def main():
#do our inputs
args = do_inputs()
global of
of = sys.stdout
if args.output:
if args.output[-4:] == '.bam':
cmd = 'samtools view -Sb - -o '+args.output
p = Popen(cmd.split(),stdin=PIPE)
of = p.stdin
else:
sys.stderr.write("ERROR: stdout and .bam are the only valid output formats\n")
sys.exit()
inf = sys.stdin
if args.input != '-':
if args.input[-3:] == '.gz':
inf = gzip.open(args.input)
else: inf = open(args.input)
sys.stderr.write("reading reference genome\n")
ref = FastaData(open(args.reference).read())
#shared = manager.dict()
shared = {}
for chr in sorted(ref.keys()):
sys.stderr.write("reading "+chr+"\n")
shared[chr] = ref[chr].upper()
ref.remove(chr)
sys.stderr.write("finished reading shared memory reference\n")
sys.stderr.write("Now make the header\n")
of.write("@HD\tVN:1.0\tSO:unknown\n")
of.write("@PG\tID:SLR\n")
for chr in sorted(shared.keys()):
of.write("@SQ\tSN:"+chr+"\tLN:"+str(len(shared[chr]))+"\n")
if args.threads > 1:
poo = Pool(processes=args.threads)
buffer = []
max_buffer = 1
z = 0
for line in inf:
z += 1
if z%1000==0: sys.stderr.write(str(z)+" \r")
buffer.append(line)
if len(buffer) >= max_buffer:
if args.threads == 1:
results = do_buffer(buffer,shared,args)
do_out(results)
else:
poo.apply_async(do_buffer,args=(buffer[:],shared,args,),callback=do_out)
buffer = []
if len(buffer) > 0:
if args.threads ==1:
results = do_buffer(buffer,shared,args)
do_out(results)
else:
poo.apply_async(do_buffer,args=(buffer[:],shared,args,),callback=do_out)
if args.threads > 1:
poo.close()
poo.join()
sys.stderr.write("\n")
if args.output:
p.communicate()
else: of.close()
# Temporary working directory step 3 of 3 - Cleanup
if not args.specific_tempdir:
rmtree(args.tempdir)
def main():
#do our inputs
args = do_inputs()
global of
of = sys.stdout
if args.output:
if args.output[-4:] == '.bam':
cmd = 'samtools view -Sb - -o ' + args.output
p = Popen(cmd.split(), stdin=PIPE)
of = p.stdin
else:
sys.stderr.write(
"ERROR: stdout and .bam are the only valid output formats\n")
sys.exit()
inf = sys.stdin
if args.input != '-':
if args.input[-3:] == '.gz':
inf = gzip.open(args.input)
else:
inf = open(args.input)
sys.stderr.write("reading reference genome\n")
ref = FastaData(open(args.reference).read())
#shared = manager.dict()
shared = {}
for chr in sorted(ref.keys()):
sys.stderr.write("reading " + chr + "\n")
shared[chr] = ref[chr].upper()
ref.remove(chr)
sys.stderr.write("finished reading shared memory reference\n")
sys.stderr.write("Now make the header\n")
of.write("@HD\tVN:1.0\tSO:unknown\n")
of.write("@PG\tID:SLR\n")
for chr in sorted(shared.keys()):
of.write("@SQ\tSN:" + chr + "\tLN:" + str(len(shared[chr])) + "\n")
if args.threads > 1:
poo = Pool(processes=args.threads)
buffer = []
max_buffer = 1
z = 0
for line in inf:
z += 1
if z % 1000 == 0: sys.stderr.write(str(z) + " \r")
buffer.append(line)
if len(buffer) >= max_buffer:
if args.threads == 1:
results = do_buffer(buffer, shared, args)
do_out(results)
else:
poo.apply_async(do_buffer,
args=(
buffer[:],
shared,
args,
),
callback=do_out)
buffer = []
if len(buffer) > 0:
if args.threads == 1:
results = do_buffer(buffer, shared, args)
do_out(results)
else:
poo.apply_async(do_buffer,
args=(
buffer[:],
shared,
args,
),
callback=do_out)
if args.threads > 1:
poo.close()
poo.join()
sys.stderr.write("\n")
if args.output:
p.communicate()
else:
of.close()
# Temporary working directory step 3 of 3 - Cleanup
if not args.specific_tempdir:
rmtree(args.tempdir)
