# Copyright 2011 by Bryan Catanzaro, bcatanzaro@acm.org
# Licensed under the Creative Commons 3.0 Unported License
# http://creativecommons.org/licenses/by/3.0/

# modified 2011 by Sara McMains

# call it on a single file:
# > python thatwhich.py foo.tex
# or call it on all the tex files in a directory:
# > find . -name '*.tex' -print -exec python thatwhich {} \; > thatwhicherrors.txt
	
import re
import sys

that = ('that', re.compile(r'.{0,35}(?<=,\s)that.{0,35}'))
which = ('which', re.compile(r'.{0,35}(?<!,\s)which.{0,35}'))

#if i knew how to pipe w/in python i'd use these but instead see kluge below --sm
which2 = ('which', re.compile(r'.{0,35}(?<!\()which.{0,35}'))
which3 = ('which', re.compile(r'.{0,35}(?<!,\s\'\'\s)which.{0,35}'))

def print_errors(string, (name, pattern)):
    errors = re.findall(pattern, string)
    if not errors:
        print('"%s" checks with no errors' % name)
    else:
        print('potential "%s" errors found:' % name)
        for error in errors:
            print error
    

def main(filename):
    string = open(filename, 'r').read()
    # Make sure every comma has Exactly one whitespace character after it
    # First, replace all commas with a comma and a space
    # This covers the case where there is no space after a comma
    normalized_string = re.sub(',', ', ', string)
    # Then, replace all strings of spaces with a single space
    # This covers the case where there are multiple spaces after a comma
    normalized_string = re.sub('\s+', ' ', normalized_string)
	# (which and ,'' which are ok too --sm
    normalized_string = re.sub('\(which', ', which', normalized_string)
    normalized_string = re.sub(',\s\'\'\swhich', ', which', normalized_string)
    print_errors(normalized_string, that)
    print_errors(normalized_string, which)
    print('-'*80)
    
if __name__ == '__main__':
    main(sys.argv[1])