SRT字幕の誤字をAIで修正する方法｜タイムコードを変えない手順とPython例

"""Apply reviewed cue-text corrections without rewriting SRT timecodes."""
import json
import re
import sys
from pathlib import Path

TIME = r"\d{2}:\d{2}:\d{2},\d{3}"
CUE = re.compile(rf"(\d+)(\r?\n)({TIME} --> {TIME})(\r?\n)([\s\S]+)")

def unique_object(pairs):
    result = {}
    for key, value in pairs:
        if key in result:
            raise ValueError(f"duplicate correction ID: {key}")
        result[key] = value
    return result

def millis(value):
    h, m, rest = value.split(":")
    s, ms = rest.split(",")
    if int(m) > 59 or int(s) > 59:
        raise ValueError("invalid timestamp")
    return ((int(h) * 60 + int(m)) * 60 + int(s)) * 1000 + int(ms)

def patch(original, corrections):
    if not isinstance(corrections, dict):
        raise ValueError("corrections must be an object")
    parts = re.split(r"(\r?\n[ \t]*\r?\n)", original)
    seen = set()
    for index in range(0, len(parts), 2):
        if not parts[index].strip():
            continue
        match = CUE.fullmatch(parts[index])
        if not match:
            raise ValueError("unsupported or malformed SRT cue")
        cue_id, nl, timing, _, body = match.groups()
        if cue_id in seen:
            raise ValueError(f"duplicate SRT ID: {cue_id}")
        seen.add(cue_id)
        start, end = timing.split(" --> ")
        if millis(start) >= millis(end):
            raise ValueError(f"invalid duration: {cue_id}")
        if cue_id in corrections:
            value = corrections[cue_id]
            if not isinstance(value, str) or not value.strip():
                raise ValueError(f"empty correction: {cue_id}")
            value = value.replace("\r\n", "\n")
            if "\r" in value or re.search(r"\n[ \t]*\n|-->|```", value):
                raise ValueError(f"invalid correction: {cue_id}")
            tail = nl if body.endswith(nl) else ""
            parts[index] = match.group(0)[:match.start(5)] + value.rstrip("\n").replace("\n", nl) + tail
    if not seen:
        raise ValueError("no cues")
    if set(corrections) - seen:
        raise ValueError("unknown correction IDs")
    return "".join(parts)

def main():
    source, edits, target = map(Path, sys.argv[1:])
    with source.open(encoding="utf-8-sig", newline="") as f:
        original = f.read()
    corrections = json.loads(edits.read_text(encoding="utf-8-sig"), object_pairs_hook=unique_object)
    result = patch(original, corrections)
    with target.open("x", encoding="utf-8", newline="") as f:
        f.write(result)
    print(f"saved: {target}; corrected cues: {len(corrections)}")

if __name__ == "__main__":
    main()