def picard_sortvcf(case_id, vcf_path, output_vcf, reference_fasta_dict, engine, logger):
files = vcf_path
step_dir = os.getcwd()
os.makedirs(step_dir, exist_ok=True)
srt_vcf_path = os.path.join(step_dir, output_vcf)
logger.info("picard_sortvcf_output=%s" % srt_vcf_path)
mem_bytes = os.sysconf("SC_PAGE_SIZE") * os.sysconf("SC_PHYS_PAGES")
mem_gib = mem_bytes / (1024.0 ** 3)
thread = int(multiprocessing.cpu_count()) - 2
java_heap = int(int(mem_gib) - thread / 2)
if pipe_util.already_step(step_dir, case_id + "_sortvcf", logger):
logger.info("already completed step `sortvcf` of: %s" % vcf_path)
else:
logger.info("running step `picard SortVcf` of: %s" % vcf_path)
home_dir = os.path.join("/home", getpass.getuser())
cmd = [
"java",
"-Djava.io.tmpdir=/tmp/job_tmp",
"-XX:ParallelGCThreads=" + str(thread),
"-Xmx" + str(java_heap) + "G",
"-d64",
"-jar",
os.path.join(home_dir, "tools/picard-tools/picard.jar"),
"SortVcf",
"OUTPUT=" + srt_vcf_path,
"SEQUENCE_DICTIONARY=" + reference_fasta_dict,
]
for i in vcf_path:
cmd.extend(["I=" + i])
output = pipe_util.do_command(cmd, logger)
metrics = time_util.parse_time(output)
met = SRT(
case_id=case_id,
tool="picard_sortvcf",
files=files,
systime=metrics["system_time"],
usertime=metrics["user_time"],
elapsed=metrics["wall_clock"],
cpu=metrics["percent_of_cpu"],
max_resident_time=metrics["maximum_resident_set_size"],
)
postgres.create_table(engine, met)
postgres.add_metrics(engine, met)
pipe_util.create_already_step(step_dir, case_id + "_SortVcf", logger)
logger.info("completed running step sortvcf of: %s" % vcf_path)
return srt_vcf_path
