Mercurial > repos > galaxyp > regex_find_replace

comparison regex.py @ 2:538933d9fccc 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/regex_find_replace commit 8871a847daed1f996c0a2069c3e876685bf8d220"
author galaxyp
date Tue, 07 Apr 2020 11:31:27 -0400
parents 60d04307b027
children 399da6b5ec21
comparison
equal deleted inserted replaced
1:209b7c5ee9d7 2:538933d9fccc
1 import sys 1 from optparse import OptionParser
2 import os
3 import re 2 import re
4 import string 3
5 import commands
6 from optparse import OptionParser
7 from tempfile import NamedTemporaryFile
8 4
9 def main(): 5 def main():
10 parser = OptionParser() 6 parser = OptionParser()
11 parser.add_option("--input", dest="input") 7 parser.add_option("--input", dest="input")
12 parser.add_option("--output", dest="output") 8 parser.add_option("--output", dest="output")
13 parser.add_option("--input_display_name", dest="input_display_name") 9 parser.add_option("--input_display_name", dest="input_display_name")
14 parser.add_option("--pattern", dest="patterns", action="append", 10 parser.add_option("--pattern", dest="patterns", action="append",
15 help="regex pattern for replacement") 11 help="regex pattern for replacement")
16 parser.add_option("--replacement", dest="replacements", action="append", 12 parser.add_option("--replacement", dest="replacements", action="append",
17 help="replacement for regex match") 13 help="replacement for regex match")
18 parser.add_option("--column", dest="column", default=None) 14 parser.add_option("--column", dest="column", default=None)
19 (options, args) = parser.parse_args() 15 (options, args) = parser.parse_args()
20 16
21 mapped_chars = { '\'' :'__sq__', '\\' : '__backslash__' } 17 mapped_chars = {'\'': '__sq__', '\\': '__backslash__'}
22 18
23 column = None 19 column = None
24 if options.column is not None: 20 if options.column is not None:
25 column = int(options.column) - 1 # galaxy tabular is 1-based, python array are zero-based 21 # galaxy tabular is 1-based, python array are zero-based
22 column = int(options.column) - 1
26 23
27 with open(options.input, 'r') as input: 24 with open(options.input, 'r') as input, open(options.output, 'w') as output:
28 with open(options.output, 'w') as output: 25 while True:
29 while True: 26 line = input.readline()
30 line = input.readline() 27 if line == "":
31 if line == "": 28 break
32 break 29 for (pattern, replacement) in zip(options.patterns, options.replacements):
33 for (pattern, replacement) in zip(options.patterns, options.replacements): 30 for key, value in mapped_chars.items():
34 for key, value in mapped_chars.items(): 31 pattern = pattern.replace(value, key)
35 pattern = pattern.replace(value, key) 32 replacement = replacement.replace(value, key)
36 replacement = replacement.replace(value, key) 33 replacement = replacement.replace("#{input_name}", options.input_display_name)
37 replacement = replacement.replace("#{input_name}", options.input_display_name) 34 if column is None:
38 if column is None: 35 line = re.sub(pattern, replacement, line)
39 line = re.sub(pattern, replacement, line) 36 else:
40 else: 37 cells = line.split("\t")
41 cells = line.split("\t") 38 if cells and len(cells) > column:
42 if cells and len(cells) > column: 39 cell = cells[column]
43 cell = cells[column] 40 cell = re.sub(pattern, replacement, cell)
44 cell = re.sub(pattern, replacement, cell) 41 cells[column] = cell
45 cells[column] = cell 42 line = "\t".join(cells)
46 line = "\t".join(cells) 43 output.write(line)
47 output.write(line) 44
48 45
49 if __name__ == "__main__": 46 if __name__ == "__main__":
50 main() 47 main()