def run(args):
"""
:run the wrapper
"""
if args.verbose:
loglevel = logging.INFO
elif args.quiet:
loglevel = logging.ERROR
elif args.debug:
loglevel = logging.DEBUG
else:
loglevel = logging.WARNING
# reinitialize logging
for handler in logging.root.handlers[:]:
logging.root.removeHandler(handler)
logging.basicConfig(filename=args.logfile,
format='%(asctime)s %(levelname)-8s %(message)s',
level=loglevel)
try:
grmpyOutput = vcfupdate.read_grmpy(args.input_grm)
vcfupdate.update_vcf_from_grmpy(args.input, grmpyOutput, args.output)
except Exception: # pylint: disable=W0703
traceback.print_exc(file=LoggingWriter(logging.ERROR))
raise
def run(args):
"""
:run the wrapper
"""
if args.verbose:
loglevel = logging.INFO
elif args.quiet:
loglevel = logging.ERROR
elif args.debug:
loglevel = logging.DEBUG
else:
loglevel = logging.WARNING
os.makedirs(args.output, exist_ok=True)
if args.scratch_dir:
os.makedirs(args.scratch_dir, exist_ok=True)
# reinitialize logging
for handler in logging.root.handlers[:]:
logging.root.removeHandler(handler)
logging.basicConfig(filename=args.logfile, format='%(asctime)s %(levelname)-8s %(message)s', level=loglevel)
# check format of manifest
with open(args.manifest) as manifest_file:
headers = {"id": False, "path": False, "idxdepth": False, "depth": False, "read length": False, "sex": False}
for line in manifest_file:
if line.startswith("#"):
line = line[1:]
line = line.strip()
fields = re.split('\t|,', line)
for field in fields:
if field not in headers:
header_str = ",".join(headers)
raise Exception("Illegal header name %s. Allowed headers:\n%s" % (field, header_str))
headers[field] = True
if not headers["id"] or not headers["path"]:
raise Exception("Missing header \"id\" or \"path\" in manifest")
if not headers["idxdepth"]:
if not headers["depth"] or not headers["read length"]:
raise Exception("Missing header \"idxdepth\", or \"depth\" and \"read length\" in manifest.")
break
# prepare input graph description
result_json_path = os.path.join(args.output, "genotypes.json.gz")
try:
graph_files = load_graph_description(args)
commandline = " -r %s" % pipes.quote(args.reference)
commandline += " -m %s" % pipes.quote(args.manifest)
commandline += " -o %s" % pipes.quote(result_json_path)
commandline += " -z"
if args.genotyping_parameters:
commandline += " -G %s" % pipes.quote(args.genotyping_parameters)
if args.max_reads_per_event:
commandline += " -M %s" % pipes.quote(str(args.max_reads_per_event))
if args.threads > 1:
commandline += " -t %s" % pipes.quote(str(args.threads))
if args.graph_sequence_matching:
commandline += " --graph-sequence-matching %s" % pipes.quote(str(args.graph_sequence_matching))
if args.klib_sequence_matching:
commandline += " --klib-sequence-matching %s" % pipes.quote(str(args.klib_sequence_matching))
if args.kmer_sequence_matching:
commandline += " --kmer-sequence-matching %s" % pipes.quote(str(args.kmer_sequence_matching))
if int(args.bad_align_uniq_kmer_len):
commandline += " --bad-align-uniq-kmer-len %s" % pipes.quote(str(args.bad_align_uniq_kmer_len))
if args.write_alignments:
alignment_directory = os.path.join(args.output, "alignments")
os.makedirs(alignment_directory, exist_ok=True)
if not os.path.isdir(alignment_directory):
raise Exception(f"Cannot create alignment output directory: {alignment_directory}")
commandline += " --alignment-output-folder !%s" % pipes.quote(alignment_directory)
if args.infer_read_haplotypes:
commandline += " --infer-read-haplotypes"
if args.verbose:
commandline += " --log-level=info"
elif args.quiet:
commandline += " --log-level=error"
elif args.debug:
commandline += " --log-level=debug"
else:
commandline += " --log-level=warning"
grmpy_log = pipes.quote(os.path.join(args.output, "grmpy.log"))
commandline += " --log-file " + grmpy_log
commandline += " --log-async no"
commandline += " -g"
for graph in graph_files:
commandline += "\n%s" % pipes.quote(graph)
response_file = tempfile.NamedTemporaryFile(dir=args.scratch_dir, mode="wt", suffix=".txt", delete=False)
response_file.write(commandline)
response_file.flush()
commandline = args.grmpy + " --response-file=%s" % pipes.quote(response_file.name)
logging.info("Starting: %s", commandline)
subprocess.check_call(commandline, shell=True, stderr=subprocess.STDOUT)
except Exception: # pylint: disable=W0703
traceback.print_exc(file=LoggingWriter(logging.ERROR))
raise
try:
if args.input.endswith("vcf") or args.input.endswith("vcf.gz"):
grmpyOutput = vcfupdate.read_grmpy(result_json_path)
result_vcf_path = os.path.join(args.output, "genotypes.vcf.gz")
vcf_input_path = os.path.join(args.output, "variants.vcf.gz")
if not os.path.exists(vcf_input_path) or not os.path.isfile(vcf_input_path):
vcf_input_path = args.input
vcfupdate.update_vcf_from_grmpy(vcf_input_path, grmpyOutput, result_vcf_path)
except Exception: # pylint: disable=W0703
traceback.print_exc(file=LoggingWriter(logging.ERROR))
raise