Convert chinese subtitle to chinese + pinyin (Output is like this: 最zùi高gāo法fǎ院yuàn在zài今jīn天tiān早zǎo上shàng)

pinyin-subtitles.py

#!/usr/bin/python3

import sys
import re
import pysrt
import pinyin


def to_pinyin(text):
    pinyins = pinyin.get(text, delimiter=" ").split(' ')
    assert len(text) == len(pinyins)
    return "".join([f"{char}{pinyin}" for char, pinyin in zip(text, pinyins)])

def process_text(text):
    return re.sub(r'[\u4e00-\u9fff]+', lambda m: to_pinyin(m.group(0)), text)

def pinyin_subtitles(filename):
    srt = pysrt.open(filename)
    for item in srt :
        item.text = process_text(item.text)
    srt.save(filename[:-4] + '.pinyin.srt', encoding="utf-8")

def usage():
    print('./pinyin-subtitles file.srt')

def main():
    if len(sys.argv) != 2 or not sys.argv[1].endswith('.srt'):
        usage()
    else:
        pinyin_subtitles(sys.argv[1])

if __name__ == '__main__':
    main()