def split_sam_by_chr(inSam, prefix):
if not isinstance(inSam, IOBase):
inSam = open(inSam)
header = ''
d_chr = {}
i = 0
j = 0
outfiles = []
for line in inSam:
if line.startswith('@'):
header += line
continue
j += 1
temp = line.split('\t')
chrom = temp[2]
if chrom in d_chr:
hname = 'f%s' % (d_chr[chrom], )
else:
d_chr[chrom] = i + 1
outfile = '%s.%s.sam' % (prefix, d_chr[chrom])
outfiles += [outfile]
hname = 'f%s' % (d_chr[chrom], )
exec('%s = open("%s", "w")' % (hname, outfile))
exec('%s.write(header)' % (hname, ))
i += 1
exec('%s.write(line)' % (hname, ))
# close files
for chrom in list(d_chr.keys()):
hname = 'f%s' % (d_chr[chrom], )
exec('%s.close()' % (hname, ))
return (j, i, j / i, outfiles)
def split_sam_by_chunk_num(inSam, prefix, chunk_num):
if not isinstance(inSam, IOBase):
inSam = open(inSam)
# open files
outfiles = []
for chunk_id in range(chunk_num):
chunk_id += 1
outfile = '%s.%s.sam' % (prefix, chunk_id)
outfiles += [outfile]
hname = 'f%s' % (chunk_id, )
exec('%s = open("%s", "w")' % (hname, outfile))
i = 0
header = ''
for line in inSam:
if line.startswith('@'):
header += line
continue
if i == 0:
for chunk_id in range(chunk_num):
chunk_id += 1
hname = 'f%s' % (chunk_id, )
exec('%s.write(header)' % (hname, ))
chunk_id = i % chunk_num + 1
hname = 'f%s' % (chunk_id, )
exec('%s.write(line)' % (hname, ))
i += 1
# close files
for chunk_id in range(chunk_num):
chunk_id += 1
exec('f%s.close()' % (chunk_id, ))
return (i, chunk_num, i / chunk_num, outfiles)
def split_fastx_by_size(inFastx,
prefix,
chunk_num,
seqfmt,
suffix,
out_random=True):
import binpacking
d_seq = {}
d_len = {}
for rc in SeqIO.parse(inFastx, seqfmt):
d_seq[rc.id] = rc
d_len[rc.id] = len(rc.seq)
bins = binpacking.to_constant_bin_number(d_len, chunk_num)
i = 0
j = 0
outfiles = []
if out_random:
import random
random.shuffle(bins)
for d_bin in bins:
chunk_id = i + 1
out_file = '%s.%s.%s%s' % (prefix, chunk_id, seqfmt, suffix)
outfiles += [out_file]
f = open(out_file, 'w')
for id in list(d_bin.keys()):
j += 1
SeqIO.write(d_seq[id], f, seqfmt)
f.close()
i += 1
return (j, chunk_num, j / chunk_num, outfiles)
def split_sam_by_chunk_size(inSam, prefix, chunk_size):
if not isinstance(inSam, IOBase):
inSam = open(inSam)
i = 0
j = 0
outfiles = []
header = ''
for line in inSam:
if line.startswith('@'):
header += line
continue
chunk_id = i / chunk_size + 1
hname = 'f%s' % (chunk_id, )
if hname in dir():
pass
else:
last_hname = 'f%s' % (chunk_id - 1, )
if last_hname in dir():
exec('%s.close()' % (last_hname, ))
outfile = '%s.%s.sam' % (prefix, chunk_id)
outfiles += [outfile]
exec('%s = open("%s", "w")' % (hname, outfile))
exec('%s.write(header)' % (hname, ))
j += 1
exec('%s.write(line)' % (hname, ))
i += 1
# close files
exec('%s.close()' % (hname, ))
return (i, j, chunk_size, outfiles)
def split_fastx_by_chunk_size(inFastx, prefix, chunk_size, seqfmt, suffix):
if not isinstance(inFastx, IOBase):
inFastx = open(inFastx)
i = 0
j = 0
outfiles = []
header = ''
for rc in parse_fastx(inFastx):
chunk_id = i / chunk_size + 1
hname = 'f%s' % (chunk_id, )
if hname in dir():
pass
else:
last_hname = 'f%s' % (chunk_id - 1, )
if last_hname in dir():
exec('%s.close()' % (last_hname, ))
outfile = '%s.%s.%s%s' % (prefix, chunk_id, seqfmt, suffix)
outfiles += [outfile]
exec('%s = open("%s", "w")' % (hname, outfile))
j += 1
exec('%s.write(rc)' % (hname, ))
i += 1
# close files
exec('%s.close()' % (hname, ))
return (i, j, chunk_size, outfiles)
def split_paf_by_chr(inPaf, prefix, suffix=''):
if not isinstance(inPaf, IOBase):
inPaf = open(inPaf)
d_chr = {}
i = 0
j = 0
outfiles = []
for line in inPaf:
j += 1
temp = line.split('\t')
chrom = temp[5]
if chrom in d_chr:
hname = 'f%s' % (d_chr[chrom], )
else:
d_chr[chrom] = i + 1
outfile = '%s.%s.paf%s' % (prefix, chrom, suffix)
outfiles += [outfile]
hname = 'f%s' % (d_chr[chrom], )
exec('%s = open("%s", "w")' % (hname, outfile))
i += 1
exec('%s.write(line)' % (hname, ))
# close files
for chrom in list(d_chr.keys()):
hname = 'f%s' % (d_chr[chrom], )
exec('%s.close()' % (hname, ))
return (j, i, j / i, outfiles)
def main():
opts, args = getopt.getopt(sys.argv[1:], 'hi:a:o:t:g:k:f:s:x:d')
input_file = ''
output_file = ''
in_accnos = sys.stdin
type = 'table'
process = 'get'
col = 1
head = 1
accnos_sep = None
sep = "\t"
dedup = False
for op, value in opts:
if op == '-i':
input_file = value
elif op == '-o':
output_file = value
elif op == '-a':
in_accnos = open(value)
elif op == '-t':
type = value
elif op == '-g':
process = value
elif op == '-k':
col = int(value)
elif op == '-f':
head = int(value)
elif op == '-x':
accnos_sep = value
elif op == '-s':
sep = value
elif op == '-d':
dedup = True
elif op == '-h':
usage()
sys.exit()
if type not in {'table', 'fasta', 'fastq', 'hmm', 'genbank'}:
raise TypeError(
"type must be one of ['table','fasta','fastq'], unexpected '%s'" %
(type, ))
usage()
sys.exit()
if process not in ['get', 'remove']:
raise TypeError(
"process must be one of ['get','remove'], unexpected '%s'" %
(process, ))
usage()
sys.exit()
get_records(input_file,
output_file,
in_accnos,
type=type,
process=process,
col=col,
head=head,
accnos_sep=accnos_sep,
sep=sep,
dedup=dedup)
def bin_split_fastx_by_chunk_num(inFastx,
prefix,
chunk_num,
seqfmt,
suffix,
window_size=1e6,
window_ovl=1e5,
tmpdir='/tmp'):
window_size, window_ovl = int(window_size), int(window_ovl)
cutSeq = '{}/cut.{}'.format(tmpdir, seqfmt)
with open(cutSeq, 'w') as f:
cut_seqs(inFastx,
f,
window_size=window_size,
window_ovl=window_ovl,
seqfmt=seqfmt)
return split_fastx_by_size(cutSeq, prefix, chunk_num, seqfmt, suffix)
def split_fastx_by_chr(inFastx, prefix, seqfmt, suffix=''):
if not isinstance(inFastx, IOBase):
inFastx = open(inFastx)
i = 0
j = 0
outfiles = []
for rc in parse_fastx(inFastx):
j += 1
last_hname = 'f%s' % (i - 1, )
if last_hname in dir():
exec('%s.close()' % (last_hname, ))
chrom = rc.split()[0][1:]
outfile = '%s.%s.%s%s' % (prefix, chrom, seqfmt, suffix)
outfiles += [outfile]
hname = 'f%s' % (i, )
exec('%s = open("%s", "w")' % (hname, outfile))
i += 1
exec('%s.write(rc)' % (hname, ))
# close files
exec('%s.close()' % (hname, ))
return (j, i, j / i, outfiles)
def split_fastx_by_chunk_num(inFastx, prefix, chunk_num, seqfmt, suffix):
if not isinstance(inFastx, IOBase):
inFastx = open(inFastx)
# open files
outfiles = []
for chunk_id in range(chunk_num):
chunk_id += 1
outfile = '%s.%s.%s%s' % (prefix, chunk_id, seqfmt, suffix)
outfiles += [outfile]
hname = 'f%s' % (chunk_id, )
exec('%s = open("%s", "w")' % (hname, outfile))
i = 0
for rc in parse_fastx(inFastx):
chunk_id = i % chunk_num + 1
hname = 'f%s' % (chunk_id, )
exec('%s.write(rc)' % (hname, ))
i += 1
# close files
for chunk_id in range(chunk_num):
chunk_id += 1
exec('f%s.close()' % (chunk_id, ))
return (i, chunk_num, i / chunk_num, outfiles)
def main():
parser = argparse.ArgumentParser(version=__version__)
parser.add_argument("-i",
"--input",
action="store",
type=str,
dest="input",
default=sys.stdin,
help="input [default=%(default)s]")
parser.add_argument("--prefix",
action="store",
dest="prefix",
default='chunk',
help="output prefix [default=%(default)s]")
parser.add_argument("-n",
"--chunk-number",
action="store",
type=int,
dest="chunk_num",
default=None,
help="number of chunk [default=%(default)s]")
parser.add_argument("-s",
"--chunk-size",
action="store",
type=int,
dest="chunk_size",
default=None,
help='size of chunk [default=%(default)s]')
parser.add_argument("-f",
"--format",
action="store",
dest="rcfmt",
default='fasta',
choices=['fasta', 'fastq', 'fastx', 'sam', 'paf'],
help="record file format [default=%(default)s]")
parser.add_argument("--gzip-output",
action="store_true",
dest="gzip_output",
default=False,
help="if gzip output [default=%(default)s]")
parser.add_argument(
"--by-size",
action="store_true",
dest="by_size",
default=False,
help=
'split by size (binpacking, only for fastx format) [default=%(default)s]'
)
parser.add_argument("--by-chrom",
action="store_true",
dest="by_chr",
default=False,
help='split by mapped chromsome [default=%(default)s]')
parser.add_argument("-pfn",
"--print-filenames",
action="store_true",
dest="print_filenames",
default=False,
help='print filenames [default=%(default)s]')
#parser.print_help()
options = parser.parse_args()
if len(sys.argv) == 1:
parser.print_help(sys.stderr)
sys.exit()
if options.chunk_num and options.chunk_size:
parser.print_help(sys.stderr)
print('chunk-number and chunk-size is not compatible.',
file=sys.stderr)
sys.exit()
elif not (options.chunk_num or options.chunk_size) and not options.by_chr:
parser.print_help(sys.stderr)
print('either chunk-number or chunk-size must be speicfied.',
file=sys.stderr)
sys.exit()
if options.gzip_output:
suffix = '.gz'
print(
'Warning: for large dataset, gzip output is very slow. You may want to diable it.',
file=sys.stderr)
else:
suffix = ''
if not isinstance(options.input, IOBase):
options.input = open(options.input)
#execute
if options.rcfmt == 'sam':
if options.chunk_num:
stats = split_sam_by_chunk_num(options.input, options.prefix,
options.chunk_num)
elif options.chunk_size:
stats = split_sam_by_chunk_size(options.input, options.prefix,
options.chunk_size)
elif options.by_chr:
stats = split_sam_by_chr(
options.input,
options.prefix,
)
elif options.rcfmt in set(['fasta', 'fastq', 'fastx']):
if options.chunk_num:
if options.by_size:
stats = split_fastx_by_size(options.input, options.prefix,
options.chunk_num, options.rcfmt,
suffix)
else:
stats = split_fastx_by_chunk_num(options.input, options.prefix,
options.chunk_num,
options.rcfmt, suffix)
elif options.chunk_size:
stats = split_fastx_by_chunk_size(options.input, options.prefix,
options.chunk_size,
options.rcfmt, suffix)
elif options.by_chr:
stats = split_fastx_by_chr(options.input, options.prefix,
options.rcfmt, suffix)
elif options.rcfmt == 'paf':
if options.by_chr:
stats = split_paf_by_chr(options.input, options.prefix, suffix)
n_records, n_chunks, per_chunk, outfiles = stats
print('total %s records, splited into %s chunks, %s per chunk' %
(n_records, n_chunks, per_chunk),
file=sys.stderr)
if options.print_filenames:
print('\n'.join(outfiles), file=sys.stdout)
def get_records(input_file,
output_file,
in_accnos,
type='table',
process='get',
sep="\t",
col=1,
head=1,
accnos_sep=None,
dedup=False):
def get_record(d_accnos, record_id):
if record_id in d_accnos:
return True
else:
return False
def remove_record(d_accnos, record_id):
if record_id in d_accnos:
return False
else:
return True
if isinstance(in_accnos, IOBase):
d_accnos = {
line.strip().split(accnos_sep)[0]
for line in in_accnos if line.strip()
}
else: # list
d_accnos = set(in_accnos)
lst_get = set([])
f = open(output_file, 'w')
if type == 'table':
i = 0
for line in open(input_file, 'r'):
i += 1
if i == head:
f.write(line)
else:
temp = line.strip().split(sep)
record_id = temp[col - 1]
if process == 'get':
if get_record(d_accnos, record_id):
if dedup and record_id in lst_get:
continue
f.write(line)
lst_get.add(record_id)
else:
continue
elif process == 'remove':
if remove_record(d_accnos, record_id):
f.write(line)
else:
lst_get.add(record_id)
continue
elif type in {'fasta', 'fastq', 'genbank'}:
for seq_record in SeqIO.parse(open(input_file), type):
record_id = seq_record.id
if process == 'get':
if get_record(d_accnos, record_id):
SeqIO.write(seq_record, f, type)
lst_get.add(record_id)
else:
continue
elif process == 'remove':
if remove_record(d_accnos, record_id):
SeqIO.write(seq_record, f, type)
else:
lst_get.add(record_id)
continue
elif type == 'hmm':
from HMMER import HMMParser
for rc in HMMParser(open(input_file)):
values = {getattr(rc, key, None) for key in ['NAME', 'ACC']}
values = {v for v in values if v}
if process == 'get':
if values & d_accnos:
rc.write(f)
lst_get = lst_get | values
elif process == 'remove':
if not (values & d_accnos):
rc.write(f)
lst_get = lst_get | values
f.close()
not_get = d_accnos - set(lst_get)
if not_get:
for not_get_id in not_get:
print(not_get_id)
def get_records(input_file,
output_file,
in_accnos,
type='table',
process='get',
col=1,
head=1,
accnos_sep=None):
def get_record(d_accnos, record_id):
if record_id in d_accnos:
return True
else:
return False
def remove_record(d_accnos, record_id):
if record_id in d_accnos:
return False
else:
return True
if isinstance(in_accnos, IOBase):
d_accnos = {line.strip().split(accnos_sep)[0] for line in in_accnos}
else: # list
d_accnos = set(in_accnos)
lst_get = []
f = open(output_file, 'w')
if type == 'table':
i = 0
for line in open(input_file, 'r'):
i += 1
if i == head:
f.write(line)
else:
temp = line.strip().split('\t')
record_id = temp[col - 1]
if process == 'get':
if get_record(d_accnos, record_id):
f.write(line)
lst_get.append(record_id)
else:
continue
elif process == 'remove':
if remove_record(d_accnos, record_id):
f.write(line)
else:
lst_get.append(record_id)
continue
elif type == 'fasta' or type == 'fastq':
for seq_record in SeqIO.parse(open(input_file), type):
record_id = seq_record.id
if process == 'get':
if get_record(d_accnos, record_id):
SeqIO.write(seq_record, f, type)
lst_get.append(record_id)
else:
continue
elif process == 'remove':
if remove_record(d_accnos, record_id):
SeqIO.write(seq_record, f, type)
else:
lst_get.append(record_id)
continue
elif type == 'hmm':
from HMMER import HMMParser
for rc in HMMParser(open(input_file)):
if process == 'get':
if rc.NAME in d_accnos or rc.ACC in d_accnos:
rc.write(f)
lst_get += [rc.NAME, rc.ACC]
elif process == 'remove':
if not (rc.NAME in d_accnos or rc.ACC in d_accnos):
rc.write(f)
lst_get += [rc.NAME, rc.ACC]
f.close()
not_get = d_accnos - set(lst_get)
if not_get:
for not_get_id in not_get:
print(not_get_id)