convert_euc2utf.py

#!/usr/bin/env python2.7
# -*- coding: utf-8 -*-

import re
import sys

NUMERIC_CODE_PATTERN = re.compile(r'&#(\d+);')


def euc_to_utf(string):
    '''
    euc_to_utf('쿠보즈카 슌스케')
    쿠보즈카 슌스케
    '''

    if string is None or 0 == len(string.strip()):
        return None

    string = (unicode(string, 'cp949')).encode('utf-8')
    numeric_code_patterns = re.findall(NUMERIC_CODE_PATTERN, string)
    # print numeric_code_patterns

    for pat in numeric_code_patterns:
        string = string.replace('&#%s;' % pat, unichr(int(pat)).encode('utf8'))
    return string


if __name__ == '__main__':
    for buf in sys.stdin:
        line = buf.rstrip()
        utf_string = euc_to_utf(line)
        print utf_string