Mercurial > repos > jjohnson > contig_annotation_tool

annotate tabpad.py @ 0:ad7507073c3f draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
author jjohnson
date Sun, 24 Nov 2019 21:56:00 -0500
parents
children 9a01840eac52
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
1 #!/usr/bin/env python
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
2
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
3 import argparse
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
4 import re
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
5
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
6
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
7 def padfile(infile, outfile, fieldcnt=None):
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
8 with open(infile, 'r') as fh:
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
9 out = open(outfile, 'w')
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
10 tabs = '\t' * fieldcnt if fieldcnt is not None else None
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
11 for i, txtline in enumerate(fh):
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
12 line = txtline.rstrip('\r\n')
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
13 fields = line.split('\t')
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
14 if not tabs:
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
15 tabs = '\t' * len(fields)
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
16 out.write('%s%s\n' % (line, tabs[len(fields):]))
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
17 out.close()
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
18
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
19
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
20 def fieldcount(infile):
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
21 fieldcnt = 0
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
22 with open(infile, 'r') as fh:
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
23 for i, line in enumerate(fh):
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
24 fieldcnt = max(fieldcnt, len(line.rstrip('\r\n').split('\t')))
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
25 return fieldcnt
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
26
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
27
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
28 def tsvname(infile):
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
29 return re.sub('\.txt$', '', infile) + '.tsv'
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
30
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
31
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
32 def __main__():
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
33 parser = argparse.ArgumentParser(
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
34 description='Pad a file with TABS for equal field size across lines')
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
35 parser.add_argument(
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
36 '-i', '--input', help='input file')
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
37 parser.add_argument(
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
38 '-o', '--output', help='output file')
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
39 parser.add_argument(
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
40 'files', nargs='*', help='.txt files')
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
41 args = parser.parse_args()
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
42
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
43 if args.input:
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
44 outfile = args.output if args.output else tsvname(args.input)
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
45 fieldcnt = fieldcount(args.input)
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
46 padfile(args.input, outfile, fieldcnt=fieldcnt)
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
47 for infile in args.files:
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
48 outfile = tsvname(infile)
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
49 fieldcnt = fieldcount(infile)
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
50 padfile(infile, outfile, fieldcnt=fieldcnt)
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
51
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
52
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
53 if __name__ == "__main__":
ad7507073c3f planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cat commit f80f020c77d04c2e13b89aaea3d784314b940931-dirty
jjohnson
parents:
diff changeset
54 __main__()