get_subtitles.py

def youtube_dl_id(file) -> str:
    if len(file) < 15:
        return ""
    # rename old youtube_dl format to new one: cargo install renamer; fd -tf . -x renamer '\-([\w\-_]{11})\.= [$1].' {}
    yt_id_regex = re.compile(r"-([\w\-_]{11})\..*$|\[([\w\-_]{11})\]\..*$", flags=re.M)
    file = str(file).strip()

    yt_ids = yt_id_regex.findall(file)
    if not yt_ids:
        return ""

    return utils.conform([*yt_ids[0]])[0]

def has_internal_subtitle(file):
    internal_sub = cmd(
        f"ffmpeg -hide_banner -nostdin -i {quote(str(file))} -c copy -map 0:s:0 -frames:s 1 -f null - -v 0",
        strict=False,
        shell=True,
    ).returncode
    if internal_sub == 0:
        return True


def get(args, file) -> None:
    try:
        if has_internal_subtitle(file) or get_external(file):
            return
    except Exception:
        pass

    try:
        yt_video_id = youtube_dl_id(file)
    except Exception as e:
        print(file)
        print(e)
        return

    run_subliminal = not args.youtube_only
    run_youtube = args.youtube_only  # for new videos I already have yt-dlp get the subtitle

    if run_youtube and yt_video_id:
        print(yt_video_id)
        cmd(
            (
                "yt-dlp --sub-lang 'en,EN,en.*,en-*,EN.*,EN-*eng,ENG,english,English,ENGLISH'"
                " --embed-subs --compat-options no-keep-subs --write-sub --write-auto-sub"
                " --no-download-archive --skip-download --limit-rate 10K"
                f" https://youtu.be/{yt_video_id}"
            ),
            cwd=str(Path(file).parent),
            strict=False,
        )

    if run_subliminal:
        print(f"[{file}] Downloading subtitles")
        cmd(
            "subliminal",
            "--opensubtitles",
            os.environ["OPEN_SUBTITLE_CREDENTIALS"],
            "download",
            "-l",
            "en",
            file,
            # strict=False
        )


def main() -> None:
    parser = argparse.ArgumentParser(prog="library subtitle")
    parser.add_argument("paths", nargs="*")
    parser.add_argument("--youtube-only", action="store_true")
    parser.add_argument("--subliminal-only", action="store_true")
    parser.add_argument("--verbose", "-v", action="count", default=0)
    args = parser.parse_args()

    video_files = get_media_files(args)

    Parallel(n_jobs=6 if args.verbose == 0 else 1)(delayed(get)(args, file) for file in video_files)


if __name__ == "__main__":
    main()

def get_subtitle_text(ydl: yt_dlp.YoutubeDL, video_path, req_sub_dict) -> str:
    def dl_sub(url):
        temp_file = tempfile.mktemp(".srt", dir=SUB_TEMP_DIR)

        try:
            ydl.dl(temp_file, {"url": url}, subtitle=True)
        except HTTPError:
            log.info("Unable to download subtitles; skipping")
            sleep(5)
            return None

        return temp_file

    urls = [d["url"] for d in list(req_sub_dict.values())]
    paths = utils.conform([dl_sub(url) for url in urls])

    subs_text = subs_to_text(video_path, paths)
    for p in paths:
        utils.trash(p)

    return subs_text

I thought this code was working but it doesn't make sense that it should work. That said, I'm too lazy to test whether or not it actually works. I think it does. Feel free to try