X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fyoutube.py;h=55eafb8663d93e07be963985c376edae37de9446;hb=822b9d9cb09429645582791dba31f4cbed7583cf;hp=c8bf98b5864fcce75d830f49e719797af6af7205;hpb=fa4ac365f69cbd51e4c9801984ebea49a12825b7;p=youtube-dl diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index c8bf98b58..55eafb866 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -351,7 +351,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor): (?:www\.)?hooktube\.com/| (?:www\.)?yourepeat\.com/| tube\.majestyc\.net/| - (?:www\.)?invidio\.us/| + (?:(?:www|dev)\.)?invidio\.us/| + (?:www\.)?invidiou\.sh/| + (?:www\.)?invidious\.snopyta\.org/| + (?:www\.)?invidious\.kabi\.tk/| + (?:www\.)?vid\.wxzm\.sx/| youtube\.googleapis\.com/) # the various hostnames, with wildcard subdomains (?:.*?\#/)? # handle anchor (#/) redirect urls (?: # the various things that can precede the ID: @@ -427,7 +431,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'vcodec': 'h264'}, '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'vcodec': 'h264'}, '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'vcodec': 'h264'}, - '138': {'ext': 'mp4', 'format_note': 'DASH video', 'vcodec': 'h264'}, # Height can vary (https://github.com/rg3/youtube-dl/issues/4559) + '138': {'ext': 'mp4', 'format_note': 'DASH video', 'vcodec': 'h264'}, # Height can vary (https://github.com/ytdl-org/youtube-dl/issues/4559) '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'vcodec': 'h264'}, '212': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'vcodec': 'h264'}, '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'vcodec': 'h264'}, @@ -480,7 +484,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # RTMP (unnamed) '_rtmp': {'protocol': 'rtmp'}, } - _SUBTITLE_FORMATS = ('ttml', 'vtt') + _SUBTITLE_FORMATS = ('srv1', 'srv2', 'srv3', 'ttml', 'vtt') _GEO_BYPASS = False @@ -692,7 +696,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'age_limit': 18, }, }, - # video_info is None (https://github.com/rg3/youtube-dl/issues/4421) + # video_info is None (https://github.com/ytdl-org/youtube-dl/issues/4421) # YouTube Red ad is not captured for creator { 'url': '__2ABJjxzNo', @@ -713,7 +717,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'DASH manifest missing', ] }, - # Olympics (https://github.com/rg3/youtube-dl/issues/4431) + # Olympics (https://github.com/ytdl-org/youtube-dl/issues/4431) { 'url': 'lqQg6PlCWgI', 'info_dict': { @@ -764,7 +768,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): }, 'skip': 'This live event has ended.', }, - # Extraction from multiple DASH manifests (https://github.com/rg3/youtube-dl/pull/6097) + # Extraction from multiple DASH manifests (https://github.com/ytdl-org/youtube-dl/pull/6097) { 'url': 'https://www.youtube.com/watch?v=FIl7x6_3R5Y', 'info_dict': { @@ -867,7 +871,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'skip': 'This video is not available.', }, { - # Multifeed video with comma in title (see https://github.com/rg3/youtube-dl/issues/8536) + # Multifeed video with comma in title (see https://github.com/ytdl-org/youtube-dl/issues/8536) 'url': 'https://www.youtube.com/watch?v=gVfLd0zydlo', 'info_dict': { 'id': 'gVfLd0zydlo', @@ -885,10 +889,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'only_matching': True, }, { - # Title with JS-like syntax "};" (see https://github.com/rg3/youtube-dl/issues/7468) + # Title with JS-like syntax "};" (see https://github.com/ytdl-org/youtube-dl/issues/7468) # Also tests cut-off URL expansion in video description (see - # https://github.com/rg3/youtube-dl/issues/1892, - # https://github.com/rg3/youtube-dl/issues/8164) + # https://github.com/ytdl-org/youtube-dl/issues/1892, + # https://github.com/ytdl-org/youtube-dl/issues/8164) 'url': 'https://www.youtube.com/watch?v=lsguqyKfVQg', 'info_dict': { 'id': 'lsguqyKfVQg', @@ -910,7 +914,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): }, }, { - # Tags with '};' (see https://github.com/rg3/youtube-dl/issues/7468) + # Tags with '};' (see https://github.com/ytdl-org/youtube-dl/issues/7468) 'url': 'https://www.youtube.com/watch?v=Ms7iBXnlUO8', 'only_matching': True, }, @@ -974,7 +978,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'only_matching': True, }, { - # YouTube Red paid video (https://github.com/rg3/youtube-dl/issues/10059) + # YouTube Red paid video (https://github.com/ytdl-org/youtube-dl/issues/10059) 'url': 'https://www.youtube.com/watch?v=i1Ko8UG-Tdo', 'only_matching': True, }, @@ -1082,7 +1086,95 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'skip_download': True, 'youtube_include_dash_manifest': False, }, - } + }, + { + # Youtube Music Auto-generated description + 'url': 'https://music.youtube.com/watch?v=MgNrAu2pzNs', + 'info_dict': { + 'id': 'MgNrAu2pzNs', + 'ext': 'mp4', + 'title': 'Voyeur Girl', + 'description': 'md5:7ae382a65843d6df2685993e90a8628f', + 'upload_date': '20190312', + 'uploader': 'Various Artists - Topic', + 'uploader_id': 'UCVWKBi1ELZn0QX2CBLSkiyw', + 'artist': 'Stephen', + 'track': 'Voyeur Girl', + 'album': 'it\'s too much love to know my dear', + 'release_date': '20190313', + 'release_year': 2019, + }, + 'params': { + 'skip_download': True, + }, + }, + { + # Youtube Music Auto-generated description + # Retrieve 'artist' field from 'Artist:' in video description + # when it is present on youtube music video + 'url': 'https://www.youtube.com/watch?v=k0jLE7tTwjY', + 'info_dict': { + 'id': 'k0jLE7tTwjY', + 'ext': 'mp4', + 'title': 'Latch Feat. Sam Smith', + 'description': 'md5:3cb1e8101a7c85fcba9b4fb41b951335', + 'upload_date': '20150110', + 'uploader': 'Various Artists - Topic', + 'uploader_id': 'UCNkEcmYdjrH4RqtNgh7BZ9w', + 'artist': 'Disclosure', + 'track': 'Latch Feat. Sam Smith', + 'album': 'Latch Featuring Sam Smith', + 'release_date': '20121008', + 'release_year': 2012, + }, + 'params': { + 'skip_download': True, + }, + }, + { + # Youtube Music Auto-generated description + # handle multiple artists on youtube music video + 'url': 'https://www.youtube.com/watch?v=74qn0eJSjpA', + 'info_dict': { + 'id': '74qn0eJSjpA', + 'ext': 'mp4', + 'title': 'Eastside', + 'description': 'md5:290516bb73dcbfab0dcc4efe6c3de5f2', + 'upload_date': '20180710', + 'uploader': 'Benny Blanco - Topic', + 'uploader_id': 'UCzqz_ksRu_WkIzmivMdIS7A', + 'artist': 'benny blanco, Halsey, Khalid', + 'track': 'Eastside', + 'album': 'Eastside', + 'release_date': '20180713', + 'release_year': 2018, + }, + 'params': { + 'skip_download': True, + }, + }, + { + # Youtube Music Auto-generated description + # handle youtube music video with release_year and no release_date + 'url': 'https://www.youtube.com/watch?v=-hcAI0g-f5M', + 'info_dict': { + 'id': '-hcAI0g-f5M', + 'ext': 'mp4', + 'title': 'Put It On Me', + 'description': 'md5:93c55acc682ae7b0c668f2e34e1c069e', + 'upload_date': '20180426', + 'uploader': 'Matt Maeson - Topic', + 'uploader_id': 'UCnEkIGqtGcQMLk73Kp-Q5LQ', + 'artist': 'Matt Maeson', + 'track': 'Put It On Me', + 'album': 'The Hearse', + 'release_date': None, + 'release_year': 2018, + }, + 'params': { + 'skip_download': True, + }, + }, ] def __init__(self, *args, **kwargs): @@ -1280,8 +1372,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # regex won't capture the whole JSON. Yet working around by trying more # concrete regex first keeping in mind proper quoted string handling # to be implemented in future that will replace this workaround (see - # https://github.com/rg3/youtube-dl/issues/7468, - # https://github.com/rg3/youtube-dl/pull/7599) + # https://github.com/ytdl-org/youtube-dl/issues/7468, + # https://github.com/ytdl-org/youtube-dl/pull/7599) r';ytplayer\.config\s*=\s*({.+?});ytplayer', r';ytplayer\.config\s*=\s*({.+?});', ) @@ -1596,7 +1688,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): add_dash_mpd(video_info) # Rental video is not rented but preview is available (e.g. # https://www.youtube.com/watch?v=yYr8q0y5Jfg, - # https://github.com/rg3/youtube-dl/issues/10532) + # https://github.com/ytdl-org/youtube-dl/issues/10532) if not video_info and args.get('ypc_vid'): return self.url_result( args['ypc_vid'], YoutubeIE.ie_key(), video_id=args['ypc_vid']) @@ -1616,7 +1708,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): # are missing from DASH manifest pointed by webpage's dashmpd, some - from DASH # manifest pointed by get_video_info's dashmpd). # The general idea is to take a union of itags of both DASH manifests (for example - # video with such 'manifest behavior' see https://github.com/rg3/youtube-dl/issues/6093) + # video with such 'manifest behavior' see https://github.com/ytdl-org/youtube-dl/issues/6093) self.report_video_info_webpage_download(video_id) for el in ('info', 'embedded', 'detailpage', 'vevo', ''): query = { @@ -1648,17 +1740,19 @@ class YoutubeIE(YoutubeBaseInfoExtractor): view_count = extract_view_count(get_video_info) if not video_info: video_info = get_video_info - if 'token' in get_video_info: + get_token = get_video_info.get('token') or get_video_info.get('account_playback_token') + if get_token: # Different get_video_info requests may report different results, e.g. # some may report video unavailability, but some may serve it without - # any complaint (see https://github.com/rg3/youtube-dl/issues/7362, + # any complaint (see https://github.com/ytdl-org/youtube-dl/issues/7362, # the original webpage as well as el=info and el=embedded get_video_info # requests report video unavailability due to geo restriction while # el=detailpage succeeds and returns valid data). This is probably # due to YouTube measures against IP ranges of hosting providers. # Working around by preferring the first succeeded video_info containing # the token if no such video_info yet was found. - if 'token' not in video_info: + token = video_info.get('token') or video_info.get('account_playback_token') + if not token: video_info = get_video_info break @@ -1667,7 +1761,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor): r'(?s)