Mercurial > repos > chrisb > gap_all_glycan_tools

comparison get_data/kegg_glycan/getkcfKEGG.py @ 0:89592faa2875 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded
author chrisb
date Wed, 23 Mar 2016 14:35:56 -0400
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:89592faa2875
1 __author__ = 'cbarnett'
2 __license__ = "MIT"
3 __version = "0.3"
4
5
6 def helper_get_G(lineentry):
7 if "G" in lineentry.upper():
8 splitline = lineentry.split()
9 for item in splitline:
10 if "G" in item:
11 return item
12 return None
13
14
15 def get_kcf_from_kegg(inputstream):
16 """
17 :param inputstream: input stream containing gl entries
18 :return: list of kcf output and list of db entry output or empty list
19 """
20 import urllib2
21
22 uri = 'http://rest.kegg.jp/get/'
23 if inputstream is None or inputstream == []:
24 raise IOError("empty input stream")
25 dbresponses = []
26 kcfresponses = []
27 for line in inputstream:
28 glentry = helper_get_G(line)
29 if glentry is not None:
30 try:
31 dbresponse = urllib2.urlopen(uri + glentry).read()
32 kcfresponse = urllib2.urlopen(uri + glentry + "/kcf").read()
33 except Exception as e:
34 raise urllib2.HTTPError(e.url, e.code, e.msg, e.hdrs, e.fp)
35 dbresponses.append(dbresponse)
36 kcfresponses.append(kcfresponse)
37 return kcfresponses, dbresponses
38
39
40 if __name__ == "__main__":
41 from optparse import OptionParser
42
43 usage = "usage: python %prog [options]\n"
44 parser = OptionParser(usage=usage)
45 parser.add_option("-i", action="store", type="string", dest="i", default="input",
46 help="single or double column text file containing GL entries")
47 parser.add_option("-k", action="store", type="string", dest="k", default="kcf.output",
48 help="kcf output file name")
49 parser.add_option("-d", action="store", type="string", dest="d", default="db.output",
50 help="KEGG db entry in text format output file name")
51 (options, args) = parser.parse_args()
52 try:
53 instream = file(options.i, 'r')
54 except Exception as e:
55 raise IOError(e, "the input file specified does not exist. Use -h flag for help")
56 kcf, db = get_kcf_from_kegg(instream)
57 try:
58 kcfout = file(options.k, 'w')
59 dbout = file(options.d, 'w')
60 except Exception as e:
61 raise IOError(e, "cannot open output files. -h flag for help")
62
63 kcfout.write("".join(kcf))
64 dbout.write("".join(db))
65 kcfout.close()
66 dbout.close()
67