Merge pull request #22027 from akesandgren/20241213164448_new_pr_Pars…

…np211

{bio}[foss/2023b] Parsnp v2.1.1

easybuild/easyconfigs/f/FastANI/FastANI-1.34-GCC-13.2.0.eb

@@ -0,0 +1,45 @@
+# easybuild easyconfig
+#
+# John Dey [email protected]
+#
+# Updated: Pavel Grochal (INUITS)
+
+easyblock = 'ConfigureMake'
+
+name = 'FastANI'
+version = '1.34'
+
+homepage = "https://github.com/ParBLiSS/FastANI"
+description = """FastANI is developed for fast alignment-free computation of
+ whole-genome Average Nucleotide Identity (ANI). ANI is defined as mean
+ nucleotide identity of orthologous gene pairs shared between two microbial
+ genomes. FastANI supports pairwise comparison of both complete and draft
+ genome assemblies."""
+
+toolchain = {'name': 'GCC', 'version': '13.2.0'}
+
+source_urls = ['https://github.com/ParBLiSS/FastANI/archive']
+sources = ['v%(version)s.zip']
+patches = ['FastANI-1.2-memcpy.patch']
+checksums = [
+    {'v1.34.zip': 'cb15540634c725cb46dded7becaff38b27a7f709c0a8589db986674effcc6180'},
+    {'FastANI-1.2-memcpy.patch': 'eebcf0b64c31ee360ca79136f644157064ac69747ed13cff70f5c9932c6bb0d5'},
+]
+
+builddependencies = [('Autotools', '20220317')]
+
+dependencies = [
+    ('GSL', '2.7'),
+    ('zlib', '1.2.13'),
+]
+
+preconfigopts = 'autoconf && '
+
+sanity_check_paths = {
+    'files': ['bin/fastANI'],
+    'dirs': []
+}
+
+sanity_check_commands = ["fastANI --help"]
+
+moduleclass = 'bio'

easybuild/easyconfigs/h/harvest-tools/harvest-tools-1.3-GCCcore-13.2.0.eb

@@ -0,0 +1,60 @@
+easyblock = 'ConfigureMake'
+
+name = 'harvest-tools'
+version = '1.3'
+
+homepage = 'https://harvest.readthedocs.io/en/latest/content/harvest-tools.html'
+description = """HarvestTools is a utility for creating and interfacing
+with Gingr files, which are efficient archives that the Harvest Suite
+uses to store reference-compressed multi-alignments, phylogenetic trees,
+filtered variants and annotations. Though designed for use with Parsnp
+and Gingr, HarvestTools can also be used for generic conversion between
+standard bioinformatics file formats."""
+
+toolchain = {'name': 'GCCcore', 'version': '13.2.0'}
+toolchainopts = {'pic': True}
+
+github_account = 'marbl'
+source_urls = [GITHUB_SOURCE]
+sources = ['v%(version)s.tar.gz']
+patches = [
+    'harvest-tools-1.3_adapt_to_newer-gcc_protobuf_and_EB.patch',
+]
+checksums = [
+    {'v1.3.tar.gz': 'ffbcf0a115c74507695fd6cee4a9d5ba27a700db36b32d226521ef8dd3309264'},
+    {'harvest-tools-1.3_adapt_to_newer-gcc_protobuf_and_EB.patch':
+     '9d04d6c942d42a8147f43f8f086f436d284925f3a3acdd9620df84d4041dff61'},
+]
+
+builddependencies = [
+    ('binutils', '2.40'),
+    ('Autotools', '20220317'),
+]
+
+dependencies = [
+    ('zlib', '1.2.13'),
+    ('protobuf', '25.3'),
+    ('CapnProto', '1.0.1.1'),
+]
+
+preconfigopts = 'autoconf && '
+
+configopts = ' '.join([
+    '--with-protobuf=$EBROOTPROTOBUF',
+    '--with-capnp=$EBROOTCAPNPROTO',
+])
+
+postinstallcmds = [
+    'ln -s harvesttools %(installdir)s/bin/harvest',
+]
+
+sanity_check_paths = {
+    'files': ['bin/harvesttools', 'lib/libharvest.%s' % SHLIB_EXT],
+    'dirs': ['include/harvest'],
+}
+
+sanity_check_commands = [
+    'harvest -h',
+]
+
+moduleclass = 'bio'

easybuild/easyconfigs/h/harvest-tools/harvest-tools-1.3_adapt_to_newer-gcc_protobuf_and_EB.patch

@@ -0,0 +1,113 @@
+commit 845bb6e29bc12cf5f989c0c64fb98002cb57b343
+Author: Ake Sandgren <[email protected]>
+Date:   Mon Dec 16 18:10:22 2024 +0100
+
+    Adapt to newer compiler/protobuf and EasyBuild
+    And actually install the binary, library and include files
+
+diff --git a/Makefile.in b/Makefile.in
+index fdb2ce0..434ba0a 100644
+--- a/Makefile.in
++++ b/Makefile.in
+@@ -1,4 +1,4 @@
+-CXXFLAGS += -std=c++11 -Isrc -I@protobuf@/include -I@capnp@/include
++CXXFLAGS += -Isrc
+
+ UNAME_S=$(shell uname -s)
+
+@@ -23,10 +23,13 @@ SOURCES=\
+
+ OBJECTS=$(SOURCES:.cpp=.o) src/harvest/pb/harvest.pb.o src/harvest/capnp/harvest.capnp.o
+
+-all : harvesttools libharvest.a
++all : harvesttools libharvest.a libharvest.so
+
+-harvesttools : libharvest.a src/harvest/memcpyWrap.o
+-	$(CXX) $(CXXFLAGS) $(CPPFLAGS) -o harvesttools src/harvest/memcpyWrap.o libharvest.a @protobuf@/lib/libprotobuf.a @capnp@/lib/libcapnp.a @capnp@/lib/libkj.a -lstdc++ -lz -lm -lpthread
++harvesttools : libharvest.so src/harvest/memcpyWrap.o
++	$(CXX) $(CXXFLAGS) $(CPPFLAGS) -o harvesttools src/harvest/memcpyWrap.o -L. -lharvest -lprotobuf -lcapnp -lkj -lstdc++ -lz -lm -lpthread
++
++libharvest.so : $(OBJECTS)
++	g++ -shared -o libharvest.so $(OBJECTS)
+
+ libharvest.a : $(OBJECTS)
+ 	ar -cr libharvest.a $(OBJECTS)
+@@ -52,27 +55,28 @@ src/harvest/pb/harvest.pb.cc src/harvest/pb/harvest.pb.h : src/harvest/pb/harves
+ src/harvest/capnp/harvest.capnp.c++ src/harvest/capnp/harvest.capnp.h : src/harvest/capnp/harvest.capnp
+ 	cd src/harvest/capnp;export PATH=@capnp@/bin/:${PATH};capnp compile -I @capnp@/include -oc++ harvest.capnp
+
+-install : libharvest.a
+-	mkdir -p @prefix@/bin/
+-	mkdir -p @prefix@/lib/
+-	mkdir -p @prefix@/include/
+-	mkdir -p @prefix@/include/harvest
+-	mkdir -p @prefix@/include/harvest/capnp
+-	mkdir -p @prefix@/include/harvest/pb
+-	ln -sf `pwd`/harvesttools @prefix@/bin/
+-	ln -sf `pwd`/libharvest.a @prefix@/lib/
+-	ln -sf `pwd`/src/harvest/exceptions.h @prefix@/include/harvest/
+-	ln -sf `pwd`/src/harvest/HarvestIO.h @prefix@/include/harvest/
+-	ln -sf `pwd`/src/harvest/capnp/harvest.capnp.h @prefix@/include/harvest/capnp/
+-	ln -sf `pwd`/src/harvest/pb/harvest.pb.h @prefix@/include/harvest/pb/
+-	ln -sf `pwd`/src/harvest/ReferenceList.h @prefix@/include/harvest/
+-	ln -sf `pwd`/src/harvest/AnnotationList.h @prefix@/include/harvest/
+-	ln -sf `pwd`/src/harvest/parse.h @prefix@/include/harvest/
+-	ln -sf `pwd`/src/harvest/PhylogenyTree.h @prefix@/include/harvest/
+-	ln -sf `pwd`/src/harvest/PhylogenyTreeNode.h @prefix@/include/harvest/
+-	ln -sf `pwd`/src/harvest/TrackList.h @prefix@/include/harvest/
+-	ln -sf `pwd`/src/harvest/LcbList.h @prefix@/include/harvest/
+-	ln -sf `pwd`/src/harvest/VariantList.h @prefix@/include/harvest/
++install : all
++	install -d @prefix@/bin/
++	install -d @prefix@/lib/
++	install -d @prefix@/include/
++	install -d @prefix@/include/harvest
++	install -d @prefix@/include/harvest/capnp
++	install -d @prefix@/include/harvest/pb
++	install harvesttools @prefix@/bin/
++	install libharvest.a @prefix@/lib/
++	install libharvest.so @prefix@/lib/
++	install src/harvest/exceptions.h @prefix@/include/harvest/
++	install src/harvest/HarvestIO.h @prefix@/include/harvest/
++	install src/harvest/capnp/harvest.capnp.h @prefix@/include/harvest/capnp/
++	install src/harvest/pb/harvest.pb.h @prefix@/include/harvest/pb/
++	install src/harvest/ReferenceList.h @prefix@/include/harvest/
++	install src/harvest/AnnotationList.h @prefix@/include/harvest/
++	install src/harvest/parse.h @prefix@/include/harvest/
++	install src/harvest/PhylogenyTree.h @prefix@/include/harvest/
++	install src/harvest/PhylogenyTreeNode.h @prefix@/include/harvest/
++	install src/harvest/TrackList.h @prefix@/include/harvest/
++	install src/harvest/LcbList.h @prefix@/include/harvest/
++	install src/harvest/VariantList.h @prefix@/include/harvest/
+
+ clean :
+ 	-rm harvesttools
+diff --git a/configure.ac b/configure.ac
+index fc8e61d..5e2fc22 100644
+--- a/configure.ac
++++ b/configure.ac
+@@ -29,8 +29,6 @@ then
+ 	AC_MSG_ERROR([Cap'n Proto compiler (capnp) not found.])
+ fi
+
+-CPPFLAGS="-I$with_protobuf/include -I$with_capnp/include -std=c++11"
+-
+ AC_CHECK_HEADER(google/protobuf/stubs/common.h, [result=1], [result=0])
+
+ if test $result == 0
+diff --git a/src/harvest/HarvestIO.cpp b/src/harvest/HarvestIO.cpp
+index c165dce..c1dbb7a 100755
+--- a/src/harvest/HarvestIO.cpp
++++ b/src/harvest/HarvestIO.cpp
+@@ -208,7 +208,11 @@ bool HarvestIO::loadHarvestProtocolBuffer(const char * file)
+ 	GzipInputStream gz(&raw_input);
+ 	CodedInputStream coded_input(&gz);
+
++#if GOOGLE_PROTOBUF_VERSION >= 3002000
++	coded_input.SetTotalBytesLimit(INT_MAX);
++#else
+ 	coded_input.SetTotalBytesLimit(INT_MAX, INT_MAX);
++#endif
+
+ 	if ( ! harvest.ParseFromCodedStream(&coded_input) )
+ 	{

easybuild/easyconfigs/p/Parsnp/Parsnp-2.1.1-foss-2023b.eb

@@ -0,0 +1,88 @@
+easyblock = 'ConfigureMake'
+
+name = 'Parsnp'
+version = '2.1.1'
+
+homepage = 'https://harvest.readthedocs.io/en/latest/content/parsnp.html'
+description = """Parsnp was designed to align the core genome of
+hundreds to thousands of bacterial genomes within a few minutes to few
+hours. Input can be both draft assemblies and finished genomes, and
+output includes variant (SNP) calls, core genome phylogeny and
+multi-alignments. Parsnp leverages contextual information provided by
+multi-alignments surrounding SNP sites for filtration/cleaning, in
+addition to existing tools for recombination detection/filtration and
+phylogenetic reconstruction."""
+
+toolchain = {'name': 'foss', 'version': '2023b'}
+toolchainopts = {'openmp': True}
+
+github_account = 'marbl'
+source_urls = [GITHUB_LOWER_SOURCE]
+sources = ['v%(version)s.tar.gz']
+patches = [
+    'Parsnp-2.1.1_fix_configure_search_for_libMUSCLE.patch',
+]
+checksums = [
+    {'v2.1.1.tar.gz': '7887ee4030dad6f30a4b735f1ee463024204030adbdcb0fa96562e6e8664474e'},
+    {'Parsnp-2.1.1_fix_configure_search_for_libMUSCLE.patch':
+     '7b601c442ed4f64b53195b9adefafbb0d9639296513e22961a8e0de672dbb3fb'},
+]
+
+builddependencies = [
+    ('Autotools', '20220317'),
+]
+
+dependencies = [
+    ('Python', '3.11.5'),
+    ('SciPy-bundle', '2023.11'),
+    ('tqdm', '4.66.2'),
+    ('Biopython', '1.84'),
+    ('FastTree', '2.1.11'),
+    ('harvest-tools', '1.3'),
+    ('RAxML', '8.2.13', '-avx2'),
+    ('pyspoa', '0.2.1'),
+    ('Mash', '2.3'),
+    ('FastANI', '1.34'),
+    ('PhiPack', '2016.06.14'),
+]
+
+preconfigopts = ' && '.join([
+    'cd muscle',
+    './autogen.sh',
+    './configure --prefix=%(installdir)s',
+    'make install',
+    'cd ..',
+    './autogen.sh',
+    '',
+])
+
+configopts = '--with-libmuscle=%(installdir)s'
+
+preinstallopts = ' && '.join([
+    'cd muscle',
+    'make install',
+    'cd ..',
+    '',
+])
+
+postinstallcmds = [
+    'cp parsnp %(installdir)s/bin',
+    'cp template.ini %(installdir)s',
+    'mkdir -p %(installdir)s/lib/python%(pyshortver)s/site-packages',
+    'cp partition.py extend.py logger.py %(installdir)s/lib/python%(pyshortver)s/site-packages',
+]
+
+sanity_check_paths = {
+    'files': ['bin/parsnp_core', 'bin/parsnp', 'template.ini'],
+    'dirs': ['lib/python%(pyshortver)s/site-packages'],
+}
+
+modextravars = {
+    'PARSNPDIR': '',
+}
+
+modextrapaths = {
+    'PYTHONPATH': 'lib/python%(pyshortver)s/site-packages',
+}
+
+moduleclass = 'bio'

easybuild/easyconfigs/p/Parsnp/Parsnp-2.1.1_fix_configure_search_for_libMUSCLE.patch

@@ -0,0 +1,27 @@
+Fix search for libMUSCLE's include files and install target
+
+Åke Sandgren, 2024-12-13
+diff -ru parsnp-2.1.1.orig/configure.ac parsnp-2.1.1/configure.ac
+--- parsnp-2.1.1.orig/configure.ac	2024-11-21 00:54:04.000000000 +0100
++++ parsnp-2.1.1/configure.ac	2024-12-13 12:33:40.964024323 +0100
+@@ -11,7 +11,7 @@
+
+ AC_LANG(C++)
+
+-CPPFLAGS="-I$with_libmuscle"
++CPPFLAGS="-I$with_libmuscle/include"
+ AC_CHECK_HEADER(libMUSCLE/muscle.h, [result=1], [result=0])
+
+ if test $result == 0
+diff -ru parsnp-2.1.1.orig/src/Makefile.am parsnp-2.1.1/src/Makefile.am
+--- parsnp-2.1.1.orig/src/Makefile.am	2024-11-21 00:54:04.000000000 +0100
++++ parsnp-2.1.1/src/Makefile.am	2024-12-13 12:55:02.385984661 +0100
+@@ -2,6 +2,6 @@
+ parsnp_core_LDFLAGS =  -fopenmp -lstdc++ -lpthread -std=gnu++0x -Wl,-rpath,$(libmuscle)/lib -L$(libmuscle)/lib -L${CONDA_PREFIX}/lib -lMUSCLE-3.7 
+ bin_PROGRAMS = parsnp_core
+ parsnp_core_SOURCES = MuscleInterface.cpp MuscleInterface.h parsnp.cpp parsnp.hh LCB.cpp LCB.hh LCR.cpp LCR.hh TMum.cpp TMum.hh Converter.cpp Converter.hh ./ext/iniFile.cpp ./ext/iniFile.h 
+-bindir = $(top_srcdir)/bin
+-libdir = $(top_srcdir)/lib
++#bindir = $(top_srcdir)/bin
++#libdir = $(top_srcdir)/lib
+

easybuild/easyconfigs/p/PhiPack/PhiPack-2016.06.14-GCC-13.2.0.eb

@@ -0,0 +1,39 @@
+# Author: Pavel Grochal (INUITS)
+# License: GPLv2
+
+easyblock = "MakeCp"
+
+name = "PhiPack"
+# No version is specified by default, copyright is 2005 latest update 2016
+# find -type f -printf '%T+ %p\n' | sort -r | head -n 1
+# 2016-06-14+06:18:29.0000000000 ./PhiPack/src/fasta.c
+# latest edit time of file in the archive
+version = "2016.06.14"
+
+homepage = "http://www.maths.otago.ac.nz/~dbryant/software.html"
+description = """The PhiPack software package implements (in C) a few tests for
+recombination and can produce refined incompatibility matrices as well."""
+
+docurls = ["http://www.maths.otago.ac.nz/~dbryant/software/phimanual.pdf"]
+software_license = "LicenseGPLv3"
+software_license_urls = ['https://www.gnu.org/licenses/lgpl-3.0.en.html']
+
+toolchain = {'name': 'GCC', 'version': '13.2.0'}
+
+source_urls = ['http://www.maths.otago.ac.nz/~dbryant/software/']
+sources = [{'download_filename': 'PhiPack.tar.gz', 'filename': SOURCE_TAR_GZ}]
+checksums = ['af43326e021f1f2e590be2cba3caa44a0963f237386e63209ccc26b5bfb02db9']
+
+start_dir = 'src/'
+buildopts = 'CXX="$CC" CXXFLAGS="$CFLAGS"'
+
+files_to_copy = [(['Phi', 'Profile', 'ppma_2_bmp/ppma_2_bmp'], 'bin')]
+
+sanity_check_paths = {
+    'files': ['bin/Phi', 'bin/Profile', 'bin/ppma_2_bmp'],
+    'dirs': [],
+}
+
+sanity_check_commands = ["Phi -f %(builddir)s/%(name)s/noro.fasta"]
+
+moduleclass = 'math'