Mercurial > repos > tomnl > frag4feature

annotate msp_split.py @ 29:15228fa5b475 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit 1800a3818988f21ef129e500818c9a087fce5875
author tomnl
date Wed, 18 Jul 2018 06:01:53 -0400
parents e7bbc00ca7e3
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
27
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
1 from __future__ import print_function
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
2 import argparse
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
3 import textwrap
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
4 import os
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
5 import re
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
6 import csv
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
7 import math
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
8
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
9 def msp_split(i, o, n):
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
10 spec_total = lcount('NAME', i)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
11 spec_lim = math.ceil(spec_total/float(n))
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
12 spec_c = 0
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
13 filelist = []
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
14 header = ''
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
15 print('spec_lim', spec_lim)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
16 with open(i, 'r') as msp_in:
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
17 for i in range(1, n+1):
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
18 with open(os.path.join(o, 'file{}.msp'.format(str(i).zfill(len(str(n))))), 'w+') as msp_out:
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
19 while spec_c <= spec_lim:
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
20 if header:
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
21 msp_out.write(header)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
22 header = ''
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
23 line = msp_in.readline()
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
24
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
25 if not line:
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
26 break # end of file
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
27
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
28 if re.match('^NAME:.*$', line, re.IGNORECASE):
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
29 header = line
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
30 spec_c += 1
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
31 else:
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
32 msp_out.write(line)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
33 spec_c = 1
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
34
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
35 return filelist
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
36
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
37 def lcount(keyword, fname):
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
38 with open(fname, 'r') as fin:
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
39 return sum([1 for line in fin if keyword in line])
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
40
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
41 def main():
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
42
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
43 p = argparse.ArgumentParser(prog='PROG',
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
44 formatter_class=argparse.RawDescriptionHelpFormatter,
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
45 description='''split msp files''',
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
46 )
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
47
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
48 p.add_argument('-i', dest='i', help='msp file', required=True)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
49 p.add_argument('-o', dest='o', help='out dir', required=True)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
50 p.add_argument('-n', dest='n',)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
51
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
52
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
53 args = p.parse_args()
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
54
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
55 if not os.path.exists(args.o):
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
56 os.makedirs(args.o)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
57 print('in file', args.i)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
58 print('out dir', args.o)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
59 print('nm files', args.n)
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
60
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
61 msp_split(args.i, args.o, int(args.n))
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
62
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
63
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
64 if __name__ == '__main__':
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
65 main()
e7bbc00ca7e3 planemo upload for repository https://github.com/computational-metabolomics/mspurity-galaxy commit f79fa34772bbab836d89cf8bad52d49285409a98
tomnl
parents:
diff changeset
66