srt parser

import re
from typing import List
import json

class Subtitle:
    def __init__(self, start: float, end: float, text: str):
        self.start = start
        self.end = end
        self.text = text

def parse_srt(file_path: str) -> List[Subtitle]:
    def timestamp_to_seconds(timestamp: str) -> float:
        hours, minutes, seconds = re.split("[:,]", timestamp)
        return int(hours) * 3600 + int(minutes) * 60 + float(seconds)

    with open(file_path, 'r', encoding='utf-8') as f:
        content = f.read()

    blocks = re.split(r'\n\s*\n', content.strip())
    subtitles = []

    for block in blocks:
        lines = block.strip().splitlines()
        if len(lines) >= 3:
            timing_line = lines[1]
            start_str, end_str = re.split(r'\s*-->\s*', timing_line)
            start = timestamp_to_seconds(start_str.replace(',', '.'))
            end = timestamp_to_seconds(end_str.replace(',', '.'))
            text = '\\n'.join(lines[2:])
            subtitles.append(Subtitle(start, end, text))

    return subtitles