trans-subs.py

trans-subs.py

import sys
import os
import codecs


# euc-kr to utf-8 with BOM
def convert(path):
    with codecs.open(path, 'r', 'euc-kr', errors='ignore') as f:
        out_lines = []
        for line in f.readlines():
            out_lines += [line]
    
    with codecs.open(path, 'w', 'utf-8') as f:
        # BOM
        f.write(u'\ufeff')
        for line in out_lines:
            f.write(line)

for r, d, fs in os.walk(sys.argv[1]):
    for f in fs:
        if f.endswith('.smi'):
            try:
                path = os.path.join(r, f)
                print(path)
                convert(path)
            except Exception as e:
                print(e)