fuzzy search regexp generator

fuzzy_search.py

# usage: 
# LC_ALL=C grep "$(python fuzzy_search.py "HELLO")"  input.file.utf8.txt

import sys

s = sys.argv[1]

su = s.decode('utf-8')

utf8_any = '[^\x80-\xbf][\x80-\xbf]*'
#utf8_any = '.'

vars = []

def escape_char(c):
    if c in '.[]*\\$^':
        return '\\' + c
    return c

def esc(r):
    return ''.join(map(escape_char, r)).encode('UTF-8')

for i in xrange(len(su)):
    vars.append(esc(su[:i]) + utf8_any + esc(su[i+1:]))
    vars.append(esc(su[:i]) + esc(su[i+1:]))
    if (i != 0):
        vars.append(esc(su[:i]) + utf8_any + esc(su[i:]))

sys.stdout.write('\\|'.join(vars))

Why not use re.escape() instead of escape_char()?

Why not use re.escape() instead of escape_char()?

grep regexes have different syntax from P(y)CRE, and have different set of chars which are "special". Symbol '|' is not special, but if it will be escaped '|' - it will have special meaning for grep.

See: sys.stdout.write('|'.join(vars)). '|' is escaped, to be 'special'.

grep is used instead python re, because grep is much faster on large regexes and large input files.

@dobrokot, makes sense, thanks