Mercurial > repos > bcclaywell > argo_navis

annotate bin/metadata_from_seqnames.py @ 0:d67268158946 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
author bcclaywell
date Mon, 12 Oct 2015 17:43:33 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
1 #!/usr/bin/env python
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
2 """Little script for parsing metadata out of sequence names given regular expressions. Supports parsing out
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
3 deme information and data information."""
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
4
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
5 import argparse
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
6 import csv
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
7 import re
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
8 from Bio import SeqIO
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
9
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
10
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
11 def get_args():
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
12 parser = argparse.ArgumentParser(description=__doc__)
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
13 parser.add_argument('inseqs', help="Input sequences in fasta format")
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
14 parser.add_argument('-d', '--deme-regex', required=True, type=re.compile,
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
15 help="Regular expression with which to parse deme information")
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
16 parser.add_argument('-t', '--time-regex', type=re.compile,
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
17 help="Regular expression with which to parse date information")
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
18 parser.add_argument('output', type=argparse.FileType('w'))
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
19 return parser.parse_args()
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
20
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
21
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
22 def main():
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
23 args = get_args()
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
24 seqreader = SeqIO.parse(args.inseqs, 'fasta')
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
25
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
26 header = ['sequence', 'deme']
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
27 if args.time_regex:
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
28 header.append('date')
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
29
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
30 outwriter = csv.DictWriter(args.output, header)
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
31 outwriter.writeheader()
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
32 for seqrec in seqreader:
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
33 seqname = seqrec.id
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
34 try:
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
35 deme = args.deme_regex.match(seqname).groups()[0]
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
36 except Exception:
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
37 raise Exception, "There was a problem parsing deme information for sequence %s. Try again." % seqname
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
38 rowdict = dict(sequence=seqname, deme=deme)
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
39 if args.time_regex:
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
40 try:
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
41 rowdict['date'] = args.time_regex.match(seqname).groups()[0]
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
42 except Exception:
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
43 raise Exception, "There was a problem parsing date information for sequence %s. Try again." % seqname
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
44 outwriter.writerow(rowdict)
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
45
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
46 args.output.close()
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
47
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
48
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
49 if __name__ == '__main__':
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
50 main()
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
51
d67268158946 planemo upload commit a3f181f5f126803c654b3a66dd4e83a48f7e203b
bcclaywell
parents:
diff changeset
52