ytsm/app/YtManagerApp/management/jobs/download_video.py

from YtManagerApp.models import Video
from YtManagerApp import scheduler
from YtManagerApp.appconfig import settings
import os
import youtube_dl
import logging
import re
from threading import Lock

log = logging.getLogger('video_downloader')
log_youtube_dl = log.getChild('youtube_dl')

_lock = Lock()


def __get_valid_path(path):
    """
    Normalizes string, converts to lowercase, removes non-alpha characters,
    and converts spaces to hyphens.
    """
    import unicodedata
    value = unicodedata.normalize('NFKD', path).encode('ascii', 'ignore').decode('ascii')
    value = re.sub('[:"*]', '', value).strip()
    value = re.sub('[?<>|]', '#', value)
    return value


def __build_youtube_dl_params(video: Video):
    # resolve path
    pattern_dict = {
        'channel': video.subscription.channel_name,
        'channel_id': video.subscription.channel_id,
        'playlist': video.subscription.name,
        'playlist_id': video.subscription.playlist_id,
        'playlist_index': "{:03d}".format(1 + video.playlist_index),
        'title': video.name,
        'id': video.video_id,
    }

    download_path = settings.get_sub(video.subscription, 'user', 'DownloadPath')
    output_pattern = __get_valid_path(settings.get_sub(
        video.subscription, 'user', 'DownloadFilePattern', vars=pattern_dict))

    output_path = os.path.join(download_path, output_pattern)
    output_path = os.path.normpath(output_path)

    youtube_dl_params = {
        'logger': log_youtube_dl,
        'format': settings.get_sub(video.subscription, 'user', 'DownloadFormat'),
        'outtmpl': output_path,
        'writethumbnail': True,
        'writedescription': True,
        'writesubtitles': settings.getboolean_sub(video.subscription, 'user', 'DownloadSubtitles'),
        'writeautomaticsub': settings.getboolean_sub(video.subscription, 'user', 'DownloadAutogeneratedSubtitles'),
        'allsubtitles': settings.getboolean_sub(video.subscription, 'user', 'DownloadSubtitlesAll'),
        'postprocessors': [
            {
                'key': 'FFmpegMetadata'
            },
        ]
    }

    sub_langs = settings.get_sub(video.subscription, 'user', 'DownloadSubtitlesLangs').split(',')
    sub_langs = [i.strip() for i in sub_langs]
    if len(sub_langs) > 0:
        youtube_dl_params['subtitleslangs'] = sub_langs

    sub_format = settings.get_sub(video.subscription, 'user', 'DownloadSubtitlesFormat')
    if len(sub_format) > 0:
        youtube_dl_params['subtitlesformat'] = sub_format

    return youtube_dl_params, output_path


def download_video(video: Video, attempt: int = 1):

    log.info('Downloading video %d [%s %s]', video.id, video.video_id, video.name)

    # Issue: if multiple videos are downloaded at the same time, a race condition appears in the mkdirs() call that
    # youtube-dl makes, which causes it to fail with the error 'Cannot create folder - file already exists'.
    # For now, allow a single download instance.
    _lock.acquire()

    try:
        max_attempts = settings.getint_sub(video.subscription, 'user', 'DownloadMaxAttempts', fallback=3)

        youtube_dl_params, output_path = __build_youtube_dl_params(video)
        with youtube_dl.YoutubeDL(youtube_dl_params) as yt:
            ret = yt.download(["https://www.youtube.com/watch?v=" + video.video_id])

        log.info('Download finished with code %d', ret)

        if ret == 0:
            video.downloaded_path = output_path
            video.save()
            log.info('Video %d [%s %s] downloaded successfully!', video.id, video.video_id, video.name)

        elif attempt <= max_attempts:
            log.warning('Re-enqueueing video (attempt %d/%d)', attempt, max_attempts)
            __schedule_download_video(video, attempt + 1)

        else:
            log.error('Multiple attempts to download video %d [%s %s] failed!', video.id, video.video_id, video.name)
            video.downloaded_path = ''
            video.save()

    finally:
        _lock.release()


def __schedule_download_video(video: Video, attempt=1):
    job = scheduler.scheduler.add_job(download_video, args=[video, attempt])
    log.info('Scheduled download video job video=(%s), attempt=%d, job=%s', video, attempt, job.id)


def schedule_download_video(video: Video):
    """
    Schedules a download video job to run immediately.
    :param video:
    :return:
    """
    __schedule_download_video(video)