Offset and scale SRT translation file

srt.py

from dataclasses import dataclass
from datetime import datetime, timedelta


TIME_FORMAT = '%H:%M:%S,%f'


def load_time(s: str) -> float:
    t = datetime.strptime(s, TIME_FORMAT)
    return (t - t.replace(hour=0, minute=0, second=0, microsecond=0)).total_seconds()


def dump_time(s: float) -> str:
    t = datetime.now().replace(hour=0, minute=0, second=0, microsecond=0)
    return (t + timedelta(seconds=s)).time().strftime(TIME_FORMAT)[:-3]


@dataclass
class Subtitle:
    index: int
    start: float
    finish: float
    text: list[str]

    def __str__(self):
        return (
            f'{self.index}\n'
            f'{dump_time(self.start)} --> {dump_time(self.finish)}\n'
        ) + '\n'.join(self.text)


def iter_subtitles(file):
    while True:
        # '1\n'
        try:
            index = int(next(file)[:-1])
        except StopIteration:
            return
        # '00:01:38,246 --> 00:01:39,855\n'
        start, finish = next(file)[:-1].split(' --> ')
        start, finish = load_time(start), load_time(finish)
        # "We're screwed.\n"
        # '\n'
        text = []
        while True:
            try:
                line = next(file)
            except StopIteration:
                break
            if len(line) == 1:
                break
            text.append(line[:-1])
        yield Subtitle(index, start, finish, text)


def main(path, start, finish):
    with open(path, 'r') as file:
        subtitles = list(iter_subtitles(file))
    subtitles_start = subtitles[0].start
    scale = (finish - start) / (subtitles[-1].finish - subtitles_start)
    for sub in subtitles:
        sub.start = (sub.start - subtitles_start) * scale + start
        sub.finish = (sub.finish - subtitles_start) * scale + start
        print(sub)
        print('')


if __name__ == '__main__':
    import argparse
    parser = argparse.ArgumentParser()
    parser.add_argument('path', help='Path of SRT file to read')
    parser.add_argument('start', type=load_time, help='Time when the first word is started (HH:MM:SS,fff)')
    parser.add_argument('finish', type=load_time, help='Time when the last word is finished (HH:MM:SS,fff)')
    main(**vars(parser.parse_args()))