Mercurial > repos > recetox > filter_compounds

comparison filter_compounds.py @ 0:987357c6941c draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/RECETOX/galaxytools/tree/master/tools/filter_compounds commit ad6ca51c5fee52d533cfd6b8ccef5e44027ed298"
author recetox
date Thu, 07 Jan 2021 11:30:13 +0000
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:987357c6941c
1 import argparse
2 import re
3
4 from openbabel import openbabel, pybel
5 openbabel.obErrorLog.StopLogging()
6
7
8 def parse_command_line():
9 parser = argparse.ArgumentParser()
10 parser.add_argument('-i', '--input', required=True, help='Input file name')
11 parser.add_argument('-o', '--output', required=True, help='Output file name')
12 parser.add_argument('-m', '--met', required=False, action='store_true', help='Remove organometallic compounds')
13 parser.add_argument('-a', '--anorg', required=False, action='store_true', help='Remove anorganic compounds')
14 return parser.parse_args()
15
16
17 def filter_compounds(args, pattern):
18 print(pattern)
19 with open(args.input, "r") as infile, open(args.output, "w") as outfile:
20 for line in infile:
21 values = line.split('\t', 1)
22
23 # check if input is list of SMILES or indexed table of SMILES
24 if values[0].isnumeric():
25 mol = pybel.readstring('smi', values[1]).write('inchi').split('/')[1] if values[1].strip() else ''
26
27 # check if both organometallic and anorganic filtering passes
28 # write original line if compound is organic without metals
29 if False not in ([bool(re.search(rf'{x}', mol)) for x in pattern]):
30 outfile.write(line)
31 else:
32 outfile.write(f'{values[0]}\t{""}\n')
33 else:
34 mol = pybel.readstring('smi', values[0]).write('inchi').split('/')[1]
35 if False not in ([bool(re.search(rf'{x}', mol)) for x in pattern]):
36 outfile.write(line)
37
38
39 def __main__():
40 """
41 Filter organometallics and/or anorganic compounds.
42 """
43 args = parse_command_line()
44
45 # check if user selected something to filter out, if not output file == input file
46 sel_pattern = []
47 if args.met is False and args.anorg is False:
48 print("No filtering selected - user did not specify what to filter out.")
49 sel_pattern = r'^[a-zA-Z]+$'
50 # select patterns for filtering
51 if args.met:
52 sel_pattern.append(r'^(?:C|N|O|P|F|S|I|B|Si|Se|Cl|Br|Li|Na|H|K|[0-9]|\.)+$')
53 if args.anorg:
54 sel_pattern.append(r'[C][^abd-z]')
55
56 filter_compounds(args, sel_pattern)
57
58
59 if __name__ == "__main__":
60 __main__()