Mercurial > repos > pjbriggs > amplicon_analysis_pipeline

comparison relabel_fasta.py @ 5:bbfc9638ba84 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
First version with (partial) bioconda deps.
author pjbriggs
date Wed, 13 Jun 2018 08:39:26 -0400
parents
children
comparison
equal deleted inserted replaced
4:013bf1e2cc8f 5:bbfc9638ba84
1 #!/usr/bin/env python
2
3 """
4 Replace FASTA labels with new labels <PREFIX>1, <PREFIX>2,
5 <PREFIX>3 etc (where <PREFIX> is a user-provided argument).
6
7 Can be used to label OTUs as OTU_1, OTU_2 etc.
8
9 This is a reimplementation of the fasta_number.py script from
10 https://drive5.com/python/fasta_number_py.html
11 """
12
13 import argparse
14
15 def relabel_fasta(fp,prefix,include_size=False):
16 """
17 """
18 # Iterate over lines in file
19 nlabel = 0
20 for line in fp:
21 # Strip trailing newlines
22 line = line.rstrip('\n')
23 if not line:
24 # Skip blank lines
25 continue
26 elif line.startswith('>'):
27 #
28 nlabel += 1
29 label = line[1:].strip()
30 if args.needsize:
31 # Extract size from the label
32 try:
33 size = filter(
34 lambda x: x.startswith("size="),
35 label.split(';'))[0]
36 except Exception as ex:
37 raise Exception("Couldn't locate 'size' in "
38 "label: %s" % label)
39 yield ">%s%d;%s" % (args.prefix,
40 nlabel,
41 size)
42 else:
43 yield ">%s%d" % (args.prefix,
44 nlabel)
45 else:
46 # Echo the line to output
47 yield line
48
49 if __name__ == "__main__":
50
51 # Set up command line parser
52 p = argparse.ArgumentParser()
53 p.add_argument("--needsize",action="store_true")
54 p.add_argument("--nosize",action="store_true")
55 p.add_argument("fasta")
56 p.add_argument("prefix")
57
58 # Process command line
59 args = p.parse_args()
60
61 # Relabel FASTA
62 with open(args.fasta,'rU') as fasta:
63 for line in relabel_fasta(fasta,
64 args.prefix,
65 include_size=args.needsize):
66 print line
67
68