Mercurial > repos > nick > duplex

annotate misc/msa_sscs_matcher.py @ 18:e4d75f9efb90 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit b'4303231da9e48b2719b4429a29b72421d24310f4\n'-dirty
author nick
date Thu, 02 Feb 2017 18:44:31 -0500
parents af383638de66
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
4
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
1 #!/usr/bin/env python
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
2 from __future__ import division
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
3 import sys
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
4 import argparse
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
5
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
6 OPT_DEFAULTS = {}
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
7 USAGE = "gunzip -c families.msa.tsv.gz | %(prog)s sscs.set.fa"
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
8 DESCRIPTION = """Find the input MSA's which produced a given set of SSCS's. Pipe the full set of
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
9 MSA's to stdin and it will filter them to the matching MSA's on stdout."""
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
10
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
11
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
12 def main(argv):
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
13
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
14 parser = argparse.ArgumentParser(usage=USAGE, description=DESCRIPTION)
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
15 parser.set_defaults(**OPT_DEFAULTS)
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
16
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
17 parser.add_argument('sscs', metavar='sscs.set.fa',
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
18 help='A set of SSCS\'s, as output from the duplex.py script with the --sscs-file option.')
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
19
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
20 args = parser.parse_args(argv[1:])
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
21
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
22 sscs = set()
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
23 with open(args.sscs) as sscs_file:
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
24 for line in sscs_file:
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
25 if line.startswith('>'):
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
26 name = line.lstrip('>').split()[0]
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
27 sscs.add(name)
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
28
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
29 for line in sys.stdin:
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
30 barcode, order, mate, rname, seq, qual = line.rstrip('\r\n').split('\t')
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
31 name = '.'.join((barcode, order, mate))
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
32 if name in sscs:
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
33 sys.stdout.write(line)
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
34
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
35
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
36 if __name__ == '__main__':
af383638de66 planemo upload commit 022984f323d3da44f70b3bf79c684cfd8dda3f61-dirty
nick
parents:
diff changeset
37 sys.exit(main(sys.argv))