{bio}[GCCcore/13.3.0,foss/2024a] Cython v3.2.4, learnMSA v3.0.0, logomaker v0.8.7, ...

easybuild/easyconfigs/c/Cython/Cython-3.2.4-GCCcore-13.3.0.eb

@@ -0,0 +1,36 @@
+easyblock = 'PythonPackage'
+
+name = 'Cython'
+version = '3.2.4'
+
+homepage = 'https://cython.org/'
+description = """
+Cython is an optimising static compiler for both the Python programming
+language and the extended Cython programming language (based on Pyrex).
+"""
+docurls = [
+    'https://cython.org/#documentation',
+    'https://github.com/cython/cython',
+]
+
+toolchain = {'name': 'GCCcore', 'version': '13.3.0'}
+
+sources = [SOURCELOWER_TAR_GZ]
+checksums = ['84226ecd313b233da27dc2eb3601b4f222b8209c3a7216d8733b031da1dc64e6']
+
+builddependencies = [
+    ('binutils', '2.42'),
+]
+
+dependencies = [
+    ('Python', '3.12.3'),
+]
+
+sanity_check_paths = {
+    'files': ['bin/cygdb', 'bin/cython', 'bin/cythonize'],
+    'dirs': ['lib/python%(pyshortver)s/site-packages'],
+}
+
+sanity_check_commands = ["cython --version"]
+
+moduleclass = 'lang'

easybuild/easyconfigs/l/learnMSA/evoten-0.0.1_pyproject.patch

@@ -0,0 +1,23 @@
+Change to how to deal with the licence, based on PR #24919
+Author: J. Saßmannshausen (Imperial College London/UK)
+diff --git a/evoten-0.0.1.orig/pyproject.toml b/evoten-0.0.1/pyproject.toml
+index 531fd01..0ab9c2f 100644
+--- a/evoten-0.0.1.orig/pyproject.toml
++++ b/evoten-0.0.1/pyproject.toml
+@@ -15,8 +15,7 @@ classifiers = [
+     "Programming Language :: Python :: 3",
+     "Operating System :: OS Independent",
+ ]
+-license = "MIT"
+-license-files = ["LICEN[CS]E*"]
++license = { file = "LICENSE" }
+ dependencies = ["numpy", "biopython"]
+
+ [project.optional-dependencies]
+@@ -41,4 +40,4 @@ Homepage = "https://github.com/felbecker/evoten"
+ Issues = "https://github.com/felbecker/evoten/issues"
+
+ [tool.hatch.build.targets.wheel]
+-packages = ["evoten"]
+\ No newline at end of file
++packages = ["evoten"]

easybuild/easyconfigs/l/learnMSA/hidten-0.0.2_pyproject.patch

@@ -0,0 +1,23 @@
+Change to how to deal with the licence, based on PR #24919
+Author: J. Saßmannshausen (Imperial College London/UK)
+diff --git a/hidten-0.0.2.orig/pyproject.toml b/hidten-0.0.2/pyproject.toml
+index a750862..4b82c1e 100644
+--- a/hidten-0.0.2.orig/pyproject.toml
++++ b/hidten-0.0.2/pyproject.toml
+@@ -16,8 +16,7 @@ classifiers = [
+     "Programming Language :: Python :: 3",
+     "Operating System :: OS Independent",
+ ]
+-license = "MIT"
+-license-files = ["LICENSE"]
++license = { file = "LICENSE" }
+ dependencies = ["numpy", "pydantic"]
+
+ [project.optional-dependencies]
+@@ -35,4 +34,4 @@ Homepage = "https://github.com/gaius-augustus/hidten"
+ Issues = "https://github.com/gaius-augustus/hidten/issues"
+
+ [tool.setuptools.packages.find]
+-exclude = ["docs*"]
+\ No newline at end of file
++exclude = ["docs*"]

easybuild/easyconfigs/l/learnMSA/learnMSA-3.0.0-foss-2024a.eb

@@ -0,0 +1,61 @@
+# This easyconfig was created by the BEAR Software team at the University of Birmingham.
+# Updated to latest version
+# Author: J. Sassmannshausen (Imperial College London/UK)
+
+easyblock = 'PythonBundle'
+
+name = 'learnMSA'
+version = '3.0.0'
+
+homepage = "https://github.com/Gaius-Augustus/learnMSA"
+description = """Learning and Aligning Large Protein Families with support of protein language models."""
+citing = """Becker F, Stanke M. learnMSA2: deep protein multiple alignments with large
+ language and hidden Markov models. Bioinformatics. 2024\n
+Becker F, Stanke M. learnMSA: learning and aligning large protein families. GigaScience. 2022"""
+
+toolchain = {'name': 'foss', 'version': '2024a'}
+
+builddependencies = [
+    ('hatchling', '1.24.2'),
+]
+
+dependencies = [
+    ('Python', '3.12.3'),
+    ('Biopython', '1.84'),
+    ('Seaborn', '0.13.2'),
+    ('networkx', '3.4.2'),
+    ('imageio', '2.36.1'),
+    ('pyfamsa', '0.7.0'),
+    ('TensorFlow', '2.18.1'),
+    ('logomaker', '0.8.7'),
+    ('SentencePiece', '0.2.1'),
+    ('Transformers', '4.55.0'),
+    ('pydantic', '2.9.1'),
+]
+
+exts_list = [
+    ('hidten', '0.0.2', {
+        'patches': ['%(name)s-%(version)s_pyproject.patch'],
+        'checksums': [
+            {'hidten-0.0.2.tar.gz': '1e488d9ea7f809636d2bec0092437b3b8202532915003488e8d8e6de666aadfb'},
+            {'hidten-0.0.2_pyproject.patch': 'f7cc4607d3baf39aff03511181c975aae64b00bafa0ec4106ae1db0413dc7485'},
+        ],
+    }),
+    ('evoten', '0.0.1', {
+        'patches': ['%(name)s-%(version)s_pyproject.patch'],
+        'checksums': [
+            {'evoten-0.0.1.tar.gz': '44b05aad42e70d145d2690a15b544f6c233a6b3bd5bf2d3df9927d3d8f5faef9'},
+            {'evoten-0.0.1_pyproject.patch': 'ef40b3c58cb76eb49b6e6b4877922dee1a8dc0b13a3016c65447b4246fa9fd8d'},
+        ],
+    }),
+    (name, version, {
+        'github_account': 'Gaius-Augustus',
+        'modulename': 'learnMSA',
+        'preinstallopts': "sed -i '/license-files/d' pyproject.toml && ",
+        'source_urls': ['https://github.com/%(github_account)s/%(name)s/archive'],
+        'sources': ['v%(version)s.tar.gz'],
+        'checksums': ['520e27b4feeba4879c65b8a58585f3731e2d562ad6f83ca735fc6a7b34719234'],
+    }),
+]
+
+moduleclass = 'bio'

easybuild/easyconfigs/l/logomaker/logomaker-0.8.7-gfbf-2024a.eb

@@ -0,0 +1,25 @@
+# This easyconfig was created by the BEAR Software team at the University of Birmingham.
+
+easyblock = 'PythonPackage'
+
+name = 'logomaker'
+version = '0.8.7'
+
+homepage = "https://github.com/jbkinney/logomaker"
+description = """Software for the visualization of sequence-function relationships."""
+docurls = 'https://logomaker.readthedocs.io/'
+
+toolchain = {'name': 'gfbf', 'version': '2024a'}
+
+sources = [SOURCE_TAR_GZ]
+checksums = ['63783ce6e24449d6f1f01ce29c4ae1b91f6e54bc198e5da40ad18fcd0efc3302']
+
+builddependencies = [('hatchling', '1.27.0')]
+
+dependencies = [
+    ('Python', '3.12.3'),
+    ('SciPy-bundle', '2024.05'),
+    ('matplotlib', '3.9.2'),
+]
+
+moduleclass = 'vis'

easybuild/easyconfigs/p/pyfamsa/pyfamsa-0.7.0-GCCcore-13.3.0.eb

@@ -0,0 +1,36 @@
+# This easyconfig was created by the BEAR Software team at the University of Birmingham.
+# Updated to latest version, Cython moved to builddependency
+# Author: J. Sassmannshausen (Imperial College London/UK)
+
+easyblock = 'PythonPackage'
+
+name = 'pyfamsa'
+version = "0.7.0"
+
+homepage = "https://github.com/althonos/pyfamsa"
+description = """Cython bindings and Python interface to FAMSA,
+an algorithm for ultra-scale multiple sequence alignments.
+"""
+docurls = 'https://pyfamsa.readthedocs.io/en/stable/'
+citing = """Deorowicz, Sebastian, Debudaj-Grabysz, Agnieszka & Gudyś, Adam.
+ ‘FAMSA: Fast and accurate multiple sequence alignment of huge protein families’.
+  Sci Rep 6, 33964 (2016). doi:10.1038/srep33964"""
+
+toolchain = {'name': 'GCCcore', 'version': '13.3.0'}
+
+sources = [SOURCE_TAR_GZ]
+checksums = ['5b1f3b60a319d5a5c92911cccec7ad7d506799aecac0926fecba4da370cf9b3d']
+
+builddependencies = [
+    ('scikit-build-core', '0.11.1'),
+    ('Python-bundle-PyPI', '2024.06'),  # required for pathspec & semantic-version /will fail if runtime
+    ('Cython', '3.2.4'),  # version required to build
+    ('binutils', '2.42'),
+]
+dependencies = [
+    ('Python', '3.12.3'),
+    ('scoring-matrices', '0.3.4'),
+]
+
+
+moduleclass = 'bio'

easybuild/easyconfigs/s/scoring-matrices/scoring-matrices-0.3.4-GCCcore-13.3.0.eb

@@ -0,0 +1,32 @@
+# This easyconfig was created by the BEAR Software team at the University of Birmingham.
+# Updated to newer toolchain, Cython moved to builddependency
+# Author: J. Sassmannshausen (Imperial College London/UK)
+
+easyblock = 'PythonPackage'
+
+name = 'scoring-matrices'
+version = '0.3.4'
+
+homepage = "https://github.com/althonos/scoring-matrices"
+description = """Dependency free, Cython-compatible scoring matrices to use with biological sequences.
+ Scoring Matrices are matrices used to score the matches and mismatches between two characters are the same position in
+ a sequence alignment. Some of these matrices are derived from substitution matrices, which uses evolutionary modeling.
+"""
+docurls = 'https://scoring-matrices.readthedocs.io/en/stable/'
+
+toolchain = {'name': 'GCCcore', 'version': '13.3.0'}
+
+sources = [f'{"_".join(name.split("-"))}-%(version)s.tar.gz']
+checksums = ['b7c97ca2ac8c11143a7d0c278916d9bc78e6ea953fee8f40692aba14253646f4']
+
+builddependencies = [
+    ('scikit-build-core', '0.11.1'),
+    ('hatchling', '1.27.0'),
+    ('Cython', '3.2.4'),  # version required to build
+    ('binutils', '2.42'),
+]
+dependencies = [
+    ('Python', '3.12.3'),
+]
+
+moduleclass = 'bio'

easybuild/easyconfigs/t/Tiberius/Tiberius-1.1.8-foss-2024a.eb

@@ -0,0 +1,45 @@
+# This easyconfig was created by the BEAR Software team at the University of Birmingham.
+# Updated to version 1.1.8 by J. Saßmannshauen (Imperial College London/UK)
+
+easyblock = 'PythonPackage'
+
+name = 'Tiberius'
+version = '1.1.8'
+
+homepage = "https://github.com/Gaius-Augustus/Tiberius"
+description = """Tiberius is a deep learning-based ab initio gene structure prediction tool that end-to-end integrates
+convolutional and long short-term memory layers with a differentiable HMM layer.
+It can be used to predict gene structures from genomic sequences only,
+while matching the accuracy of tool that use extrinsic evidence."""
+
+docurls = 'https://github.com/Gaius-Augustus/Tiberius/blob/main/docs/training_large_data.md'
+
+citing = """Gabriel, Lars, et al. "Tiberius: End-to-End Deep Learning with an HMM for Gene Prediction." 2024,
+https://academic.oup.com/bioinformatics/article/40/12/btae685/7903281
+ https://www.biorxiv.org/content/early/2024/07/23/2024.07.21.604459
+"""
+
+toolchain = {'name': 'foss', 'version': '2024a'}
+
+github_account = 'Gaius-Augustus'
+source_urls = [GITHUB_SOURCE]
+sources = ['v%(version)s.tar.gz']
+patches = ['%(name)s-%(version)s_pyproject.patch']
+checksums = [
+    {'v1.1.8.tar.gz': '31b350863af4ed8ef5802b2d7ee70aebe7e742666efe76b06f3f948e37916a51'},
+    {'Tiberius-1.1.8_pyproject.patch': '098177d8275de2b0bce62cc65009d7b6c84ff49697a80bf4b5605414be911896'},
+]
+
+builddependencies = [('hatchling', '1.24.2')]
+
+dependencies = [
+    ('Python', '3.12.3'),
+    ('learnMSA', '3.0.0'),
+    ('SciPy-bundle', '2024.05'),
+    ('Python-bundle-PyPI', '2024.06'),
+    ('Biopython', '1.84'),
+    ('pyBigWig', '0.3.24'),
+    ('bcbio-gff', '0.7.1'),
+]
+
+moduleclass = 'bio'

easybuild/easyconfigs/t/Tiberius/Tiberius-1.1.8_pyproject.patch

@@ -0,0 +1,16 @@
+Change to how to deal with the licence, based on PR #24919
+Author: J. Saßmannshausen (Imperial College London/UK)
+diff --git a/Tiberius-1.1.8.orig/pyproject.toml b/Tiberius-1.1.8/pyproject.toml
+index d87b27c..5060778 100644
+--- a/Tiberius-1.1.8.orig/pyproject.toml
++++ b/Tiberius-1.1.8/pyproject.toml
+@@ -19,8 +19,7 @@ classifiers = [
+     "Programming Language :: Python :: 3",
+     "Operating System :: OS Independent",
+ ]
+-license = "MIT"
+-license-files = ["LICEN[CS]E*"]
++license = { file = "LICENSE" }
+ dependencies = [
+   "rich",
+   "pyyaml"

easybuild/easyconfigs/t/Tiberius/Tiberius-2.0.6-foss-2024a.eb

@@ -0,0 +1,46 @@
+# This easyconfig was created by the BEAR Software team at the University of Birmingham.
+# Updated to version 2.0.6 by J. Saßmannshauen (Imperial College London/UK)
+
+easyblock = 'PythonPackage'
+
+name = 'Tiberius'
+version = '2.0.6'
+
+homepage = "https://github.com/Gaius-Augustus/Tiberius"
+description = """Tiberius is a deep learning-based ab initio gene structure prediction tool that end-to-end integrates
+convolutional and long short-term memory layers with a differentiable HMM layer.
+It can be used to predict gene structures from genomic sequences only,
+while matching the accuracy of tool that use extrinsic evidence."""
+
+docurls = 'https://github.com/Gaius-Augustus/Tiberius/blob/main/docs/training_large_data.md'
+
+citing = """Gabriel, Lars, et al. "Tiberius: End-to-End Deep Learning with an HMM for Gene Prediction." 2024,
+https://academic.oup.com/bioinformatics/article/40/12/btae685/7903281
+ https://www.biorxiv.org/content/early/2024/07/23/2024.07.21.604459
+"""
+
+toolchain = {'name': 'foss', 'version': '2024a'}
+
+github_account = 'Gaius-Augustus'
+source_urls = [GITHUB_SOURCE]
+sources = ['v%(version)s.tar.gz']
+patches = ['%(name)s-1.1.8_pyproject.patch']
+checksums = [
+    {'v2.0.6.tar.gz': '7124ad6f4361be54677db19b5941204feed3d2117382b601ec243ef20a77fbde'},
+    {'Tiberius-1.1.8_pyproject.patch': '098177d8275de2b0bce62cc65009d7b6c84ff49697a80bf4b5605414be911896'},
+]
+
+builddependencies = [('hatchling', '1.24.2')]
+
+dependencies = [
+    ('Python', '3.12.3'),
+    ('learnMSA', '3.0.0'),
+    ('SciPy-bundle', '2024.05'),
+    ('Python-bundle-PyPI', '2024.06'),
+    ('Biopython', '1.84'),
+    ('pyBigWig', '0.3.24'),
+    ('bcbio-gff', '0.7.1'),
+]
+
+
+moduleclass = 'bio'