Mercurial > repos > dannon > metaphlan

comparison metaphlan2krona.py @ 6:e951f9d38339 default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Added metaphlan2krona
author Dannon Baker <dannonbaker@me.com>
date Tue, 08 Apr 2014 14:16:46 -0400
parents
children
comparison
equal deleted inserted replaced
5:73f082e9fa2d 6:e951f9d38339
1 #!/usr/bin/env python
2
3 # ==============================================================================
4 # Conversion script: from MetaPhlAn output to Krona text input file
5 # Author: Daniel Brami (daniel.brami@gmail.com)
6 # ==============================================================================
7
8 import sys
9 import optparse
10 import re
11
12 def main():
13 #Parse Command Line
14 parser = optparse.OptionParser()
15 parser.add_option( '-p', '--profile', dest='profile', default='', action='store', help='The input file is the MetaPhlAn standard result file' )
16 parser.add_option( '-k', '--krona', dest='krona', default='krona.out', action='store', help='the Krona output file name' )
17 ( options, spillover ) = parser.parse_args()
18
19 if not options.profile or not options.krona:
20 parser.print_help()
21 sys.exit()
22
23 re_candidates = re.compile(r"s__|unclassified\t")
24 re_replace = re.compile(r"\w__")
25 re_bar = re.compile(r"\|")
26
27 metaPhLan = list()
28 with open(options.profile,'r') as f:
29 metaPhLan = f.readlines()
30 f.close()
31
32 krona_tmp = options.krona
33 metaPhLan_FH = open(krona_tmp, 'w')
34
35 for aline in (metaPhLan):
36 if(re.search(re_candidates, aline)):
37 x=re.sub(re_replace, '\t', aline)
38 x=re.sub(re_bar, '', x)
39
40 x_cells = x.split('\t')
41 lineage = '\t'.join(x_cells[0:(len(x_cells) -1)])
42 abundance = float(x_cells[-1].rstrip('\n'))
43
44 metaPhLan_FH.write('%s\n'%(str(abundance) + '\t' + lineage))
45
46 metaPhLan_FH.close()
47
48 if __name__ == '__main__':
49 main()