convert_to_utf8.py

#!/usr/bin/env python

import codecs
import os
from os.path import join
import re
import sys


if len(sys.argv) < 3:
    print "usage: convert_to_utf8 <path> <charset>"
    sys.exit(0)
    
srcDir = sys.argv[1]

if not os.path.isdir(srcDir):
    print "'%s' is not a directory" % srcDir
    sys.exit(0)

encoding = sys.argv[2]

try:
    codecs.lookup(encoding)
except LookupError:
    print "'%s' is not a valid encoding" % encoding
    sys.exit(0)

r = re.compile('.+\.(a4d|a4p|a4l|ini|inc|htm|html|shtml|css|js|php)$')

for root, dirs, files in os.walk(srcDir):
    if '.git' in dirs:
        dirs.remove('.git')
        
    for srcFile in [f for f in files if r.search(f)]:
        path = join(root, srcFile)
        print "Processing %s" % path
        f = open(path, 'rb')
        uni = unicode(f.read(), encoding)
        f.close()
        f = open(path, 'wb')
        f.write(uni.encode('utf-8'))
        f.close()