CHANGELOG.md

Change Log

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog and this project adheres to Semantic Versioning.

Unreleased

Added

• ReSpecTh 2.x support in ReSpecTh_to_ChemKED: the reference is read from the bibliographyLink child elements (referenceDOI, details) with no DOI lookup, falling back to the 1.x doi and preferredKey attributes
• Conversion of laminar burning velocity and speciation measurements, alongside ignition delay. ReSpecTh's four speciation experiment types all map to ChemKED speciation measurement
• Mapping tables in converters.py for ReSpecTh experiment types, apparatus kinds and modes, common property names, speciation independent variables, and ignition targets and types
• apparatus_kind argument to ReSpecTh_to_ChemKED and -ak/--apparatus-kind to respth2ck, for files whose apparatus ReSpecTh records only as the generic flame
• uncertainty and evaluated standard deviation properties and dataGroup columns are converted and attached to the quantity they describe, whether that is a datapoint value, a common property, a species amount, or a concentration profile
• ReSpecTh comment elements are converted to ChemKED comments
• isvalid_experiment validation rule cross-checking experiment-type against the datapoint contents and the apparatus kind, which the anyof datapoints schema cannot do on its own
• Round-trip tests asserting the converter reproduces each checked-in ReSpecTh/ChemKED fixture pair
• resolve_reference maps an uncertainty that points at a column by its label or id, and one that references initial composition, onto the quantity it describes; both were silently dropped before
• 0.5.0 added to the allowed chemked-version values, and a test asserting the current __version__ is always among them, since get_file_metadata stamps it into every converted file

Changed

• The tubular reactor apparatus kind is renamed flow reactor, the name ReSpecTh, the ChemKED database, and the speciation literature all use. tubular reactor is still accepted as converter input and maps to flow reactor
• The mole and mass fraction sum check uses a relative tolerance of 1e-3 (composition_sum_tolerance) rather than the NumPy default of 1e-5, so compositions reported to a few digits are not rejected for round-off
• An auxiliary profile against a residence-time axis is recorded with the axis name time, which auxiliary-profiles.independent.name allows, since a residence time is a time coordinate

Fixed

• A composition amount of exactly zero, meaning a species measured as absent, is accepted when it carries uncertainty metadata. It was accepted without metadata and rejected with it, and isvalid_composition already allowed it. Only a negative value is an error for a property with no configured units
• The unsupported-composition-units error lists the units the converter actually accepts, built from composition_units rather than hand-written, so it cannot fall behind the table again
• Every text file is now opened with an explicit UTF-8 encoding. Reading or writing a ChemKED or ReSpecTh file whose author names contain non-ASCII characters previously depended on the platform's preferred encoding, and failed on Windows. convert_to_ReSpecTh could also rewrite its pretty-printed XML in the platform encoding while declaring UTF-8 in the file. The ruff rule PLW1514 now enforces this
• Relative uncertainties and relative evaluated standard deviations are now checked to be dimensionless. They previously bypassed unit validation entirely, so uncertainty-type: relative with uncertainty: 0.1 kelvin validated and was then misread at load time
• An uncertainty value Pint cannot parse is reported as a validation error instead of raising UndefinedUnitError out of the validator
• Unit strings that ReSpecTh writes with bare negative exponents, such as ms-1 and kg m-2 s-1, are normalized for Pint instead of escaping as a DimensionalityError from outside the guarded block
• get_ignition_type accepted only 5 of the 10 ignition types and 6 of the 18 ignition targets the schema allows; the vocabulary is now in step with ignition_delay_schema.yaml and is checked by a test
• experimentType is matched case-insensitively, so the lowercase spelling used throughout ReSpecTh 2.x is recognized
• ignitionType is only required for ignition delay files
• composition_schema.yaml now references the shared uncertainty-metadata definition instead of a hand-copied duplicate that had drifted to type: float, so a species amount accepts the same uncertainty forms as every other quantity. The definition is hoisted to its own top-level anchor in value_unit_schema.yaml
• Composition units accept concentrations (mol/cm3, mol/m3, mol/dm3, mol/L), the misspelling mole faction found in part of the ReSpecTh corpus, and a mixture of mole percent and mole fraction, which is reconciled onto mole fraction

0.5.0 - 2026-07-02

Added

• Add codemeta file
• GitHub Actions CI workflow testing on Python 3.10, 3.11, 3.12, 3.13, and 3.14 on Linux, and Python 3.14 on macOS and Windows
• GitHub Actions workflow for deploying documentation to GitHub Pages, and separate workflow for publishing to PyPI on version tags using Trusted Publishers
• Updates dependency groups: pytest-cov in test, ruff in lint
• Adds ruff, mypy, and pre-commit-hooks to .pre-commit-config.yaml
• httpx2 as a direct dependency for HTTP calls in orcid.py
• pytest testpaths, filterwarnings, and coverage configuration in pyproject.toml
• live_api pytest marker and tests/test_live_api.py with contract tests that verify mock data still matches real Crossref and ORCID API responses
• GitHub Actions workflow (.github/workflows/live-api.yml) running live API contract tests on a weekly schedule and on manual dispatch
• tests/_mock_data.py centralising shared Crossref and ORCID mock data for use by both conftest.py and test_live_api.py
• mock_orcid_api and mock_all_apis fixtures in tests/conftest.py
• Development and testing documentation in docs/development.rst and CONTRIBUTING.md

Changed

• replaces os.path with pathlib.Path operations
• replaces codemeta.json file with CITATION.cff
• All tests that previously called external APIs directly now use mock fixtures; addopts = "-m 'not live_api'" in pyproject.toml excludes live API tests from the default run
• switched to Coveralls for code coverage
• Directly use the Markdown formatting of the README on pypi, rather than converting to reST
• Composition type is included in the pandas data-frame resulting from to_dataframe()
• Migrated from setup.py/setup.cfg to pyproject.toml with hatchling build backend
• pandas moved from optional dataframes extra into main dependencies
• Moved source to src/pyked/ layout; tests moved to top-level tests/ directory
• orcid.py now uses httpx2 instead of requests for HTTP calls, and updated to use ORCID public API v3.0
• Updated conda recipe to use load_file_regex for version, updated all dependency pins, require Python >= 3.10
• Updated README badges; added GitHub Actions CI badge; updated Codecov to main branch
• Updated docs/conf.py: replaced deprecated pkg_resources with importlib.metadata; replaced deprecated autodoc_default_flags with autodoc_default_options; updated intersphinx mappings; fixed language = None deprecation; removed legacy Travis CI environment check
• Updated paths and URLs in example notebooks and ck-tutorial.rst

Fixed

• filter in chemked.py incompatible with Python 3.14; replaced with list comprehension
• ORCID URL stripping in converters.py used lstrip() (strips characters, not a prefix) causing malformed ORCIDs when Crossref returns https:// URLs; replaced with rfind('/') approach consistent with validation.py
• Test assertions for Crossref author names updated to match current API response format
• Test path assertions for ReSpecTh conversion detail string corrected after test directory migration
• pandas.util.testing (removed in pandas 1.0) replaced with pandas.testing in test fixtures
• Removed no_internet() helper and internet_missing skip marker from test_validation.py; tests now mock APIs instead of skipping when offline
• Duplicate compression_time entry in DataPoint docstring causing Sphinx build failure

Removed

• appveyor.yml replaced by GitHub Actions
• requirements.txt replaced by pyproject.toml dependencies
• MANIFEST.in not needed with hatchling build backend

0.4.1 - 2018-03-09

Added

• Documentation for old versions is available on the Releases page of the docs

Changed

Fixed

• Doctr deploys on tags now
• Syntax changes for example files in the documentation

0.4.0 - 2018-03-07

Added

• New method to instantiate a ChemKED class directly from a ReSpecTh XML file
• The __version__ attribute can be imported from the top-level module
• New time-histories field to replace the volume-history. This field allows specification of several other relevant parameters besides volume.
• Added rcm-data field and moved compressed-temperature, compressed-pressure, and compression-time to this field
• Added stroke, clearance, and compression-ratio to the rcm-data field
• Added conda-forge instructions to the installation documentation
• Allow alpha versions to be specified during testing

Changed

• Crossref lookups via Habanero now comply with the "be-nice" policy
• Removed UnboundLocalError from error processing for reference validation
• Switch to flake8 for style checking in CI services
• file-author field is now a list called file-authors
• ReSpecTh->ChemKED converter function now returns a dictionary, while the command-line entry points write out files
• Require Habanero>=0.6.0 to support the mailto argument
• Require pytest>=3.2.0 to support the pytest.mark.filterwarnings decorator
• Deprecate the volume-history field in the ChemKED YAML file and replace with time-histories
• ORCID lookups are now done by a function in the local orcid.py module, removing an external dependency
• Composition in a DataPoint is now stored in a dictionary of namedtuples (called Composition) rather than a list of dictionaries

Fixed

• Crossref lookups in the converters use the common API instance from validation
• d/dt max extrapolated ignition type can be converted to/from ReSpecTh
• Tests now check for appropriate warnings and ignore unrelated warnings

0.3.0 - 2017-10-09

Added

• New extrapolated ignition type, where the maximum slope is extrapolated to the baseline
• Tests that the composition type is stored properly in the DataPoint
• species_conversion dictionary can be passed to the get_cantera_mole_fraction and get_cantera_mass_fraction functions to change the name of a species in the output string
• Jupyter Notebook examples of usage

Removed

• Removes elemental-composition as a synonym for atomic-composition

Fixed

• Fixes test_incorrect_doi_period_at_end docstring

Changed

• Conda builds are now noarch - one package for all Pythons!
• pip installs now require Python compatible with 3.5
• Appveyor runs a single job and no longer builds conda packages
• Remove journal from required fields in the reference

0.2.1 - 2017-08-31

Fixed

• Fixes Cantera convenience output functions

0.2.0 - 2017-08-10

Added

• Adds ChemKED method to write new file, with tests
• Adds converters to and from ReSpecTh files, with tests
• Adds command-line entry points for converter scripts
• Add docs for converters

Fixed

• ignition_type dictionary in DataPoint is now deepcopyd

0.1.6 - 2017-07-17

Added

• Added logo files to repo
• Added first_stage_ignition_delay, compressed_pressure, and compressed_temperature as properties

Changed

• Added Zenodo collection DOI to CITATION.md

0.1.5 - 2017-05-22

Added

• Schema can now be split into multiple files via !include directive

Fixed

• Remove Python 2.7 classifier from setup.py
• DataFrame output for datapoints lists with multiple compositions (i.e., a species not in all compositions)

Changed

• Improved tests with no internet
• Improved tests with no warning

0.1.4 - 2017-04-21

Added

• Add skip_validation keyword argument to the ChemKED initializer

Removed

• Python 2.7 support is removed again

0.1.3 - 2017-04-13

Added

• Add back Python 2.7 support
• Add Appveyor builds for Windows conda packages

0.1.2 - 2017-04-13

Added

• Tests of the composition uncertainty in the DataPoint
• Tests of the values in the references
• Packaging for conda and PyPI
• Add Anaconda-Server badge to README

Changed

• All fixed DOIs in CITATION.md are now specified with placeholders

0.1.1 - 2017-04-02

Added

• Added Zenodo DOI badge to README
• Added CITATION file, and mention of license to README

Fixed

• Fixed chemked-version bug in schema introduced in 0.1.0

0.1.0 - 2017-04-02

Added

• First minor release of PyKED, supporting autoignition experiments.
• Basic API documentation is available via https://pr-omethe-us.github.io/PyKED/