#235 smafa + cd-hit + polars

235_singleM/SingleM-0.16.0-foss-2023a.eb

@@ -11,60 +11,100 @@ finding bias in genome recovery, computing ecological diversity metrics, and lin
 
 toolchain = {'name': 'foss', 'version': '2023a'}
 
+# DEPS
+    # OK - python>=3.7 #, <3.10 # https://github.com/google-research/google-research/issues/779 means scann (at least 1.2.8) fails to install via pip for 3.10. Older python versions than 3.7 may also work, not tested.
+    # OK - diamond>=2.1.7 # Getting segmentation fault on test_insert_prefilter otherwise
+    # OK - biopython
+    # OK - hmmer
+    # OK - orfm
+    # OK - mfqe
+    # ok - extern
+    # ok - graftm>=0.15.0 # Version needed for --translation-table
+    # OK - krona
+    # OK - pplacer
+    # OK - sra-tools
+    # OK - ncbi-ngs-sdk - in SRA-Toolkit
+    # OK - sqlite - in python
+    # ok - squarify
+    # OK - mafft
+    # OK - seqmagick
+    # OK - expressbetadiversity
+    # OK - cd-hit -> created
+    # OK - fasttree
+    # ok - fastalite
+    # OK - jinja2 - pypi bundle
+    # OK - pip
+    # OK - sqlalchemy
+    # OK - pandas
+    # ok - bird_tool_utils_python>=0.4.1
+    # ok - zenodo_backpack
+    # OK - smafa>=0.7.0 -> created
+    # ok - pyranges
+    # OK - polars>=0.19.3 # supplement with taxon genome length requires this, for strip_chars
+    # OK - prodigal
+    # OK - tqdm
+    # OK - pyarrow
+    # # - galah >= 0.4.0 # 0.4.0 is the first version with checkm2 quality inputs, but not yet released 
+    # - pytest # For testing only
+    # - ipython # For testing only
+    # - sqlparse # Don't understand why, but this isn't being installed before tests are run on GH actions. Required indirectly (by taxtastic).
+
 dependencies = [
     ('Python', '3.11.3'), # with SQLite
+    ('Python-bundle-PyPI', '2023.06'),
     ('SciPy-bundle', '2023.07'),  # for pandas
-    ('matplotlib', '3.7.2'),
+    # ('matplotlib', '3.7.2'),
     ('Biopython', '1.83'), 
-    ('polars', '0.19.19'), # polars >= 0.19.3 
+    ('polars', '0.20.2'), # polars >= 0.19.3 
     ('tqdm', '4.66.1'),
     ('Arrow', '14.0.1'), # for pyarrow
     ('DIAMOND', '2.1.8'),
     ('HMMER', '3.4'),
     ('MAFFT', '7.520', '-with-extensions'),
-    ('tqdm', '4.66.1'),
     ('OrfM', '0.7.1'),
     ('mfqe', '0.5.0'),
     ('KronaTools', '2.8.1'),
-    ('pplacer', '1.1.alpha19', '', True),
+    ('pplacer', '1.1.alpha19', '', SYSTEM),
     ('SRA-Toolkit', '3.0.10'),
     ('Seqmagick', '0.8.6'),
     ('ExpressBetaDiversity', '1.0.10'),
+    ('FastTree', '2.1.11'),
+    ('SQLAlchemy', '2.0.25'),
+    ('CD-HIT', '4.8.1'),
+    ('smafa', '0.8.0'),
+    ('prodigal', '2.6.3'),
+    #
     # ('DendroPy', '4.4.0'),
     # ('PyYAML', '5.1'),
     # ('PostgreSQL', '11.3', versionsuffix),  # required for psycopg2
-    # ('smafa', '0.4.0'),
     # ('VSEARCH', '2.13.4'),
     # ('fxtract', '2.3'),
 ]
 
 use_pip = True
 sanity_pip_check = True
 
-# required because we're building Python packages using Intel compilers on top of Python built with GCC
-# check_ldshared = True
-
- # ('zenodo_backpack', '0.2.0', {
-#     'checksums': ['5843a04b353ae613b474932c2a12d1fbfd529674074e73927ddc131eba7f1da8'],
-# }),
-
-# ('bird_tool_utils', '0.4.1', {
-#     'checksums': ['6fe80f9608626427e8d382c5341c24088d61f17336fb6ce834d40aa4577499b5'],
-# }),
 exts_list = [
-    ('Jinja2', '3.1.3'),
-    ('fastalite', '0.4.1'),
+    ('zenodo_backpack', '0.3.1', {
+        'checksums': ['dc91f3c427f976465789746e94736abfa536cf42dc2e49b6d6067382a9a39b26'],
+    }),
+    ('bird_tool_utils', '0.4.1', {
+        'checksums': ['6fe80f9608626427e8d382c5341c24088d61f17336fb6ce834d40aa4577499b5'],
+    }),
+    ('fastalite', '0.4.1', {
+        'checksums': ['e85413ee22bdb3fe0f73f5226771cf71eb33074ccdf8bbefff3a1bc6242de37c'],
+    }),
+    ('squarify', '0.4.3', {
+        'checksums': ['54091f6ad175f7f201f8934574e647ce1b50dedc478c5fd968688eb7d7469f95'],
+    }),
     ('pyranges', '0.0.129', {
         'checksums': ['bee83b4fad0062be9586668c6b0fc4270d5e761951975e018202993680071fb3'],
     }),
-    ('SQLAlchemy', '2.0.25', {
-        'checksums': ['a2c69a7664fb2d54b8682dd774c3b54f67f84fa123cf84dda2a5f40dcaa04e08'],
-    }),
     ('extern', '0.4.1', {
         'checksums': ['0ff01adc2ad423f3d1e31641024b3974569fb0127b4d925bc6bed1cb86b6b1e4'],
     }),
-    ('graftm', '0.15.0', {
-        'checksums': ['794b2c6a8fff36c81db8b9a544ac021ab1397c569638bf7651ce5329db36de21'],
+    ('graftm', '0.15.1', {
+        'checksums': ['80d828c311d2d6067977cfad5b6bac7cbc5d223ef8ab770d676b39bf2bc75163'],
     }),
     ('singlem', version, {
         'checksums': ['64e43a6a40795d68ff5aed7dfff9a94532b862f25a28c27de7d588d64a8c7f79'],

235_singleM/cdhit.eb

@@ -0,0 +1,39 @@
+# This file is an EasyBuild reciPY as per https://github.com/easybuilders/easybuild
+# changed toolchain and Perl version
+# Updated by: Thomas Eylenbosch(Gluo N.V.), Pavel Tománek (Inuits)
+
+easyblock = 'MakeCp'
+
+name = 'CD-HIT'
+version = '4.8.1'
+
+homepage = 'http://weizhongli-lab.org/cd-hit/'
+description = """ CD-HIT is a very widely used program for clustering and 
+ comparing protein or nucleotide sequences."""
+
+toolchain = {'name': 'GCC', 'version': '12.3.0'}
+toolchainopts = {'openmp': True}
+
+source_urls = ['https://github.com/weizhongli/cdhit/releases/download/V%(version)s/']
+sources = ['%(namelower)s-v%(version)s-2019-0228.tar.gz']
+checksums = ['26172dba3040d1ae5c73ff0ac6c3be8c8e60cc49fc7379e434cdf9cb1e7415de']
+
+dependencies = [
+    ('Perl', '5.36.1'),
+    ('zlib', '1.2.13'),
+]
+
+buildopts = ' CC="$CXX" CCFLAGS="$CPPFLAGS $CXXFLAGS"'
+
+local_list_of_executables = ['cd-hit', 'cd-hit-est', 'cd-hit-2d', 'cd-hit-est-2d', 'cd-hit-div', 'cd-hit-454']
+
+files_to_copy = [(local_list_of_executables, 'bin'), (['*.pl'], 'bin'), 'README', 'doc', 'license.txt']
+
+fix_perl_shebang_for = ['bin/*.pl']
+
+sanity_check_paths = {
+    'files': ['bin/%s' % x for x in local_list_of_executables],
+    'dirs': [],
+}
+
+moduleclass = 'bio'

235_singleM/crates_smafa.txt

@@ -0,0 +1,105 @@
+[('smafa', '0.8.0')]
+crates = [
+    ('addr2line', '0.22.0'),
+    ('adler', '1.0.2'),
+    ('aho-corasick', '1.1.3'),
+    ('anstream', '0.6.14'),
+    ('anstyle', '1.0.7'),
+    ('anstyle-parse', '0.2.4'),
+    ('anstyle-query', '1.1.0'),
+    ('anstyle-wincon', '3.0.3'),
+    ('assert_cli', '0.6.3'),
+    ('atomic-polyfill', '1.0.3'),
+    ('autocfg', '1.3.0'),
+    ('backtrace', '0.3.73'),
+    ('bird_tool_utils', '0.4.1'),
+    ('bird_tool_utils-man', '0.4.0'),
+    ('bitflags', '2.6.0'),
+    ('buffer-redux', '1.0.1'),
+    ('bytecount', '0.6.8'),
+    ('byteorder', '1.5.0'),
+    ('bzip2', '0.4.4'),
+    ('bzip2-sys', '0.1.11+1.0.8'),
+    ('cargo-husky', '1.5.0'),
+    ('cc', '1.1.5'),
+    ('cfg-if', '1.0.0'),
+    ('clap', '4.5.9'),
+    ('clap_builder', '4.5.9'),
+    ('clap_lex', '0.7.1'),
+    ('cobs', '0.2.3'),
+    ('colorchoice', '1.0.1'),
+    ('colored', '1.9.4'),
+    ('crc32fast', '1.4.2'),
+    ('critical-section', '1.1.2'),
+    ('difference', '2.0.0'),
+    ('embedded-io', '0.4.0'),
+    ('env_logger', '0.10.2'),
+    ('environment', '0.1.1'),
+    ('errno', '0.3.9'),
+    ('failure', '0.1.8'),
+    ('failure_derive', '0.1.8'),
+    ('fastrand', '2.1.0'),
+    ('flate2', '1.0.30'),
+    ('gimli', '0.29.0'),
+    ('hash32', '0.2.1'),
+    ('heapless', '0.7.17'),
+    ('hermit-abi', '0.3.9'),
+    ('humantime', '2.1.0'),
+    ('is-terminal', '0.4.12'),
+    ('is_terminal_polyfill', '1.70.0'),
+    ('itoa', '1.0.11'),
+    ('lazy_static', '1.5.0'),
+    ('libc', '0.2.155'),
+    ('linux-raw-sys', '0.4.14'),
+    ('lock_api', '0.4.12'),
+    ('log', '0.4.22'),
+    ('lzma-sys', '0.1.20'),
+    ('memchr', '2.7.4'),
+    ('miniz_oxide', '0.7.4'),
+    ('needletail', '0.5.1'),
+    ('object', '0.36.1'),
+    ('pkg-config', '0.3.30'),
+    ('postcard', '1.0.8'),
+    ('proc-macro2', '1.0.86'),
+    ('quote', '1.0.36'),
+    ('regex', '1.10.5'),
+    ('regex-automata', '0.4.7'),
+    ('regex-syntax', '0.8.4'),
+    ('roff', '0.1.0'),
+    ('rustc-demangle', '0.1.24'),
+    ('rustc_version', '0.4.0'),
+    ('rustix', '0.38.34'),
+    ('ryu', '1.0.18'),
+    ('scopeguard', '1.2.0'),
+    ('semver', '1.0.23'),
+    ('serde', '1.0.204'),
+    ('serde_derive', '1.0.204'),
+    ('serde_json', '1.0.120'),
+    ('spin', '0.9.8'),
+    ('stable_deref_trait', '1.2.0'),
+    ('strsim', '0.11.1'),
+    ('syn', '1.0.109'),
+    ('syn', '2.0.71'),
+    ('synstructure', '0.12.6'),
+    ('tempfile', '3.10.1'),
+    ('termcolor', '1.4.1'),
+    ('unicode-ident', '1.0.12'),
+    ('unicode-xid', '0.2.4'),
+    ('utf8parse', '0.2.2'),
+    ('version-compare', '0.1.1'),
+    ('winapi', '0.3.9'),
+    ('winapi-i686-pc-windows-gnu', '0.4.0'),
+    ('winapi-util', '0.1.8'),
+    ('winapi-x86_64-pc-windows-gnu', '0.4.0'),
+    ('windows-sys', '0.52.0'),
+    ('windows-targets', '0.52.6'),
+    ('windows_aarch64_gnullvm', '0.52.6'),
+    ('windows_aarch64_msvc', '0.52.6'),
+    ('windows_i686_gnu', '0.52.6'),
+    ('windows_i686_gnullvm', '0.52.6'),
+    ('windows_i686_msvc', '0.52.6'),
+    ('windows_x86_64_gnu', '0.52.6'),
+    ('windows_x86_64_gnullvm', '0.52.6'),
+    ('windows_x86_64_msvc', '0.52.6'),
+    ('xz2', '0.1.7'),
+]