translate_srt.py

# -*- coding: utf-8 -*-
import re
import sys


NEWLINE = '\r\n'
TIME_STR = re.compile(r'\d{2}\:\d{2}\:\d{2}[\,\.]\d{0,3}\s*--\>\s*\d{2}\:\d{2}\:\d{2}[\,\.]\d{0,3}')
TAGS_PATTERN = re.compile(r'\<[^\>]*\>')


def strip_tags(text):
    return TAGS_PATTERN.sub(u'', text or u'').strip()


def translate_srt(from_lang, to_lang, input_file, output):
    import mtranslate
    script = open(input_file, 'rb').read().split(NEWLINE + NEWLINE)

    chunk1 = []
    chunk2 = []
    for i, item in enumerate(script):
        lines = item.replace('\xe2\x99\xaa', '\'').split(NEWLINE)
        chunk1.append(NEWLINE.join(lines[:2]))
        chunk2.append(strip_tags('|'.join(lines[2:])))
        if len(chunk1) >= 50 or i == len(script) - 1:
            translated = mtranslate.translate(
                ' ||| '.join(chunk2), to_language=to_lang, from_language=from_lang)
            translated = translated.encode('utf-8').split('|||')
            # Try to fix
            translated += ['...'] * (len(chunk1) - len(translated))
            for i1, i2 in zip(chunk1, translated):
                output.write(i1)
                output.write(NEWLINE)
                output.write(NEWLINE.join(s.strip() for s in i2.split('|')))
                output.write(NEWLINE)
                output.write(NEWLINE)
                output.flush()
            chunk1 = []
            chunk2 = []


if __name__ == '__main__':
    if len(sys.argv) != 4:
        sys.stderr.write('''Usage:

    python translate_srt.py <from_language> <to_language> input_file
''')
        sys.exit(1)

    translate_srt(sys.argv[1], sys.argv[2], sys.argv[3], sys.stdout)
    sys.exit(0)