From 01c92973ddebe6429f5a03855e41c412889c96dc Mon Sep 17 00:00:00 2001 From: Remita Amine Date: Mon, 23 Nov 2020 22:44:34 +0100 Subject: [PATCH] [youtube] imporve music metadata and license extraction(closes #26013) --- youtube_dl/extractor/youtube.py | 30 +++++++++++++++++++++++++++++- 1 file changed, 29 insertions(+), 1 deletion(-) diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 7324d8080..1e3ff7d44 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -2162,7 +2162,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # Youtube Music Auto-generated description release_date = release_year = None if video_description: - mobj = re.search(r'(?s)Provided to YouTube by [^\n]+\n+(?P[^·]+)·(?P[^\n]+)\n+(?P[^\n]+)(?:.+?℗\s*(?P\d{4})(?!\d))?(?:.+?Released on\s*:\s*(?P\d{4}-\d{2}-\d{2}))?(.+?\nArtist\s*:\s*(?P[^\n]+))?', video_description) + mobj = re.search(r'(?s)(?P[^·\n]+)·(?P[^\n]+)\n+(?P[^\n]+)(?:.+?℗\s*(?P\d{4})(?!\d))?(?:.+?Released on\s*:\s*(?P\d{4}-\d{2}-\d{2}))?(.+?\nArtist\s*:\s*(?P[^\n]+))?.+\nAuto-generated by YouTube\.\s*$', video_description) if mobj: if not track: track = mobj.group('track').strip() @@ -2179,6 +2179,34 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if release_year: release_year = int(release_year) + yt_initial_data = self._extract_yt_initial_data(video_id, video_webpage) + contents = try_get(yt_initial_data, lambda x: x['contents']['twoColumnWatchNextResults']['results']['results']['contents'], list) or [] + for content in contents: + rows = try_get(content, lambda x: x['videoSecondaryInfoRenderer']['metadataRowContainer']['metadataRowContainerRenderer']['rows'], list) or [] + multiple_songs = False + for row in rows: + if try_get(row, lambda x: x['metadataRowRenderer']['hasDividerLine']) is True: + multiple_songs = True + break + for row in rows: + mrr = row.get('metadataRowRenderer') or {} + mrr_title = try_get( + mrr, lambda x: x['title']['simpleText'], compat_str) + mrr_contents = try_get( + mrr, lambda x: x['contents'][0], dict) or {} + mrr_contents_text = try_get(mrr_contents, [lambda x: x['simpleText'], lambda x: x['runs'][0]['text']], compat_str) + if not (mrr_title and mrr_contents_text): + continue + if mrr_title == 'License': + video_license = mrr_contents_text + elif not multiple_songs: + if mrr_title == 'Album': + album = mrr_contents_text + elif mrr_title == 'Artist': + artist = mrr_contents_text + elif mrr_title == 'Song': + track = mrr_contents_text + m_episode = re.search( r']+id="watch7-headline"[^>]*>\s*]*>.*?>(?P[^<]+)\s*S(?P\d+)\s*•\s*E(?P\d+)', video_webpage)