def _assign_taxonomy(self, extracted_reads, assignment_method):
graftm_align_directory_base = os.path.join(self._working_directory, 'graftm_aligns')
os.mkdir(graftm_align_directory_base)
commands = []
for singlem_package, sample_names, tmp_grafts in extracted_reads.each_package_wise():
tmpnames = list([tg.name for tg in tmp_grafts if tg])
if len(tmpnames) > 0:
cmd = "%s "\
"--threads %i "\
"--forward %s "\
"--graftm_package %s "\
"--output_directory %s/%s "\
"--max_samples_for_krona 0 "\
"--assignment_method %s" % (
self._graftm_command_prefix(singlem_package.is_protein_package()),
self._num_threads,
' '.join(tmpnames),
singlem_package.graftm_package_path(),
graftm_align_directory_base,
singlem_package.graftm_package_basename(),
assignment_method)
commands.append(cmd)
extern.run_many(commands, num_threads=1)
logging.info("Finished running taxonomic assignment with graftm")
return SingleMPipeTaxonomicAssignmentResult(graftm_align_directory_base)
def _align(self, search_result):
graftm_separate_directory_base = os.path.join(self._working_directory, 'graftm_separates')
os.mkdir(graftm_separate_directory_base)
logging.info("Running separate alignments in GraftM..")
commands = []
def command(singlem_package, hit_files, is_protein):
return self._graftm_command_prefix(is_protein) + \
"--threads %i "\
"--forward %s "\
"--graftm_package %s --output_directory %s/%s "\
"--search_only" % (
1, #use 1 thread since most likely better to parallelise processes with extern
' '.join(hit_files),
singlem_package.graftm_package_path(),
graftm_separate_directory_base,
os.path.basename(singlem_package.graftm_package_path()))
# Gather commands for aligning protein packages
for singlem_package in self._singlem_package_database.protein_packages():
commands.append(command(singlem_package, search_result.protein_hit_paths().values(), True))
# Gather commands for aligning nucleotide packages.
for singlem_package in self._singlem_package_database.nucleotide_packages():
temporary_hit_files = [tf for _, tf in \
search_result.direction_corrected_nucleotide_read_files()]
commands.append(command(singlem_package, temporary_hit_files, False))
extern.run_many(commands, num_threads=self._num_threads)
return SingleMPipeAlignSearchResult(
graftm_separate_directory_base, search_result.samples_with_hits())
def _assign_taxonomy(self, extracted_reads, assignment_method):
graftm_align_directory_base = os.path.join(self._working_directory,
'graftm_aligns')
os.mkdir(graftm_align_directory_base)
commands = []
all_tmp_files = []
# Run each one at a time serially so that the number of threads is
# respected, to save RAM as one DB needs to be loaded at once, and so
# fewer open files are needed, so that the open file count limit is
# eased.
for singlem_package, readsets in extracted_reads.each_package_wise():
tmp_files = []
for readset in readsets:
if len(readset.sequences) > 0:
tmp = tempfile.NamedTemporaryFile(prefix='singlem.%s' %
readset.sample_name,
suffix=".fasta")
# Record basename (remove .fasta) so that the graftm output
# file is recorded for later on in pipe.
tmpbase = os.path.basename(tmp.name[:-6])
readset.tmpfile_basename = tmpbase
seqio = SequenceIO()
seqio.write_fasta(readset.sequences, tmp)
tmp.flush()
tmp_files.append(tmp)
if len(tmp_files) > 0:
tmpnames = list([tg.name for tg in tmp_files])
cmd = "%s "\
"--threads %i "\
"--forward %s "\
"--graftm_package %s "\
"--output_directory %s/%s "\
"--max_samples_for_krona 0 "\
"--assignment_method %s" % (
self._graftm_command_prefix(singlem_package.is_protein_package()),
self._num_threads,
' '.join(tmpnames),
singlem_package.graftm_package_path(),
graftm_align_directory_base,
singlem_package.graftm_package_basename(),
assignment_method)
commands.append(cmd)
all_tmp_files.append(tmp_files)
extern.run_many(commands, num_threads=1)
for tmp_files in all_tmp_files:
[t.close() for t in tmp_files]
logging.info("Finished running taxonomic assignment with GraftM")
return SingleMPipeTaxonomicAssignmentResult(
graftm_align_directory_base)
def test_multi_with_exception(self):
with self.assertRaises(ExternCalledProcessError) as ex:
extern.run_many(['seq 2','cat /notafile'])
self.assertEqual('Command cat /notafile returned non-zero exit status 1.\nSTDERR was: cat: /notafile: No such file or directory\nSTDOUT was: ',
str(ex.exception))
def test_multi_with_many_threads(self):
commands = ['seq 2','seq 3 4']*50
answers = ['1\n2\n','3\n4\n']*50
self.assertEqual(answers, extern.run_many(commands, num_threads=10))
def test_multi_hello_world(self):
self.assertEqual(['1\n2\n'], extern.run_many(['seq 2'], num_threads=1))
