#!/usr/bin/python
# -*- coding: utf-8 -*-
# vim: fileencoding=utf-8

import os , sys

DIR_NAME = 'C:\\Program Files\\'
OUTPUT_NAME = 'result_file_encoding_list.txt'

TARGET_ENCODING_LIST = [
	'utf-8',
	'shift-jis',
	'euc-jp',
	'iso2022-jp'
]

FLAG_STDOUT = True
#FLAG_STDOUT = False

import os, sys

write = sys.stdout.write

def guess_charset(data):
	file = lambda d, encoding: d.decode(encoding) and encoding
	for enc in TARGET_ENCODING_LIST:
		try:
			file(data, enc)
			return enc
		except:
			pass
	return 'binary'

out = open(OUTPUT_NAME, 'w')
for dirpath, dirs, files in os.walk(DIR_NAME):
	for fn in files:
		path = os.path.join(dirpath, fn)
		fobj = file(path, 'rU')
		data = fobj.read()
		fobj.close()
		try:
			enc = guess_charset(data)
		except:
			continue
		str = path + ',' + enc + '\n'
		try:
			if FLAG_STDOUT == True:
				write(str)
			out.write(str)
		except:
			continue

Recommended Posts