progress with scenicplus (pyfasta)

146_clair3/Clair3-remove_htslib.patch

@@ -1,3 +1,16 @@
+diff -ruN Clair3-1.0.4_orig/build.py Clair3-1.0.4/build.py
+--- Clair3-1.0.4_orig/build.py	2023-09-07 14:28:29.842168186 +0100
++++ Clair3-1.0.4/build.py	2023-09-08 12:31:06.300472155 +0100
+@@ -52,8 +52,7 @@
+             'clair3_pileup.c',
+             'clair3_full_alignment.c')],
+     extra_compile_args=extra_compile_args,
+-    extra_link_args=extra_link_args,
+-    extra_objects=[os.path.join(htslib_dir, 'libhts.a')]
++    extra_link_args=extra_link_args
+ )
+
+ cdef = [
 diff -ruN Clair3-1.0.4_orig/Makefile Clair3-1.0.4/Makefile
 --- Clair3-1.0.4_orig/Makefile	2023-09-07 14:28:29.842168186 +0100
 +++ Clair3-1.0.4/Makefile	2023-09-07 14:35:52.320773700 +0100

146_clair3/clair3.eb

@@ -33,14 +33,12 @@ dependencies = [
     ('cURL', '7.83.0'),
     ('zstd', '1.5.2'),
     ('HTSlib', '1.15.1'),
+    ('libdeflate', '1.10'),
 ]
 
 files_to_copy = [
-    (['angsd', 'misc/supersim', 'misc/thetaStat', 'misc/realSFS', 'misc/msToGlf',
-      'misc/smartCount', 'misc/printIcounts', 'misc/contamination', 'misc/splitgl',
-      'misc/NGSadmix', 'misc/contamination2', 'misc/haploToPlink', 'misc/ngsPSMC',
-      'misc/ibs'], 'bin'),
-    'doc',
+    (['run_clair3.sh', 'longphase', 'clair3.py', 'clair3'], 'bin'),
+    (['libclair3.so', 'libclair3.o', 'libclair3.cpython-310-x86_64-linux-gnu.so', 'libclair3.c', 'longphase', ], 'lib'),
 ]
 
 moduleclass = 'bio'

64_scenicplus/adjust_text.eb

@@ -0,0 +1,27 @@
+easyblock = 'PythonPackage'
+
+name = 'adjustText'
+version = '0.7.3'
+
+homepage = 'https://github.com/Phlya/adjustText'
+description = "A small library for automatically adjustment of text position in matplotlib plots to minimize overlaps."
+
+toolchain = {'name': 'foss', 'version': '2022a'}
+
+sources = [SOURCE_TAR_GZ]
+checksums = ['b90e275a95b4d980cbbac7967914b8d66477c09bc346a0b3c9e2125bba664b06']
+
+dependencies = [
+    ('Python', '3.10.4'),
+    ('matplotlib', '3.5.2'),
+    ('SciPy-bundle', '2022.05'),
+]
+
+download_dep_fail = True
+use_pip = True
+
+sanity_pip_check = True
+
+options = {'modulename': 'adjustText'}
+
+moduleclass = 'tools'

64_scenicplus/gensim-4.2.0-foss-2022a.eb

@@ -0,0 +1,28 @@
+easyblock = 'PythonBundle'
+
+name = 'gensim'
+version = '4.3.2'
+
+homepage = 'https://radimrehurek.com/gensim'
+description = """Gensim is a Python library for topic modelling, document indexing and similarity retrieval with
+ large corpora."""
+
+toolchain = {'name': 'foss', 'version': '2022a'}
+
+dependencies = [
+    ('Python', '3.10.4'),
+    ('SciPy-bundle', '2022.05'),  # for numpy, pandas
+]
+
+use_pip = True
+
+exts_list = [
+    ('smart_open', '6.4.0', {
+    }),
+    (name, version, {
+    }),
+]
+
+sanity_pip_check = True
+
+moduleclass = 'lib'

64_scenicplus/old_logz/easybuild-run_cmd-84ywy4qt.log

@@ -0,0 +1,7 @@
+# output for command: /apps/gent/RHEL8/cascadelake-ib/software/Python/3.10.4-GCCcore-11.3.0/bin/python -c "import pyfasta"
+
+Traceback (most recent call last):
+  File "<string>", line 1, in <module>
+  File "/scratch/gent/vo/001/gvo00117/easybuild/RHEL8/cascadelake-ampere-ib/software/pyfasta/0.5.2-foss-2022a/lib/python3.10/site-packages/pyfasta/__init__.py", line 3, in <module>
+    from fasta import Fasta, complement, DuplicateHeaderException
+ModuleNotFoundError: No module named 'fasta'

64_scenicplus/pyfasta-0.5.2-foss-2022a.eb

@@ -0,0 +1,29 @@
+easyblock = 'PythonPackage'
+
+name = 'pyfasta'
+version = '0.5.2'
+
+homepage = 'https://pypi.org/project/pyfasta/'
+description = """fast, memory-efficient, pythonic (and command-line) access to fasta sequence files"""
+
+toolchain = {'name': 'foss', 'version': '2022a'}
+
+sources = [SOURCE_TAR_GZ]
+
+use_pip = True
+download_dep_fail = True
+
+dependencies = [
+    ('Python', '3.10.4'),
+    ('SciPy-bundle', '2022.05'),
+]
+
+sanity_check_paths = {
+    'files': ['bin/pyfasta'],
+    'dirs': ['lib/python%(pyshortver)s/site-packages'],
+}
+# sanity_check_commands = ['pyfasta extract --help']
+
+sanity_pip_check = True
+
+moduleclass = 'tools'

64_scenicplus/scenicplus.eb

@@ -0,0 +1,115 @@
+easyblock = 'PythonBundle'
+
+name = 'scenicplus'
+version = '1.0.0'
+
+homepage = 'https://github.com/aertslab/scenicplus'
+description = """SCENIC+ is a python package to build gene regulatory networks (GRNs)
+using combined or separate single-cell gene expression (scRNA-seq) and single-cell
+chromatin accessibility (scATAC-seq) data."""
+
+toolchain = {'name': 'foss', 'version': '2022a'}
+
+builddependencies = [
+    ('binutils', '2.38'),
+]
+
+dependencies = [
+    ('Python', '3.10.4'),
+    ('SciPy-bundle', '2022.05'),
+    ('tqdm', '4.64.0'),
+    ('Ray-project', '2.2.0'),
+    ('attr', '2.5.1'),
+    ('typing-extensions', '4.3.0'),
+    ('matplotlib', '3.5.2'),
+    ('scikit-learn', '1.1.2'),
+    ('networkx', '2.8.4'),
+    ('Seaborn', '0.12.1'),
+    ('pyBigWig', '0.3.18'),
+    ('plotly.py', '5.12.0'),
+    ('MACS2', '2.2.9.1'),
+    ('lxml', '4.9.1'),
+    ('Kaleido', '0.2.1'),
+    ('IPython', '8.5.0'),
+    ('scanpy', '1.9.1'),
+    ('Cython', '0.29.33'),
+    ('statsmodels', '0.13.1'),
+    ('h5py', '3.7.0'),
+    ('adjustText', '0.7.3'),
+    ('umap-learn', '0.5.3'),
+    ('ipympl', '0.9.3'),
+    ('leidenalg', '0.9.1'),
+    ('loompy', '3.0.7'),
+    ('numba', '0.56.4'),
+    ('polars', '0.15.6'),
+    ('pybedtools', '0.9.0'),
+]
+
+use_pip = True
+sanity_pip_check = True
+
+exts_list = [
+    ('pyranges', '0.0.129'),
+    ('interval-tree', '0.3.4'),
+    ('ncls', '0.0.68'),
+    ('sorted_nearest', '0.0.39'),
+    ('natsort', '8.4.0'),
+    ('pyranges', '0.0.129'),
+    ('cattrs', '23.1.2'),
+    ('url-normalize', '1.4.3'),
+    ('requests-cache', '1.1.0'),
+    ('pybiomart', '0.2.0'),
+    ('frozendict', '2.3.8', {
+        'checksums': ['5526559eca8f1780a4ee5146896f59afc31435313560208dd394a3a5e537d3ff'],
+    }),
+    ('arboreto', '0.1.6', {
+        # the source on PyPI is missing 'requirements.txt', fetching from github instead
+        'source_urls': ['https://github.com/aertslab/arboreto/archive/refs/tags/'],
+        'sources': [{'download_filename': '%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
+        'checksums': ['f94c0bc084ca61422e8b74ecdb80585a45b168ac682f18dd3c52c9097ee29caa'],
+    }),
+    ('gseapy', '1.0.6'),
+    ('ctxcore', '0.2.0', {
+        'sources': ['ctxcore-0.2.0-py3-none-any.whl'],
+        'checksums': ['b90570377e26280c4861ebad1f4cee2fe598167c5d4bd12c1b713f03c9682627'],
+    }),
+    ('tspex', '0.6.3'),
+    ('jsonpickle', '2.2.0', {
+        'checksums': ['7b272918b0554182e53dc340ddd62d9b7f902fec7e7b05620c04f3ccef479a0e'],
+    }),
+    ('pyvis', '0.3.2'),
+        ('progressbar2', '4.2.0', {
+        'modulename': 'progressbar',
+        'checksums': ['1393922fcb64598944ad457569fbeb4b3ac189ef50b5adb9cef3284e87e394ce'],
+    }),
+    ('pygam', '0.9.0'),
+    ('mizani', '0.10.0'),
+    ('patsy', '0.5.2', {
+        'checksums': ['5053de7804676aba62783dbb0f23a2b3d74e35e5bfa238b88b7cbf148a38b69d'],
+    }),
+    ('plotnine', '0.12.3'),
+    ('anndata', '0.8.0', {
+    }),
+    ('bbknn', '1.5.1'),
+    ('mudata', '0.2.0', {
+    }),
+    ('annoy', '1.17.1'),
+    ('pynndescent', '0.5.8', {
+        'checksums': ['a7c552569bf604a101fd54bba1d27c12389e065945dee3a6777a518c63a46f2b'],
+    }),
+    ('harmonypy', '0.0.9'),
+    ('lda', '2.0.0'),
+    ('pyOpenSSL', '21.0.0', {
+        'modulename': 'OpenSSL',
+        'checksums': ['5e2d8c5e46d0d865ae933bef5230090bdaf5506281e9eec60fa250ee80600cb3'],
+    }),
+    ('pycisTopic', '1.0.2', {
+        'source_urls': ['https://github.com/aertslab/pycisTopic/archive/refs/tags/'],
+        'sources': [{'download_filename': 'v%(version)s.tar.gz', 'filename': SOURCE_TAR_GZ}],
+    })
+    (name, version, {
+        'checksums': ['f8ecaf91bded57e12676d0373bf27acc154dea56c386f6ad05b476bc2c6ddc4f'],
+    }),
+]
+
+moduleclass = 'bio'