", + " state_params.pop('event_map', [])\n", " self.bbox_state = BBoxVideoState(\n", " uuid=self.bbox_state._uuid,\n", " **state_params\n", @@ -609,8 +625,8 @@ " # \"BBoxAnnotatorController\" has no attribute \"update_storage_labels\"\n", " self.controller.update_storage_labels(change, index) # type: ignore\n", "\n", - " def on_save_btn_clicked(self, bbox_coords: Dict):\n", - " self.controller.save_current_annotations(bbox_coords)\n", + " def on_save_btn_clicked(self, bbox_coords: List[BboxVideoCoordinate]):\n", + " self.controller.save_current_annotations(bbox_coords) # type: ignore\n", "\n", " def _update_state_id(self, merged_ids: List[str], bbox_coords: List[BboxVideoCoordinate]):\n", " merged_id = \"-\".join(merged_ids)\n", @@ -1158,7 +1174,7 @@ " coords = trajectory_fixture.bbox_state.coords\n", " trajectory_fixture.view.right_menu[0].btn_delete.click()\n", " assert del_coord not in trajectory_fixture.view.right_menu._bbox_canvas_state.bbox_coords\n", - " assert '0' not in trajectory_fixture.bbox_state.trajectories.track" + " assert '0' not in dict(trajectory_fixture.bbox_state.trajectories)" ] }, { diff --git a/nbs/20_image_classification_user_story.ipynb b/nbs/20_image_classification_user_story.ipynb index ca69614..8105b72 100644 --- a/nbs/20_image_classification_user_story.ipynb +++ b/nbs/20_image_classification_user_story.ipynb @@ -10,12 +10,24 @@ "#all_slow" ] }, + { + "cell_type": "markdown", + "id": "f82b9d31", + "metadata": {}, + "source": [ + "# Image classification - Real project example with CIFAR-10 dataset\n", + "\n", + "This notebook will exemplify how to do image classification in Ipyannotator using one of the most commonly used datasets in deep learning: [CIFAR-10](https://www.cs.toronto.edu/~kriz/cifar.html). The dataset contains 60000 32x32 images in 10 classes." + ] + }, { "cell_type": "markdown", "id": "b464e38d-ba40-40b1-b45d-2f1140750cfa", "metadata": {}, "source": [ - "### setup data for a fictive greenfield project" + "## Setup data for a fictive greenfield project\n", + "\n", + "The first step is to download the dataset. The next cell will use the [pooch](https://github.com/fatiando/pooch) library to easily fetch the data files from s3." ] }, { @@ -43,6 +55,14 @@ ")" ] }, + { + "cell_type": "markdown", + "id": "f48accc5", + "metadata": {}, + "source": [ + "Pooch retrieves the data to your local machine. The next cell will display the exact path where the files were downloaded." + ] + }, { "cell_type": "code", "execution_count": null, @@ -53,6 +73,16 @@ "file_path" ] }, + { + "cell_type": "markdown", + "id": "62775b51", + "metadata": {}, + "source": [ + "Since the CITAR-10 dataset is downloaded as a compressed `tar` file, the next cells will extract the files. \n", + "\n", + "Ipyannotator has some internal tools to manipulate data, which is the case of the `_extract_tar` function used below to extract the files and move them to a new folder `tmp`." + ] + }, { "cell_type": "code", "execution_count": null, @@ -73,6 +103,26 @@ "_extract_tar(file_path, Path('/tmp'))" ] }, + { + "cell_type": "markdown", + "id": "e3c7d484", + "metadata": {}, + "source": [ + "Ipyannotator uses the following path setup:\n", + "\n", + "```\n", + "project_root\n", + "│\n", + "│─── images\n", + "│\n", + "└─── results\n", + "```\n", + "\n", + "The `project root` is the folder that contains folders for the image raw data and the annotation results. `Images` is the folder that contains all images that can displayed by the navigator and are used to create the dataset by the annotator. The `results` folder stores the dataset. The folder names can be chosen by the user. By default Ipyannotator uses `images` and `results`.\n", + "\n", + "The next cell defines a project root called `user_project` and creates a new folder called `images` inside of it." + ] + }, { "cell_type": "code", "execution_count": null, @@ -84,6 +134,16 @@ "(project_root / 'images').mkdir(parents=True, exist_ok=True)" ] }, + { + "cell_type": "markdown", + "id": "4549175b", + "metadata": {}, + "source": [ + "Once the folder structure is created, the files are downloaded and extracted, they will be moved to the `images` folder. \n", + "\n", + "The next cell copies the 200 random images from the CIFAR-10 dataset to the Ipyannotator path structure." + ] + }, { "cell_type": "code", "execution_count": null, @@ -93,7 +153,7 @@ "source": [ "import shutil\n", "import random\n", - "# copy some random images\n", + "\n", "classes = \"airplane automobile bird cat deer dog frog horse ship truck\".split()\n", "for i in range(1, 200):\n", " rand_class = random.randint(0, 9)\n", @@ -115,9 +175,9 @@ "id": "d972ee2f-4c15-422a-8526-38fa228e2c13", "metadata": {}, "source": [ - "I start with a bunch of images which I need to classify. However, at the start I don't know which classes it contains (the definition of the classes might even be something ambiguous I need to come up with during the project).\n", + "In the current step we have 200 images from random classes and we need to classify them. The first step is to have a look at the images before checking which classes need to be set in the classification.\n", "\n", - "So I use ipyannotator to take a first look at the images." + "Ipyannotator uses an API to ensure easy access to the annotators. The next cell will import the `Annotator` factory, that provides a simple function `InputImage` to explore images." ] }, { @@ -131,41 +191,58 @@ "from ipyannotator.annotator import Annotator" ] }, + { + "cell_type": "markdown", + "id": "e786e563", + "metadata": {}, + "source": [ + "CIFAR-10 uses 32x32 px color images. The small size of the images makes the visualization difficult. Therefore, the `fit_canvas` property will be used in the next cell to improve the visual appearance, displaying the image at the same size of the `InputImage`." + ] + }, { "cell_type": "code", "execution_count": null, - "id": "eeecc88b", + "id": "6cf0de9c-0a16-4a4f-a21d-9c036d4d3727", "metadata": {}, "outputs": [], "source": [ - "# todo: move Settings to ipyannotator.annotator:\n", - "# fix legacy-daaset-factory:\n", - "# cannot import name 'Settings' from partially initialized module 'ipyannotator.annotator'\n", - "# (most likely due to a circular import)\n", - "from ipyannotator.base import Settings" + "input_ = InputImage(image_width=100, image_height=100, image_dir='images', fit_canvas=True)" + ] + }, + { + "cell_type": "markdown", + "id": "a66d118c", + "metadata": {}, + "source": [ + "To use the `Annotator` factory, a simple pair of `Input/Output` is used. Omitting the output, Ipyannotator will use `NoOutput` as default. In this case, the user can only navigate across the input images and labels/classes are not displayed in the explore function. " ] }, { "cell_type": "code", "execution_count": null, - "id": "6cf0de9c-0a16-4a4f-a21d-9c036d4d3727", + "id": "58021121-989d-437e-8f02-61519e2a1f83", "metadata": {}, "outputs": [], "source": [ - "input_ = InputImage(image_dir=Path('images'))\n", - "# gottcha, image_dir is relative to project dir\n", - "# -> better doc &" + "Annotator(input_).explore()" + ] + }, + { + "cell_type": "markdown", + "id": "ebcdb3cf", + "metadata": {}, + "source": [ + "As mentioned before, the Ipyannotator path setup provides some default names for the folders. These names can be changed using the `Settings` property. The next cells demonstrates how to use the settings property to customize the folder structure." ] }, { "cell_type": "code", "execution_count": null, - "id": "58021121-989d-437e-8f02-61519e2a1f83", + "id": "4962f95b", "metadata": {}, "outputs": [], "source": [ - "Annotator(input_).explore()\n", - "# should work without specifiying setting just with resonable default values" + "from ipyannotator.base import Settings" ] }, { @@ -175,7 +252,11 @@ "metadata": {}, "outputs": [], "source": [ - "settings = Settings(project_path=Path('user_project'))" + "settings = Settings(\n", + " project_path=Path('user_project'),\n", + " image_dir='images',\n", + " result_dir='results'\n", + ")" ] }, { @@ -185,8 +266,7 @@ "metadata": {}, "outputs": [], "source": [ - "anni = Annotator(input_, settings=settings)\n", - "# should work without specifiying outputs" + "anni = Annotator(input_, settings=settings)" ] }, { @@ -199,6 +279,16 @@ "anni.explore()" ] }, + { + "cell_type": "markdown", + "id": "7de0b05a", + "metadata": {}, + "source": [ + "Once the user has gained an overview on the input image dataset, the user can define classes to label the images. Using `OutputLabel` you can define the classes that will be used to label the images. \n", + "\n", + "The `class_labels` property at `OutputLabel` allows an array of classes to be used in the classification. Since CIFAR-10 uses 10 classes, these are going to be used in the next cells." + ] + }, { "cell_type": "code", "execution_count": null, @@ -206,14 +296,8 @@ "metadata": {}, "outputs": [], "source": [ - "from ipyannotator.mltypes import OutputImageLabel\n", - "\n", - "# todo: add check for empty label dir, to give autogeneration a try.\n", - "# Currently supported only 'class_autogenerated_' ^^ Ok, it's not very obvious haha\n", - "# It all comes from messy Storage stuff, which should recieave a bunch of refactorings... =(\n", - "\n", - "output_ = OutputImageLabel(label_dir=Path('class_autogenerated_'))\n", - "# label_dir is relateve again (I guess)" + "from ipyannotator.mltypes import OutputLabel\n", + "output_ = OutputLabel(class_labels=classes)" ] }, { @@ -233,11 +317,15 @@ "metadata": {}, "outputs": [], "source": [ - "anni.explore()\n", - "# failes with FileNotFoundError:\n", - "# surprise explore only works for labeled data ???\n", - "# labels need to be created first? -> user warning, what to do, actually should work without labels\n", - "# need proper path consistency checking and constructive user warnings / errors" + "anni.explore()" + ] + }, + { + "cell_type": "markdown", + "id": "229594cd", + "metadata": {}, + "source": [ + "To create your own dataset you just have to call the `create` step at the `Annotator` factory. This step will allow users to associate the classes to a image." ] }, { @@ -246,12 +334,22 @@ "id": "31eb7b67-7e50-44c4-9c30-df901cf3a647", "metadata": {}, "outputs": [], + "source": [ + "anni.create()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "id": "48a6c31b", + "metadata": {}, + "outputs": [], "source": [] } ], "metadata": { "kernelspec": { - "display_name": "Python 3 (ipykernel)", + "display_name": "Python 3", "language": "python", "name": "python3" } diff --git a/poetry.lock b/poetry.lock index 8f4446e..d75e73a 100644 --- a/poetry.lock +++ b/poetry.lock @@ -1,3 +1,11 @@ +[[package]] +name = "alabaster" +version = "0.7.12" +description = "A configurable sidebar-enabled Sphinx theme" +category = "dev" +optional = false +python-versions = "*" + [[package]] name = "anyio" version = "3.5.0" @@ -142,6 +150,21 @@ category = "main" optional = false python-versions = "*" +[[package]] +name = "beautifulsoup4" +version = "4.10.0" +description = "Screen-scraping library" +category = "dev" +optional = false +python-versions = ">3.0.0" + +[package.dependencies] +soupsieve = ">1.2" + +[package.extras] +html5lib = ["html5lib"] +lxml = ["lxml"] + [[package]] name = "black" version = "22.1.0" @@ -271,7 +294,7 @@ python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*, !=3.4.*" [[package]] name = "docutils" -version = "0.18.1" +version = "0.17.1" description = "Docutils -- Python Documentation Utilities" category = "dev" optional = false @@ -392,6 +415,17 @@ python-versions = ">=3.7" [package.dependencies] gitdb = ">=4.0.1,<5" +[[package]] +name = "greenlet" +version = "1.1.2" +description = "Lightweight in-process concurrent programming" +category = "dev" +optional = false +python-versions = ">=2.7,!=3.0.*,!=3.1.*,!=3.2.*,!=3.3.*,!=3.4.*" + +[package.extras] +docs = ["sphinx"] + [[package]] name = "idna" version = "3.3" @@ -425,6 +459,14 @@ linting = ["black", "flake8"] test = ["invoke", "pytest", "pytest-cov"] tifffile = ["tifffile"] +[[package]] +name = "imagesize" +version = "1.3.0" +description = "Getting image size from png/jpeg/jpeg2000/gif file" +category = "dev" +optional = false +python-versions = ">=2.7, !=3.0.*, !=3.1.*, !=3.2.*, !=3.3.*" + [[package]] name = "importlib-metadata" version = "4.10.1" @@ -686,6 +728,27 @@ nbconvert = "*" notebook = "*" qtconsole = "*" +[[package]] +name = "jupyter-cache" +version = "0.4.3" +description = "A defined interface for working with a cache of jupyter notebooks." +category = "dev" +optional = false +python-versions = ">=3.6" + +[package.dependencies] +attrs = "*" +nbclient = ">=0.2,<0.6" +nbdime = "*" +nbformat = "*" +sqlalchemy = ">=1.3.12,<1.5" + +[package.extras] +cli = ["click", "click-completion", "click-log", "tabulate", "pyyaml"] +code_style = ["flake8 (>=3.7.0,<3.8.0)", "black", "pre-commit (==1.17.0)"] +rtd = ["myst-nb (>=0.7,<1.0)", "sphinx-copybutton", "pydata-sphinx-theme"] +testing = ["ipykernel", "coverage", "pytest (>=3.6,<4)", "pytest-cov", "pytest-regressions", "matplotlib", "numpy", "sympy", "pandas", "nbformat (>=5.1)"] + [[package]] name = "jupyter-client" version = "6.1.12" @@ -764,6 +827,21 @@ websocket-client = "*" [package.extras] test = ["coverage", "pytest (>=6.0)", "pytest-cov", "pytest-mock", "requests", "pytest-tornasync", "pytest-console-scripts", "ipykernel"] +[[package]] +name = "jupyter-sphinx" +version = "0.3.2" +description = "Jupyter Sphinx Extensions" +category = "dev" +optional = false +python-versions = ">= 3.6" + +[package.dependencies] +IPython = "*" +ipywidgets = ">=7.0.0" +nbconvert = ">=5.5" +nbformat = "*" +Sphinx = ">=2" + [[package]] name = "jupyterlab" version = "3.2.9" @@ -860,6 +938,25 @@ category = "dev" optional = false python-versions = ">=3.6" +[[package]] +name = "markdown-it-py" +version = "1.1.0" +description = "Python port of markdown-it. Markdown parsing, done right!" +category = "dev" +optional = false +python-versions = "~=3.6" + +[package.dependencies] +attrs = ">=19,<22" + +[package.extras] +code_style = ["pre-commit (==2.6)"] +compare = ["commonmark (>=0.9.1,<0.10.0)", "markdown (>=3.2.2,<3.3.0)", "mistletoe-ebp (>=0.10.0,<0.11.0)", "mistune (>=0.8.4,<0.9.0)", "panflute (>=1.12,<2.0)"] +linkify = ["linkify-it-py (>=1.0,<2.0)"] +plugins = ["mdit-py-plugins"] +rtd = ["myst-nb (==0.13.0a1)", "pyyaml", "sphinx (>=2,<4)", "sphinx-copybutton", "sphinx-panels (>=0.4.0,<0.5.0)", "sphinx-book-theme"] +testing = ["coverage", "psutil", "pytest (>=3.6,<4)", "pytest-benchmark (>=3.2,<4.0)", "pytest-cov", "pytest-regressions"] + [[package]] name = "markupsafe" version = "2.0.1" @@ -906,6 +1003,22 @@ category = "dev" optional = false python-versions = "*" +[[package]] +name = "mdit-py-plugins" +version = "0.2.8" +description = "Collection of plugins for markdown-it-py" +category = "dev" +optional = false +python-versions = "~=3.6" + +[package.dependencies] +markdown-it-py = ">=1.0,<2.0" + +[package.extras] +code_style = ["pre-commit (==2.6)"] +rtd = ["myst-parser (==0.14.0a3)", "sphinx-book-theme (>=0.1.0,<0.2.0)"] +testing = ["coverage", "pytest (>=3.6,<4)", "pytest-cov", "pytest-regressions"] + [[package]] name = "mistune" version = "0.8.4" @@ -939,6 +1052,55 @@ category = "main" optional = false python-versions = "*" +[[package]] +name = "myst-nb" +version = "0.13.2" +description = "A Jupyter Notebook Sphinx reader built on top of the MyST markdown parser." +category = "dev" +optional = false +python-versions = ">=3.6" + +[package.dependencies] +docutils = ">=0.15,<0.18" +importlib-metadata = "*" +ipython = "*" +ipywidgets = ">=7.0.0,<8" +jupyter-cache = ">=0.4.1,<0.5.0" +jupyter-sphinx = ">=0.3.2,<0.4.0" +myst-parser = ">=0.15.2,<0.16.0" +nbconvert = ">=5.6,<7" +nbformat = ">=5.0,<6.0" +pyyaml = "*" +sphinx = ">=3.1,<5" +sphinx-togglebutton = ">=0.3.0,<0.4.0" + +[package.extras] +code_style = ["pre-commit (>=2.12,<3.0)"] +rtd = ["alabaster", "altair", "bokeh", "coconut (>=1.4.3,<1.5.0)", "ipykernel (>=5.5,<6.0)", "ipywidgets", "jupytext (>=1.11.2,<1.12.0)", "matplotlib", "numpy", "pandas", "plotly", "sphinx-book-theme (>=0.1.0,<0.2.0)", "sphinx-copybutton", "sphinx-panels (>=0.4.1,<0.5.0)", "sphinxcontrib-bibtex", "sympy"] +testing = ["coverage (<5.0)", "ipykernel (>=5.5,<6.0)", "ipython (<8)", "jupytext (>=1.11.2,<1.12.0)", "matplotlib (>=3.3.0,<3.4.0)", "numpy", "pandas (<1.4)", "pytest (>=5.4,<6.0)", "pytest-cov (>=2.8,<3.0)", "pytest-regressions", "sympy"] + +[[package]] +name = "myst-parser" +version = "0.15.2" +description = "An extended commonmark compliant parser, with bridges to docutils & sphinx." +category = "dev" +optional = false +python-versions = ">=3.6" + +[package.dependencies] +docutils = ">=0.15,<0.18" +jinja2 = "*" +markdown-it-py = ">=1.0.0,<2.0.0" +mdit-py-plugins = ">=0.2.8,<0.3.0" +pyyaml = "*" +sphinx = ">=3.1,<5" + +[package.extras] +code_style = ["pre-commit (>=2.12,<3.0)"] +linkify = ["linkify-it-py (>=1.0,<2.0)"] +rtd = ["ipython", "sphinx-book-theme (>=0.1.0,<0.2.0)", "sphinx-panels (>=0.5.2,<0.6.0)", "sphinxcontrib-bibtex (>=2.1,<3.0)", "sphinxext-rediraffe (>=0.2,<1.0)", "sphinxcontrib.mermaid (>=0.6.3,<0.7.0)", "sphinxext-opengraph (>=0.4.2,<0.5.0)"] +testing = ["beautifulsoup4", "coverage", "docutils (>=0.17.0,<0.18.0)", "pytest (>=3.6,<4)", "pytest-cov", "pytest-regressions"] + [[package]] name = "nbclassic" version = "0.3.5" @@ -1371,6 +1533,25 @@ typing-extensions = ">=3.7.4.3" dotenv = ["python-dotenv (>=0.10.4)"] email = ["email-validator (>=1.0.3)"] +[[package]] +name = "pydata-sphinx-theme" +version = "0.8.0" +description = "Bootstrap-based Sphinx theme from the PyData community" +category = "dev" +optional = false +python-versions = ">=3.7" + +[package.dependencies] +beautifulsoup4 = "*" +docutils = "!=0.17.0" +sphinx = "*" + +[package.extras] +doc = ["numpydoc", "myst-parser", "pandas", "pytest", "pytest-regressions", "sphinxext-rediraffe", "sphinx-sitemap", "jupyter-sphinx", "plotly", "numpy", "xarray"] +test = ["pytest", "pydata-sphinx-theme"] +coverage = ["pytest-cov", "codecov", "pydata-sphinx-theme"] +dev = ["pyyaml", "pre-commit", "nox", "pydata-sphinx-theme"] + [[package]] name = "pyflakes" version = "2.4.0" @@ -1729,6 +1910,172 @@ category = "main" optional = false python-versions = ">=3.5" +[[package]] +name = "snowballstemmer" +version = "2.2.0" +description = "This package provides 29 stemmers for 28 languages generated from Snowball algorithms." +category = "dev" +optional = false +python-versions = "*" + +[[package]] +name = "soupsieve" +version = "2.3.1" +description = "A modern CSS selector implementation for Beautiful Soup." +category = "dev" +optional = false +python-versions = ">=3.6" + +[[package]] +name = "sphinx" +version = "4.4.0" +description = "Python documentation generator" +category = "dev" +optional = false +python-versions = ">=3.6" + +[package.dependencies] +alabaster = ">=0.7,<0.8" +babel = ">=1.3" +colorama = {version = ">=0.3.5", markers = "sys_platform == \"win32\""} +docutils = ">=0.14,<0.18" +imagesize = "*" +importlib-metadata = {version = ">=4.4", markers = "python_version < \"3.10\""} +Jinja2 = ">=2.3" +packaging = "*" +Pygments = ">=2.0" +requests = ">=2.5.0" +snowballstemmer = ">=1.1" +sphinxcontrib-applehelp = "*" +sphinxcontrib-devhelp = "*" +sphinxcontrib-htmlhelp = ">=2.0.0" +sphinxcontrib-jsmath = "*" +sphinxcontrib-qthelp = "*" +sphinxcontrib-serializinghtml = ">=1.1.5" + +[package.extras] +docs = ["sphinxcontrib-websupport"] +lint = ["flake8 (>=3.5.0)", "isort", "mypy (>=0.931)", "docutils-stubs", "types-typed-ast", "types-requests"] +test = ["pytest", "pytest-cov", "html5lib", "cython", "typed-ast"] + +[[package]] +name = "sphinx-togglebutton" +version = "0.3.0" +description = "Toggle page content and collapse admonitions in Sphinx." +category = "dev" +optional = false +python-versions = "*" + +[package.dependencies] +docutils = "*" +sphinx = "*" + +[package.extras] +sphinx = ["myst-parser", "sphinx-book-theme", "sphinx-design"] + +[[package]] +name = "sphinxcontrib-applehelp" +version = "1.0.2" +description = "sphinxcontrib-applehelp is a sphinx extension which outputs Apple help books" +category = "dev" +optional = false +python-versions = ">=3.5" + +[package.extras] +lint = ["flake8", "mypy", "docutils-stubs"] +test = ["pytest"] + +[[package]] +name = "sphinxcontrib-devhelp" +version = "1.0.2" +description = "sphinxcontrib-devhelp is a sphinx extension which outputs Devhelp document." +category = "dev" +optional = false +python-versions = ">=3.5" + +[package.extras] +lint = ["flake8", "mypy", "docutils-stubs"] +test = ["pytest"] + +[[package]] +name = "sphinxcontrib-htmlhelp" +version = "2.0.0" +description = "sphinxcontrib-htmlhelp is a sphinx extension which renders HTML help files" +category = "dev" +optional = false +python-versions = ">=3.6" + +[package.extras] +lint = ["flake8", "mypy", "docutils-stubs"] +test = ["pytest", "html5lib"] + +[[package]] +name = "sphinxcontrib-jsmath" +version = "1.0.1" +description = "A sphinx extension which renders display math in HTML via JavaScript" +category = "dev" +optional = false +python-versions = ">=3.5" + +[package.extras] +test = ["pytest", "flake8", "mypy"] + +[[package]] +name = "sphinxcontrib-qthelp" +version = "1.0.3" +description = "sphinxcontrib-qthelp is a sphinx extension which outputs QtHelp document." +category = "dev" +optional = false +python-versions = ">=3.5" + +[package.extras] +lint = ["flake8", "mypy", "docutils-stubs"] +test = ["pytest"] + +[[package]] +name = "sphinxcontrib-serializinghtml" +version = "1.1.5" +description = "sphinxcontrib-serializinghtml is a sphinx extension which outputs \"serialized\" HTML files (json and pickle)." +category = "dev" +optional = false +python-versions = ">=3.5" + +[package.extras] +lint = ["flake8", "mypy", "docutils-stubs"] +test = ["pytest"] + +[[package]] +name = "sqlalchemy" +version = "1.4.31" +description = "Database Abstraction Library" +category = "dev" +optional = false +python-versions = "!=3.0.*,!=3.1.*,!=3.2.*,!=3.3.*,!=3.4.*,!=3.5.*,>=2.7" + +[package.dependencies] +greenlet = {version = "!=0.4.17", markers = "python_version >= \"3\" and (platform_machine == \"aarch64\" or platform_machine == \"ppc64le\" or platform_machine == \"x86_64\" or platform_machine == \"amd64\" or platform_machine == \"AMD64\" or platform_machine == \"win32\" or platform_machine == \"WIN32\")"} + +[package.extras] +aiomysql = ["greenlet (!=0.4.17)", "aiomysql"] +aiosqlite = ["typing_extensions (!=3.10.0.1)", "greenlet (!=0.4.17)", "aiosqlite"] +asyncio = ["greenlet (!=0.4.17)"] +asyncmy = ["greenlet (!=0.4.17)", "asyncmy (>=0.2.3)"] +mariadb_connector = ["mariadb (>=1.0.1)"] +mssql = ["pyodbc"] +mssql_pymssql = ["pymssql"] +mssql_pyodbc = ["pyodbc"] +mypy = ["sqlalchemy2-stubs", "mypy (>=0.910)"] +mysql = ["mysqlclient (>=1.4.0,<2)", "mysqlclient (>=1.4.0)"] +mysql_connector = ["mysql-connector-python"] +oracle = ["cx_oracle (>=7,<8)", "cx_oracle (>=7)"] +postgresql = ["psycopg2 (>=2.7)"] +postgresql_asyncpg = ["greenlet (!=0.4.17)", "asyncpg"] +postgresql_pg8000 = ["pg8000 (>=1.16.6)"] +postgresql_psycopg2binary = ["psycopg2-binary"] +postgresql_psycopg2cffi = ["psycopg2cffi"] +pymysql = ["pymysql (<1)", "pymysql"] +sqlcipher = ["sqlcipher3-binary"] + [[package]] name = "stack-data" version = "0.1.4" @@ -1977,9 +2324,13 @@ testing = ["pytest (>=6)", "pytest-checkdocs (>=2.4)", "pytest-flake8", "pytest- [metadata] lock-version = "1.1" python-versions = "^3.8" -content-hash = "1e1fd5aae2e2d7ddd694f9e96edbd1899f83236636ad77e37257b290abe8db63" +content-hash = "9cf569820b77a1286cf470136d74d1b38a0cfe4a98df0452e56cb276d352cd14" [metadata.files] +alabaster = [ + {file = "alabaster-0.7.12-py2.py3-none-any.whl", hash = "sha256:446438bdcca0e05bd45ea2de1668c1d9b032e1a9154c2c259092d77031ddd359"}, + {file = "alabaster-0.7.12.tar.gz", hash = "sha256:a661d72d58e6ea8a57f7a86e37d86716863ee5e92788398526d58b26a4e4dc02"}, +] anyio = [ {file = "anyio-3.5.0-py3-none-any.whl", hash = "sha256:b5fa16c5ff93fa1046f2eeb5bbff2dad4d3514d6cda61d02816dba34fa8c3c2e"}, {file = "anyio-3.5.0.tar.gz", hash = "sha256:a0aeffe2fb1fdf374a8e4b471444f0f3ac4fb9f5a5b542b48824475e0042a5a6"}, @@ -2047,6 +2398,10 @@ backcall = [ {file = "backcall-0.2.0-py2.py3-none-any.whl", hash = "sha256:fbbce6a29f263178a1f7915c1940bde0ec2b2a967566fe1c65c1dfb7422bd255"}, {file = "backcall-0.2.0.tar.gz", hash = "sha256:5cbdbf27be5e7cfadb448baf0aa95508f91f2bbc6c6437cd9cd06e2a4c215e1e"}, ] +beautifulsoup4 = [ + {file = "beautifulsoup4-4.10.0-py3-none-any.whl", hash = "sha256:9a315ce70049920ea4572a4055bc4bd700c940521d36fc858205ad4fcde149bf"}, + {file = "beautifulsoup4-4.10.0.tar.gz", hash = "sha256:c23ad23c521d818955a4151a67d81580319d4bf548d3d49f4223ae041ff98891"}, +] black = [ {file = "black-22.1.0-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:1297c63b9e1b96a3d0da2d85d11cd9bf8664251fd69ddac068b98dc4f34f73b6"}, {file = "black-22.1.0-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:2ff96450d3ad9ea499fc4c60e425a1439c2120cbbc1ab959ff20f7c76ec7e866"}, @@ -2178,8 +2533,8 @@ defusedxml = [ {file = "defusedxml-0.7.1.tar.gz", hash = "sha256:1bb3032db185915b62d7c6209c5a8792be6a32ab2fedacc84e01b52c51aa3e69"}, ] docutils = [ - {file = "docutils-0.18.1-py2.py3-none-any.whl", hash = "sha256:23010f129180089fbcd3bc08cfefccb3b890b0050e1ca00c867036e9d161b98c"}, - {file = "docutils-0.18.1.tar.gz", hash = "sha256:679987caf361a7539d76e584cbeddc311e3aee937877c87346f31debc63e9d06"}, + {file = "docutils-0.17.1-py2.py3-none-any.whl", hash = "sha256:cf316c8370a737a022b72b56874f6602acf974a37a9fba42ec2876387549fc61"}, + {file = "docutils-0.17.1.tar.gz", hash = "sha256:686577d2e4c32380bb50cbb22f575ed742d58168cee37e99117a854bcd88f125"}, ] entrypoints = [ {file = "entrypoints-0.4-py3-none-any.whl", hash = "sha256:f174b5ff827504fd3cd97cc3f8649f3693f51538c7e4bdf3ef002c8429d42f9f"}, @@ -2217,6 +2572,63 @@ gitpython = [ {file = "GitPython-3.1.26-py3-none-any.whl", hash = "sha256:26ac35c212d1f7b16036361ca5cff3ec66e11753a0d677fb6c48fa4e1a9dd8d6"}, {file = "GitPython-3.1.26.tar.gz", hash = "sha256:fc8868f63a2e6d268fb25f481995ba185a85a66fcad126f039323ff6635669ee"}, ] +greenlet = [ + {file = "greenlet-1.1.2-cp27-cp27m-macosx_10_14_x86_64.whl", hash = "sha256:58df5c2a0e293bf665a51f8a100d3e9956febfbf1d9aaf8c0677cf70218910c6"}, + {file = "greenlet-1.1.2-cp27-cp27m-manylinux1_x86_64.whl", hash = "sha256:aec52725173bd3a7b56fe91bc56eccb26fbdff1386ef123abb63c84c5b43b63a"}, + {file = "greenlet-1.1.2-cp27-cp27m-manylinux2010_x86_64.whl", hash = "sha256:833e1551925ed51e6b44c800e71e77dacd7e49181fdc9ac9a0bf3714d515785d"}, + {file = "greenlet-1.1.2-cp27-cp27m-win32.whl", hash = "sha256:aa5b467f15e78b82257319aebc78dd2915e4c1436c3c0d1ad6f53e47ba6e2713"}, + {file = "greenlet-1.1.2-cp27-cp27m-win_amd64.whl", hash = "sha256:40b951f601af999a8bf2ce8c71e8aaa4e8c6f78ff8afae7b808aae2dc50d4c40"}, + {file = "greenlet-1.1.2-cp27-cp27mu-manylinux1_x86_64.whl", hash = "sha256:95e69877983ea39b7303570fa6760f81a3eec23d0e3ab2021b7144b94d06202d"}, + {file = "greenlet-1.1.2-cp27-cp27mu-manylinux2010_x86_64.whl", hash = "sha256:356b3576ad078c89a6107caa9c50cc14e98e3a6c4874a37c3e0273e4baf33de8"}, + {file = "greenlet-1.1.2-cp310-cp310-macosx_10_14_x86_64.whl", hash = "sha256:8639cadfda96737427330a094476d4c7a56ac03de7265622fcf4cfe57c8ae18d"}, + {file = "greenlet-1.1.2-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:97e5306482182170ade15c4b0d8386ded995a07d7cc2ca8f27958d34d6736497"}, + {file = "greenlet-1.1.2-cp310-cp310-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:e6a36bb9474218c7a5b27ae476035497a6990e21d04c279884eb10d9b290f1b1"}, + {file = "greenlet-1.1.2-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:abb7a75ed8b968f3061327c433a0fbd17b729947b400747c334a9c29a9af6c58"}, + {file = "greenlet-1.1.2-cp310-cp310-musllinux_1_1_x86_64.whl", hash = "sha256:b336501a05e13b616ef81ce329c0e09ac5ed8c732d9ba7e3e983fcc1a9e86965"}, + {file = "greenlet-1.1.2-cp310-cp310-win_amd64.whl", hash = "sha256:14d4f3cd4e8b524ae9b8aa567858beed70c392fdec26dbdb0a8a418392e71708"}, + {file = "greenlet-1.1.2-cp35-cp35m-macosx_10_14_x86_64.whl", hash = "sha256:17ff94e7a83aa8671a25bf5b59326ec26da379ace2ebc4411d690d80a7fbcf23"}, + {file = "greenlet-1.1.2-cp35-cp35m-manylinux1_x86_64.whl", hash = "sha256:9f3cba480d3deb69f6ee2c1825060177a22c7826431458c697df88e6aeb3caee"}, + {file = "greenlet-1.1.2-cp35-cp35m-manylinux2010_x86_64.whl", hash = "sha256:fa877ca7f6b48054f847b61d6fa7bed5cebb663ebc55e018fda12db09dcc664c"}, + {file = "greenlet-1.1.2-cp35-cp35m-win32.whl", hash = "sha256:7cbd7574ce8e138bda9df4efc6bf2ab8572c9aff640d8ecfece1b006b68da963"}, + {file = "greenlet-1.1.2-cp35-cp35m-win_amd64.whl", hash = "sha256:903bbd302a2378f984aef528f76d4c9b1748f318fe1294961c072bdc7f2ffa3e"}, + {file = "greenlet-1.1.2-cp36-cp36m-macosx_10_14_x86_64.whl", hash = "sha256:049fe7579230e44daef03a259faa24511d10ebfa44f69411d99e6a184fe68073"}, + {file = "greenlet-1.1.2-cp36-cp36m-manylinux1_x86_64.whl", hash = "sha256:dd0b1e9e891f69e7675ba5c92e28b90eaa045f6ab134ffe70b52e948aa175b3c"}, + {file = "greenlet-1.1.2-cp36-cp36m-manylinux2010_x86_64.whl", hash = "sha256:7418b6bfc7fe3331541b84bb2141c9baf1ec7132a7ecd9f375912eca810e714e"}, + {file = "greenlet-1.1.2-cp36-cp36m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:f9d29ca8a77117315101425ec7ec2a47a22ccf59f5593378fc4077ac5b754fce"}, + {file = "greenlet-1.1.2-cp36-cp36m-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:21915eb821a6b3d9d8eefdaf57d6c345b970ad722f856cd71739493ce003ad08"}, + {file = "greenlet-1.1.2-cp36-cp36m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:eff9d20417ff9dcb0d25e2defc2574d10b491bf2e693b4e491914738b7908168"}, + {file = "greenlet-1.1.2-cp36-cp36m-musllinux_1_1_x86_64.whl", hash = "sha256:b8c008de9d0daba7b6666aa5bbfdc23dcd78cafc33997c9b7741ff6353bafb7f"}, + {file = "greenlet-1.1.2-cp36-cp36m-win32.whl", hash = "sha256:32ca72bbc673adbcfecb935bb3fb1b74e663d10a4b241aaa2f5a75fe1d1f90aa"}, + {file = "greenlet-1.1.2-cp36-cp36m-win_amd64.whl", hash = "sha256:f0214eb2a23b85528310dad848ad2ac58e735612929c8072f6093f3585fd342d"}, + {file = "greenlet-1.1.2-cp37-cp37m-macosx_10_14_x86_64.whl", hash = "sha256:b92e29e58bef6d9cfd340c72b04d74c4b4e9f70c9fa7c78b674d1fec18896dc4"}, + {file = "greenlet-1.1.2-cp37-cp37m-manylinux1_x86_64.whl", hash = "sha256:fdcec0b8399108577ec290f55551d926d9a1fa6cad45882093a7a07ac5ec147b"}, + {file = "greenlet-1.1.2-cp37-cp37m-manylinux2010_x86_64.whl", hash = "sha256:93f81b134a165cc17123626ab8da2e30c0455441d4ab5576eed73a64c025b25c"}, + {file = "greenlet-1.1.2-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:1e12bdc622676ce47ae9abbf455c189e442afdde8818d9da983085df6312e7a1"}, + {file = "greenlet-1.1.2-cp37-cp37m-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:8c790abda465726cfb8bb08bd4ca9a5d0a7bd77c7ac1ca1b839ad823b948ea28"}, + {file = "greenlet-1.1.2-cp37-cp37m-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:f276df9830dba7a333544bd41070e8175762a7ac20350786b322b714b0e654f5"}, + {file = "greenlet-1.1.2-cp37-cp37m-musllinux_1_1_x86_64.whl", hash = "sha256:8c5d5b35f789a030ebb95bff352f1d27a93d81069f2adb3182d99882e095cefe"}, + {file = "greenlet-1.1.2-cp37-cp37m-win32.whl", hash = "sha256:64e6175c2e53195278d7388c454e0b30997573f3f4bd63697f88d855f7a6a1fc"}, + {file = "greenlet-1.1.2-cp37-cp37m-win_amd64.whl", hash = "sha256:b11548073a2213d950c3f671aa88e6f83cda6e2fb97a8b6317b1b5b33d850e06"}, + {file = "greenlet-1.1.2-cp38-cp38-macosx_10_14_x86_64.whl", hash = "sha256:9633b3034d3d901f0a46b7939f8c4d64427dfba6bbc5a36b1a67364cf148a1b0"}, + {file = "greenlet-1.1.2-cp38-cp38-manylinux1_x86_64.whl", hash = "sha256:eb6ea6da4c787111adf40f697b4e58732ee0942b5d3bd8f435277643329ba627"}, + {file = "greenlet-1.1.2-cp38-cp38-manylinux2010_x86_64.whl", hash = "sha256:f3acda1924472472ddd60c29e5b9db0cec629fbe3c5c5accb74d6d6d14773478"}, + {file = "greenlet-1.1.2-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e859fcb4cbe93504ea18008d1df98dee4f7766db66c435e4882ab35cf70cac43"}, + {file = "greenlet-1.1.2-cp38-cp38-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:00e44c8afdbe5467e4f7b5851be223be68adb4272f44696ee71fe46b7036a711"}, + {file = "greenlet-1.1.2-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ec8c433b3ab0419100bd45b47c9c8551248a5aee30ca5e9d399a0b57ac04651b"}, + {file = "greenlet-1.1.2-cp38-cp38-musllinux_1_1_x86_64.whl", hash = "sha256:2bde6792f313f4e918caabc46532aa64aa27a0db05d75b20edfc5c6f46479de2"}, + {file = "greenlet-1.1.2-cp38-cp38-win32.whl", hash = "sha256:288c6a76705dc54fba69fbcb59904ae4ad768b4c768839b8ca5fdadec6dd8cfd"}, + {file = "greenlet-1.1.2-cp38-cp38-win_amd64.whl", hash = "sha256:8d2f1fb53a421b410751887eb4ff21386d119ef9cde3797bf5e7ed49fb51a3b3"}, + {file = "greenlet-1.1.2-cp39-cp39-macosx_10_14_x86_64.whl", hash = "sha256:166eac03e48784a6a6e0e5f041cfebb1ab400b394db188c48b3a84737f505b67"}, + {file = "greenlet-1.1.2-cp39-cp39-manylinux1_x86_64.whl", hash = "sha256:572e1787d1460da79590bf44304abbc0a2da944ea64ec549188fa84d89bba7ab"}, + {file = "greenlet-1.1.2-cp39-cp39-manylinux2010_x86_64.whl", hash = "sha256:be5f425ff1f5f4b3c1e33ad64ab994eed12fc284a6ea71c5243fd564502ecbe5"}, + {file = "greenlet-1.1.2-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b1692f7d6bc45e3200844be0dba153612103db241691088626a33ff1f24a0d88"}, + {file = "greenlet-1.1.2-cp39-cp39-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:7227b47e73dedaa513cdebb98469705ef0d66eb5a1250144468e9c3097d6b59b"}, + {file = "greenlet-1.1.2-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:7ff61ff178250f9bb3cd89752df0f1dd0e27316a8bd1465351652b1b4a4cdfd3"}, + {file = "greenlet-1.1.2-cp39-cp39-musllinux_1_1_x86_64.whl", hash = "sha256:0051c6f1f27cb756ffc0ffbac7d2cd48cb0362ac1736871399a739b2885134d3"}, + {file = "greenlet-1.1.2-cp39-cp39-win32.whl", hash = "sha256:f70a9e237bb792c7cc7e44c531fd48f5897961701cdaa06cf22fc14965c496cf"}, + {file = "greenlet-1.1.2-cp39-cp39-win_amd64.whl", hash = "sha256:013d61294b6cd8fe3242932c1c5e36e5d1db2c8afb58606c5a67efce62c1f5fd"}, + {file = "greenlet-1.1.2.tar.gz", hash = "sha256:e30f5ea4ae2346e62cedde8794a56858a67b878dd79f7df76a0767e356b1744a"}, +] idna = [ {file = "idna-3.3-py3-none-any.whl", hash = "sha256:84d9dd047ffa80596e0f246e2eab0b391788b0503584e8945f2368256d2735ff"}, {file = "idna-3.3.tar.gz", hash = "sha256:9d643ff0a55b762d5cdb124b8eaa99c66322e2157b69160bc32796e824360e6d"}, @@ -2225,6 +2637,10 @@ imageio = [ {file = "imageio-2.14.1-py3-none-any.whl", hash = "sha256:4bc1257abe5d8c9ef89132dccd9d783c1c0bdbbcfb98c0e5fe84e8b7b9ee4975"}, {file = "imageio-2.14.1.tar.gz", hash = "sha256:709c18f800981e4286abe4bd86b6c9b5bb6e285b6b933b5ba0962ef8e7994058"}, ] +imagesize = [ + {file = "imagesize-1.3.0-py2.py3-none-any.whl", hash = "sha256:1db2f82529e53c3e929e8926a1fa9235aa82d0bd0c580359c67ec31b2fddaa8c"}, + {file = "imagesize-1.3.0.tar.gz", hash = "sha256:cd1750d452385ca327479d45b64d9c7729ecf0b3969a58148298c77092261f9d"}, +] importlib-metadata = [ {file = "importlib_metadata-4.10.1-py3-none-any.whl", hash = "sha256:899e2a40a8c4a1aec681feef45733de8a6c58f3f6a0dbed2eb6574b4387a77b6"}, {file = "importlib_metadata-4.10.1.tar.gz", hash = "sha256:951f0d8a5b7260e9db5e41d429285b5f451e928479f19d80818878527d36e95e"}, @@ -2294,6 +2710,10 @@ jupyter = [ {file = "jupyter-1.0.0.tar.gz", hash = "sha256:d9dc4b3318f310e34c82951ea5d6683f67bed7def4b259fafbfe4f1beb1d8e5f"}, {file = "jupyter-1.0.0.zip", hash = "sha256:3e1f86076bbb7c8c207829390305a2b1fe836d471ed54be66a3b8c41e7f46cc7"}, ] +jupyter-cache = [ + {file = "jupyter-cache-0.4.3.tar.gz", hash = "sha256:4c9b5431b1d320bc68440c21fa0a155bbeb29c5b979bef72222e244a7bcd54fc"}, + {file = "jupyter_cache-0.4.3-py3-none-any.whl", hash = "sha256:6d5d662d81f565d18009e8dcfd3a56fb876af47eafead2a19ef0045aba8ffe3b"}, +] jupyter-client = [ {file = "jupyter_client-6.1.12-py3-none-any.whl", hash = "sha256:e053a2c44b6fa597feebe2b3ecb5eea3e03d1d91cc94351a52931ee1426aecfc"}, {file = "jupyter_client-6.1.12.tar.gz", hash = "sha256:c4bca1d0846186ca8be97f4d2fa6d2bae889cce4892a167ffa1ba6bd1f73e782"}, @@ -2310,6 +2730,10 @@ jupyter-server = [ {file = "jupyter_server-1.13.4-py3-none-any.whl", hash = "sha256:3a1df2e27a322e84c028e52272e6ff72fd875f9a74c84409263c5c2f1afbf6fa"}, {file = "jupyter_server-1.13.4.tar.gz", hash = "sha256:5fb5a219385338b1d13a013a68f54688b6a69ecff4e757fd230e27ecacdbf212"}, ] +jupyter-sphinx = [ + {file = "jupyter_sphinx-0.3.2-py3-none-any.whl", hash = "sha256:301e36d0fb3007bb5802f6b65b60c24990eb99c983332a2ab6eecff385207dc9"}, + {file = "jupyter_sphinx-0.3.2.tar.gz", hash = "sha256:37fc9408385c45326ac79ca0452fbd7ae2bf0e97842d626d2844d4830e30aaf2"}, +] jupyterlab = [ {file = "jupyterlab-3.2.9-py3-none-any.whl", hash = "sha256:729d1f06e97733070badc04152aecf9fb2cd036783eebbd9123ff58aab83a8f5"}, {file = "jupyterlab-3.2.9.tar.gz", hash = "sha256:65ddc34e5da1a764606e38c4f70cf9d4ac1c05182813cf0ab2dfea312c701124"}, @@ -2415,6 +2839,10 @@ lazy-object-proxy = [ {file = "lazy_object_proxy-1.7.1-cp39-cp39-win_amd64.whl", hash = "sha256:677ea950bef409b47e51e733283544ac3d660b709cfce7b187f5ace137960d61"}, {file = "lazy_object_proxy-1.7.1-pp37.pp38-none-any.whl", hash = "sha256:d66906d5785da8e0be7360912e99c9188b70f52c422f9fc18223347235691a84"}, ] +markdown-it-py = [ + {file = "markdown-it-py-1.1.0.tar.gz", hash = "sha256:36be6bb3ad987bfdb839f5ba78ddf094552ca38ccbd784ae4f74a4e1419fc6e3"}, + {file = "markdown_it_py-1.1.0-py3-none-any.whl", hash = "sha256:98080fc0bc34c4f2bcf0846a096a9429acbd9d5d8e67ed34026c03c61c464389"}, +] markupsafe = [ {file = "MarkupSafe-2.0.1-cp310-cp310-macosx_10_9_universal2.whl", hash = "sha256:d8446c54dc28c01e5a2dbac5a25f071f6653e6e40f3a8818e8b45d790fe6ef53"}, {file = "MarkupSafe-2.0.1-cp310-cp310-macosx_10_9_x86_64.whl", hash = "sha256:36bc903cbb393720fad60fc28c10de6acf10dc6cc883f3e24ee4012371399a38"}, @@ -2531,6 +2959,10 @@ mccabe = [ {file = "mccabe-0.6.1-py2.py3-none-any.whl", hash = "sha256:ab8a6258860da4b6677da4bd2fe5dc2c659cff31b3ee4f7f5d64e79735b80d42"}, {file = "mccabe-0.6.1.tar.gz", hash = "sha256:dd8d182285a0fe56bace7f45b5e7d1a6ebcbf524e8f3bd87eb0f125271b8831f"}, ] +mdit-py-plugins = [ + {file = "mdit-py-plugins-0.2.8.tar.gz", hash = "sha256:5991cef645502e80a5388ec4fc20885d2313d4871e8b8e320ca2de14ac0c015f"}, + {file = "mdit_py_plugins-0.2.8-py3-none-any.whl", hash = "sha256:1833bf738e038e35d89cb3a07eb0d227ed647ce7dd357579b65343740c6d249c"}, +] mistune = [ {file = "mistune-0.8.4-py2.py3-none-any.whl", hash = "sha256:88a1051873018da288eee8538d476dffe1262495144b33ecb586c4ab266bb8d4"}, {file = "mistune-0.8.4.tar.gz", hash = "sha256:59a3429db53c50b5c6bcc8a07f8848cb00d7dc8bdb431a4ab41920d201d4756e"}, @@ -2561,6 +2993,14 @@ mypy-extensions = [ {file = "mypy_extensions-0.4.3-py2.py3-none-any.whl", hash = "sha256:090fedd75945a69ae91ce1303b5824f428daf5a028d2f6ab8a299250a846f15d"}, {file = "mypy_extensions-0.4.3.tar.gz", hash = "sha256:2d82818f5bb3e369420cb3c4060a7970edba416647068eb4c5343488a6c604a8"}, ] +myst-nb = [ + {file = "myst-nb-0.13.2.tar.gz", hash = "sha256:81e0a4f186bb35c487f5443c7005a474d68ffb58f518f469102d1db7b452066a"}, + {file = "myst_nb-0.13.2-py3-none-any.whl", hash = "sha256:1b9ea3a04c9e0eee05145aa297d2feeabb94c4e23e3047b92efa011ddba4f4b4"}, +] +myst-parser = [ + {file = "myst-parser-0.15.2.tar.gz", hash = "sha256:f7f3b2d62db7655cde658eb5d62b2ec2a4631308137bd8d10f296a40d57bbbeb"}, + {file = "myst_parser-0.15.2-py3-none-any.whl", hash = "sha256:40124b6f27a4c42ac7f06b385e23a9dcd03d84801e9c7130b59b3729a554b1f9"}, +] nbclassic = [ {file = "nbclassic-0.3.5-py3-none-any.whl", hash = "sha256:012d18efb4e24fe9af598add0dcaa621c1f8afbbbabb942fb583dd7fbb247fc8"}, {file = "nbclassic-0.3.5.tar.gz", hash = "sha256:99444dd63103af23c788d9b5172992f12caf8c3098dd5a35c787f0df31490c29"}, @@ -2624,6 +3064,7 @@ numpy = [ {file = "numpy-1.22.2-cp39-cp39-win32.whl", hash = "sha256:8cf33634b60c9cef346663a222d9841d3bbbc0a2f00221d6bcfd0d993d5543f6"}, {file = "numpy-1.22.2-cp39-cp39-win_amd64.whl", hash = "sha256:59153979d60f5bfe9e4c00e401e24dfe0469ef8da6d68247439d3278f30a180f"}, {file = "numpy-1.22.2-pp38-pypy38_pp73-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:4a176959b6e7e00b5a0d6f549a479f869829bfd8150282c590deee6d099bbb6e"}, + {file = "numpy-1.22.2.zip", hash = "sha256:076aee5a3763d41da6bef9565fdf3cb987606f567cd8b104aded2b38b7b47abf"}, ] packaging = [ {file = "packaging-21.3-py3-none-any.whl", hash = "sha256:ef103e05f519cdc783ae24ea4e2e0f508a9c99b2d4969652eed6a2e1ea5bd522"}, @@ -2790,6 +3231,10 @@ pydantic = [ {file = "pydantic-1.9.0-py3-none-any.whl", hash = "sha256:085ca1de245782e9b46cefcf99deecc67d418737a1fd3f6a4f511344b613a5b3"}, {file = "pydantic-1.9.0.tar.gz", hash = "sha256:742645059757a56ecd886faf4ed2441b9c0cd406079c2b4bee51bcc3fbcd510a"}, ] +pydata-sphinx-theme = [ + {file = "pydata_sphinx_theme-0.8.0-py3-none-any.whl", hash = "sha256:fbcbb833a07d3ad8dd997dd40dc94da18d98b41c68123ab0182b58fe92271204"}, + {file = "pydata_sphinx_theme-0.8.0.tar.gz", hash = "sha256:9f72015d9c572ea92e3007ab221a8325767c426783b6b9941813e65fa988dc90"}, +] pyflakes = [ {file = "pyflakes-2.4.0-py2.py3-none-any.whl", hash = "sha256:3bb3a3f256f4b7968c9c788781e4ff07dce46bdf12339dcda61053375426ee2e"}, {file = "pyflakes-2.4.0.tar.gz", hash = "sha256:05a85c2872edf37a4ed30b0cce2f6093e1d0581f8c19d7393122da7e25b2b24c"}, @@ -3087,6 +3532,84 @@ sniffio = [ {file = "sniffio-1.2.0-py3-none-any.whl", hash = "sha256:471b71698eac1c2112a40ce2752bb2f4a4814c22a54a3eed3676bc0f5ca9f663"}, {file = "sniffio-1.2.0.tar.gz", hash = "sha256:c4666eecec1d3f50960c6bdf61ab7bc350648da6c126e3cf6898d8cd4ddcd3de"}, ] +snowballstemmer = [ + {file = "snowballstemmer-2.2.0-py2.py3-none-any.whl", hash = "sha256:c8e1716e83cc398ae16824e5572ae04e0d9fc2c6b985fb0f900f5f0c96ecba1a"}, + {file = "snowballstemmer-2.2.0.tar.gz", hash = "sha256:09b16deb8547d3412ad7b590689584cd0fe25ec8db3be37788be3810cbf19cb1"}, +] +soupsieve = [ + {file = "soupsieve-2.3.1-py3-none-any.whl", hash = "sha256:1a3cca2617c6b38c0343ed661b1fa5de5637f257d4fe22bd9f1338010a1efefb"}, + {file = "soupsieve-2.3.1.tar.gz", hash = "sha256:b8d49b1cd4f037c7082a9683dfa1801aa2597fb11c3a1155b7a5b94829b4f1f9"}, +] +sphinx = [ + {file = "Sphinx-4.4.0-py3-none-any.whl", hash = "sha256:5da895959511473857b6d0200f56865ed62c31e8f82dd338063b84ec022701fe"}, + {file = "Sphinx-4.4.0.tar.gz", hash = "sha256:6caad9786055cb1fa22b4a365c1775816b876f91966481765d7d50e9f0dd35cc"}, +] +sphinx-togglebutton = [ + {file = "sphinx-togglebutton-0.3.0.tar.gz", hash = "sha256:005594ceb82c3da382d7b3a20aa0ceabc79648fc14f85bbef1424d1409112831"}, + {file = "sphinx_togglebutton-0.3.0-py3-none-any.whl", hash = "sha256:a6f37dc04fab6f07154c598973fcb4615171885d90a0e4575f80203c9ddef44a"}, +] +sphinxcontrib-applehelp = [ + {file = "sphinxcontrib-applehelp-1.0.2.tar.gz", hash = "sha256:a072735ec80e7675e3f432fcae8610ecf509c5f1869d17e2eecff44389cdbc58"}, + {file = "sphinxcontrib_applehelp-1.0.2-py2.py3-none-any.whl", hash = "sha256:806111e5e962be97c29ec4c1e7fe277bfd19e9652fb1a4392105b43e01af885a"}, +] +sphinxcontrib-devhelp = [ + {file = "sphinxcontrib-devhelp-1.0.2.tar.gz", hash = "sha256:ff7f1afa7b9642e7060379360a67e9c41e8f3121f2ce9164266f61b9f4b338e4"}, + {file = "sphinxcontrib_devhelp-1.0.2-py2.py3-none-any.whl", hash = "sha256:8165223f9a335cc1af7ffe1ed31d2871f325254c0423bc0c4c7cd1c1e4734a2e"}, +] +sphinxcontrib-htmlhelp = [ + {file = "sphinxcontrib-htmlhelp-2.0.0.tar.gz", hash = "sha256:f5f8bb2d0d629f398bf47d0d69c07bc13b65f75a81ad9e2f71a63d4b7a2f6db2"}, + {file = "sphinxcontrib_htmlhelp-2.0.0-py2.py3-none-any.whl", hash = "sha256:d412243dfb797ae3ec2b59eca0e52dac12e75a241bf0e4eb861e450d06c6ed07"}, +] +sphinxcontrib-jsmath = [ + {file = "sphinxcontrib-jsmath-1.0.1.tar.gz", hash = "sha256:a9925e4a4587247ed2191a22df5f6970656cb8ca2bd6284309578f2153e0c4b8"}, + {file = "sphinxcontrib_jsmath-1.0.1-py2.py3-none-any.whl", hash = "sha256:2ec2eaebfb78f3f2078e73666b1415417a116cc848b72e5172e596c871103178"}, +] +sphinxcontrib-qthelp = [ + {file = "sphinxcontrib-qthelp-1.0.3.tar.gz", hash = "sha256:4c33767ee058b70dba89a6fc5c1892c0d57a54be67ddd3e7875a18d14cba5a72"}, + {file = "sphinxcontrib_qthelp-1.0.3-py2.py3-none-any.whl", hash = "sha256:bd9fc24bcb748a8d51fd4ecaade681350aa63009a347a8c14e637895444dfab6"}, +] +sphinxcontrib-serializinghtml = [ + {file = "sphinxcontrib-serializinghtml-1.1.5.tar.gz", hash = "sha256:aa5f6de5dfdf809ef505c4895e51ef5c9eac17d0f287933eb49ec495280b6952"}, + {file = "sphinxcontrib_serializinghtml-1.1.5-py2.py3-none-any.whl", hash = "sha256:352a9a00ae864471d3a7ead8d7d79f5fc0b57e8b3f95e9867eb9eb28999b92fd"}, +] +sqlalchemy = [ + {file = "SQLAlchemy-1.4.31-cp27-cp27m-macosx_10_14_x86_64.whl", hash = "sha256:c3abc34fed19fdeaead0ced8cf56dd121f08198008c033596aa6aae7cc58f59f"}, + {file = "SQLAlchemy-1.4.31-cp27-cp27m-manylinux_2_5_x86_64.manylinux1_x86_64.whl", hash = "sha256:8d0949b11681380b4a50ac3cd075e4816afe9fa4a8c8ae006c1ca26f0fa40ad8"}, + {file = "SQLAlchemy-1.4.31-cp27-cp27m-win32.whl", hash = "sha256:f3b7ec97e68b68cb1f9ddb82eda17b418f19a034fa8380a0ac04e8fe01532875"}, + {file = "SQLAlchemy-1.4.31-cp27-cp27m-win_amd64.whl", hash = "sha256:81f2dd355b57770fdf292b54f3e0a9823ec27a543f947fa2eb4ec0df44f35f0d"}, + {file = "SQLAlchemy-1.4.31-cp27-cp27mu-manylinux_2_5_x86_64.manylinux1_x86_64.whl", hash = "sha256:4ad31cec8b49fd718470328ad9711f4dc703507d434fd45461096da0a7135ee0"}, + {file = "SQLAlchemy-1.4.31-cp310-cp310-macosx_10_15_x86_64.whl", hash = "sha256:05fa14f279d43df68964ad066f653193187909950aa0163320b728edfc400167"}, + {file = "SQLAlchemy-1.4.31-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:dccff41478050e823271642837b904d5f9bda3f5cf7d371ce163f00a694118d6"}, + {file = "SQLAlchemy-1.4.31-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:57205844f246bab9b666a32f59b046add8995c665d9ecb2b7b837b087df90639"}, + {file = "SQLAlchemy-1.4.31-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:ea8210090a816d48a4291a47462bac750e3bc5c2442e6d64f7b8137a7c3f9ac5"}, + {file = "SQLAlchemy-1.4.31-cp310-cp310-win32.whl", hash = "sha256:2e216c13ecc7fcdcbb86bb3225425b3ed338e43a8810c7089ddb472676124b9b"}, + {file = "SQLAlchemy-1.4.31-cp310-cp310-win_amd64.whl", hash = "sha256:e3a86b59b6227ef72ffc10d4b23f0fe994bef64d4667eab4fb8cd43de4223bec"}, + {file = "SQLAlchemy-1.4.31-cp36-cp36m-macosx_10_14_x86_64.whl", hash = "sha256:2fd4d3ca64c41dae31228b80556ab55b6489275fb204827f6560b65f95692cf3"}, + {file = "SQLAlchemy-1.4.31-cp36-cp36m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:6f22c040d196f841168b1456e77c30a18a3dc16b336ddbc5a24ce01ab4e95ae0"}, + {file = "SQLAlchemy-1.4.31-cp36-cp36m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:c0c7171aa5a57e522a04a31b84798b6c926234cb559c0939840c3235cf068813"}, + {file = "SQLAlchemy-1.4.31-cp36-cp36m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d046a9aeba9bc53e88a41e58beb72b6205abb9a20f6c136161adf9128e589db5"}, + {file = "SQLAlchemy-1.4.31-cp36-cp36m-win32.whl", hash = "sha256:d86132922531f0dc5a4f424c7580a472a924dd737602638e704841c9cb24aea2"}, + {file = "SQLAlchemy-1.4.31-cp36-cp36m-win_amd64.whl", hash = "sha256:ca68c52e3cae491ace2bf39b35fef4ce26c192fd70b4cd90f040d419f70893b5"}, + {file = "SQLAlchemy-1.4.31-cp37-cp37m-macosx_10_14_x86_64.whl", hash = "sha256:cf2cd387409b12d0a8b801610d6336ee7d24043b6dd965950eaec09b73e7262f"}, + {file = "SQLAlchemy-1.4.31-cp37-cp37m-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:bb4b15fb1f0aafa65cbdc62d3c2078bea1ceecbfccc9a1f23a2113c9ac1191fa"}, + {file = "SQLAlchemy-1.4.31-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:c317ddd7c586af350a6aef22b891e84b16bff1a27886ed5b30f15c1ed59caeaa"}, + {file = "SQLAlchemy-1.4.31-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:3c7ed6c69debaf6198fadb1c16ae1253a29a7670bbf0646f92582eb465a0b999"}, + {file = "SQLAlchemy-1.4.31-cp37-cp37m-win32.whl", hash = "sha256:6a01ec49ca54ce03bc14e10de55dfc64187a2194b3b0e5ac0fdbe9b24767e79e"}, + {file = "SQLAlchemy-1.4.31-cp37-cp37m-win_amd64.whl", hash = "sha256:330eb45395874cc7787214fdd4489e2afb931bc49e0a7a8f9cd56d6e9c5b1639"}, + {file = "SQLAlchemy-1.4.31-cp38-cp38-macosx_10_14_x86_64.whl", hash = "sha256:5e9c7b3567edbc2183607f7d9f3e7e89355b8f8984eec4d2cd1e1513c8f7b43f"}, + {file = "SQLAlchemy-1.4.31-cp38-cp38-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:de85c26a5a1c72e695ab0454e92f60213b4459b8d7c502e0be7a6369690eeb1a"}, + {file = "SQLAlchemy-1.4.31-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:975f5c0793892c634c4920057da0de3a48bbbbd0a5c86f5fcf2f2fedf41b76da"}, + {file = "SQLAlchemy-1.4.31-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:d5c20c8415173b119762b6110af64448adccd4d11f273fb9f718a9865b88a99c"}, + {file = "SQLAlchemy-1.4.31-cp38-cp38-win32.whl", hash = "sha256:b35dca159c1c9fa8a5f9005e42133eed82705bf8e243da371a5e5826440e65ca"}, + {file = "SQLAlchemy-1.4.31-cp38-cp38-win_amd64.whl", hash = "sha256:b7b20c88873675903d6438d8b33fba027997193e274b9367421e610d9da76c08"}, + {file = "SQLAlchemy-1.4.31-cp39-cp39-macosx_10_15_x86_64.whl", hash = "sha256:85e4c244e1de056d48dae466e9baf9437980c19fcde493e0db1a0a986e6d75b4"}, + {file = "SQLAlchemy-1.4.31-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:e79e73d5ee24196d3057340e356e6254af4d10e1fc22d3207ea8342fc5ffb977"}, + {file = "SQLAlchemy-1.4.31-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl", hash = "sha256:15a03261aa1e68f208e71ae3cd845b00063d242cbf8c87348a0c2c0fc6e1f2ac"}, + {file = "SQLAlchemy-1.4.31-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:0ddc5e5ccc0160e7ad190e5c61eb57560f38559e22586955f205e537cda26034"}, + {file = "SQLAlchemy-1.4.31-cp39-cp39-win32.whl", hash = "sha256:289465162b1fa1e7a982f8abe59d26a8331211cad4942e8031d2b7db1f75e649"}, + {file = "SQLAlchemy-1.4.31-cp39-cp39-win_amd64.whl", hash = "sha256:9e4fb2895b83993831ba2401b6404de953fdbfa9d7d4fa6a4756294a83bbc94f"}, + {file = "SQLAlchemy-1.4.31.tar.gz", hash = "sha256:582b59d1e5780a447aada22b461e50b404a9dc05768da1d87368ad8190468418"}, +] stack-data = [ {file = "stack_data-0.1.4-py3-none-any.whl", hash = "sha256:02cc0683cbc445ae4ca8c4e3a0e58cb1df59f252efb0aa016b34804a707cf9bc"}, {file = "stack_data-0.1.4.tar.gz", hash = "sha256:7769ed2482ce0030e00175dd1bf4ef1e873603b6ab61cd3da443b410e64e9477"}, diff --git a/pyproject.toml b/pyproject.toml index eb081dd..acc28b2 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,6 +1,6 @@ [tool.poetry] name = "ipyannotator" -version = "0.7.0" +version = "0.8.0" description = "The infinitely hackable annotation framework" authors = ["palaimon.io "] license = "Apache License 2.0" @@ -42,6 +42,9 @@ nbqa = "^1.2.3" pylint = "^2.12.2" mypy = "^0.931" autopep8 = "^1.6.0" +myst-nb = "^0.13.2" +Sphinx = "^4.4.0" +pydata-sphinx-theme = "^0.8.0" [build-system] diff --git a/settings.ini b/settings.ini index 35b7870..ae1b7ef 100644 --- a/settings.ini +++ b/settings.ini @@ -9,7 +9,7 @@ author = Palaimon author_email = oss@mail.palaimon.io copyright = Palaimon GmbH branch = master -version = 0.7.0 +version = 0.8.0 min_python = 3.7 audience = Developers language = English diff --git a/voila.Dockerfile b/voila.Dockerfile index 4532499..3b8e6c5 100644 --- a/voila.Dockerfile +++ b/voila.Dockerfile @@ -41,4 +41,4 @@ EXPOSE 8080 ENTRYPOINT ["poetry", "run", "voila", "--enable_nbextensions=True", "--no-browser", "--port=8080"] -CMD ["nbs/09_viola_example.ipynb"] \ No newline at end of file +CMD ["nbs/09_voila_example.ipynb"] \ No newline at end of file