X-Git-Url: http://git.bitcoin.ninja/index.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fyoutube.py;h=92b9f3ae4eab412ef42f30cd29201dfd6783e3a7;hb=e9bd0f772b28176e86cfe8c641b6281a96be2ee4;hp=21731188af1b413fb05578f02a8154e393396fa9;hpb=3cfd000849208b58dab4f78d1486d3f24552009e;p=youtube-dl diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py index 21731188a..92b9f3ae4 100644 --- a/youtube_dl/extractor/youtube.py +++ b/youtube_dl/extractor/youtube.py @@ -20,21 +20,24 @@ from ..compat import ( compat_urllib_parse_unquote, compat_urllib_parse_unquote_plus, compat_urllib_parse_urlparse, - compat_urllib_request, compat_urlparse, compat_str, ) from ..utils import ( clean_html, encode_dict, + error_to_compat_str, ExtractorError, float_or_none, get_element_by_attribute, get_element_by_id, int_or_none, + mimetype2ext, orderedSet, parse_duration, + remove_quotes, remove_start, + sanitized_Request, smuggle_url, str_to_int, unescapeHTML, @@ -114,7 +117,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): login_data = compat_urllib_parse.urlencode(encode_dict(login_form_strs)).encode('ascii') - req = compat_urllib_request.Request(self._LOGIN_URL, login_data) + req = sanitized_Request(self._LOGIN_URL, login_data) login_results = self._download_webpage( req, None, note='Logging in', errnote='unable to log in', fatal=False) @@ -147,7 +150,7 @@ class YoutubeBaseInfoExtractor(InfoExtractor): tfa_data = compat_urllib_parse.urlencode(encode_dict(tfa_form_strs)).encode('ascii') - tfa_req = compat_urllib_request.Request(self._TWOFACTOR_URL, tfa_data) + tfa_req = sanitized_Request(self._TWOFACTOR_URL, tfa_data) tfa_results = self._download_webpage( tfa_req, None, note='Submitting TFA code', errnote='unable to submit tfa', fatal=False) @@ -178,15 +181,13 @@ class YoutubeBaseInfoExtractor(InfoExtractor): return -class YoutubePlaylistBaseInfoExtractor(InfoExtractor): - # Extract the video ids from the playlist pages +class YoutubeEntryListBaseInfoExtractor(InfoExtractor): + # Extract entries from page with "Load more" button def _entries(self, page, playlist_id): more_widget_html = content_html = page for page_num in itertools.count(1): - for video_id, video_title in self.extract_videos_from_page(content_html): - yield self.url_result( - video_id, 'Youtube', video_id=video_id, - video_title=video_title) + for entry in self._process_page(content_html): + yield entry mobj = re.search(r'data-uix-load-more-href="/?(?P[^"]+)"', more_widget_html) if not mobj: @@ -203,6 +204,12 @@ class YoutubePlaylistBaseInfoExtractor(InfoExtractor): break more_widget_html = more['load_more_widget_html'] + +class YoutubePlaylistBaseInfoExtractor(YoutubeEntryListBaseInfoExtractor): + def _process_page(self, content): + for video_id, video_title in self.extract_videos_from_page(content): + yield self.url_result(video_id, 'Youtube', video_id, video_title) + def extract_videos_from_page(self, page): ids_in_page = [] titles_in_page = [] @@ -224,6 +231,19 @@ class YoutubePlaylistBaseInfoExtractor(InfoExtractor): return zip(ids_in_page, titles_in_page) +class YoutubePlaylistsBaseInfoExtractor(YoutubeEntryListBaseInfoExtractor): + def _process_page(self, content): + for playlist_id in re.findall(r'href="/?playlist\?list=(.+?)"', content): + yield self.url_result( + 'https://www.youtube.com/playlist?list=%s' % playlist_id, 'YoutubePlaylist') + + def _real_extract(self, url): + playlist_id = self._match_id(url) + webpage = self._download_webpage(url, playlist_id) + title = self._og_search_title(webpage, fatal=False) + return self.playlist_result(self._entries(webpage, playlist_id), playlist_id, title) + + class YoutubeIE(YoutubeBaseInfoExtractor): IE_DESC = 'YouTube.com' _VALID_URL = r"""(?x)^ @@ -241,7 +261,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): |(?: # or the v= param in all its forms (?:(?:watch|movie)(?:_popup)?(?:\.php)?/?)? # preceding watch(_popup|.php) or nothing (like /?v=xxxx) (?:\?|\#!?) # the params delimiter ? or # or #! - (?:.*?&)?? # any other preceding param (like /?s=tuff&v=xxxx) + (?:.*?[&;])?? # any other preceding param (like /?s=tuff&v=xxxx or ?s=tuff&v=V36LpHqtcDY) v= ) )) @@ -258,55 +278,55 @@ class YoutubeIE(YoutubeBaseInfoExtractor): $""" _NEXT_URL_RE = r'[\?&]next_url=([^&]+)' _formats = { - '5': {'ext': 'flv', 'width': 400, 'height': 240}, - '6': {'ext': 'flv', 'width': 450, 'height': 270}, - '13': {'ext': '3gp'}, - '17': {'ext': '3gp', 'width': 176, 'height': 144}, - '18': {'ext': 'mp4', 'width': 640, 'height': 360}, - '22': {'ext': 'mp4', 'width': 1280, 'height': 720}, - '34': {'ext': 'flv', 'width': 640, 'height': 360}, - '35': {'ext': 'flv', 'width': 854, 'height': 480}, - '36': {'ext': '3gp', 'width': 320, 'height': 240}, - '37': {'ext': 'mp4', 'width': 1920, 'height': 1080}, - '38': {'ext': 'mp4', 'width': 4096, 'height': 3072}, - '43': {'ext': 'webm', 'width': 640, 'height': 360}, - '44': {'ext': 'webm', 'width': 854, 'height': 480}, - '45': {'ext': 'webm', 'width': 1280, 'height': 720}, - '46': {'ext': 'webm', 'width': 1920, 'height': 1080}, - '59': {'ext': 'mp4', 'width': 854, 'height': 480}, - '78': {'ext': 'mp4', 'width': 854, 'height': 480}, - - - # 3d videos - '82': {'ext': 'mp4', 'height': 360, 'format_note': '3D', 'preference': -20}, - '83': {'ext': 'mp4', 'height': 480, 'format_note': '3D', 'preference': -20}, - '84': {'ext': 'mp4', 'height': 720, 'format_note': '3D', 'preference': -20}, - '85': {'ext': 'mp4', 'height': 1080, 'format_note': '3D', 'preference': -20}, - '100': {'ext': 'webm', 'height': 360, 'format_note': '3D', 'preference': -20}, - '101': {'ext': 'webm', 'height': 480, 'format_note': '3D', 'preference': -20}, - '102': {'ext': 'webm', 'height': 720, 'format_note': '3D', 'preference': -20}, + '5': {'ext': 'flv', 'width': 400, 'height': 240, 'acodec': 'mp3', 'abr': 64, 'vcodec': 'h263'}, + '6': {'ext': 'flv', 'width': 450, 'height': 270, 'acodec': 'mp3', 'abr': 64, 'vcodec': 'h263'}, + '13': {'ext': '3gp', 'acodec': 'aac', 'vcodec': 'mp4v'}, + '17': {'ext': '3gp', 'width': 176, 'height': 144, 'acodec': 'aac', 'abr': 24, 'vcodec': 'mp4v'}, + '18': {'ext': 'mp4', 'width': 640, 'height': 360, 'acodec': 'aac', 'abr': 96, 'vcodec': 'h264'}, + '22': {'ext': 'mp4', 'width': 1280, 'height': 720, 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264'}, + '34': {'ext': 'flv', 'width': 640, 'height': 360, 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264'}, + '35': {'ext': 'flv', 'width': 854, 'height': 480, 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264'}, + '36': {'ext': '3gp', 'width': 320, 'height': 240, 'acodec': 'aac', 'abr': 32, 'vcodec': 'mp4v'}, + '37': {'ext': 'mp4', 'width': 1920, 'height': 1080, 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264'}, + '38': {'ext': 'mp4', 'width': 4096, 'height': 3072, 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264'}, + '43': {'ext': 'webm', 'width': 640, 'height': 360, 'acodec': 'vorbis', 'abr': 128, 'vcodec': 'vp8'}, + '44': {'ext': 'webm', 'width': 854, 'height': 480, 'acodec': 'vorbis', 'abr': 128, 'vcodec': 'vp8'}, + '45': {'ext': 'webm', 'width': 1280, 'height': 720, 'acodec': 'vorbis', 'abr': 192, 'vcodec': 'vp8'}, + '46': {'ext': 'webm', 'width': 1920, 'height': 1080, 'acodec': 'vorbis', 'abr': 192, 'vcodec': 'vp8'}, + '59': {'ext': 'mp4', 'width': 854, 'height': 480, 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264'}, + '78': {'ext': 'mp4', 'width': 854, 'height': 480, 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264'}, + + + # 3D videos + '82': {'ext': 'mp4', 'height': 360, 'format_note': '3D', 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264', 'preference': -20}, + '83': {'ext': 'mp4', 'height': 480, 'format_note': '3D', 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264', 'preference': -20}, + '84': {'ext': 'mp4', 'height': 720, 'format_note': '3D', 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264', 'preference': -20}, + '85': {'ext': 'mp4', 'height': 1080, 'format_note': '3D', 'acodec': 'aac', 'abr': 192, 'vcodec': 'h264', 'preference': -20}, + '100': {'ext': 'webm', 'height': 360, 'format_note': '3D', 'acodec': 'vorbis', 'abr': 128, 'vcodec': 'vp8', 'preference': -20}, + '101': {'ext': 'webm', 'height': 480, 'format_note': '3D', 'acodec': 'vorbis', 'abr': 192, 'vcodec': 'vp8', 'preference': -20}, + '102': {'ext': 'webm', 'height': 720, 'format_note': '3D', 'acodec': 'vorbis', 'abr': 192, 'vcodec': 'vp8', 'preference': -20}, # Apple HTTP Live Streaming - '92': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10}, - '93': {'ext': 'mp4', 'height': 360, 'format_note': 'HLS', 'preference': -10}, - '94': {'ext': 'mp4', 'height': 480, 'format_note': 'HLS', 'preference': -10}, - '95': {'ext': 'mp4', 'height': 720, 'format_note': 'HLS', 'preference': -10}, - '96': {'ext': 'mp4', 'height': 1080, 'format_note': 'HLS', 'preference': -10}, - '132': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'preference': -10}, - '151': {'ext': 'mp4', 'height': 72, 'format_note': 'HLS', 'preference': -10}, + '92': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 48, 'vcodec': 'h264', 'preference': -10}, + '93': {'ext': 'mp4', 'height': 360, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264', 'preference': -10}, + '94': {'ext': 'mp4', 'height': 480, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 128, 'vcodec': 'h264', 'preference': -10}, + '95': {'ext': 'mp4', 'height': 720, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 256, 'vcodec': 'h264', 'preference': -10}, + '96': {'ext': 'mp4', 'height': 1080, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 256, 'vcodec': 'h264', 'preference': -10}, + '132': {'ext': 'mp4', 'height': 240, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 48, 'vcodec': 'h264', 'preference': -10}, + '151': {'ext': 'mp4', 'height': 72, 'format_note': 'HLS', 'acodec': 'aac', 'abr': 24, 'vcodec': 'h264', 'preference': -10}, # DASH mp4 video - '133': {'ext': 'mp4', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '134': {'ext': 'mp4', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '138': {'ext': 'mp4', 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, # Height can vary (https://github.com/rg3/youtube-dl/issues/4559) - '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '298': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'}, - '299': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'h264'}, - '266': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'h264'}, + '133': {'ext': 'mp4', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40}, + '134': {'ext': 'mp4', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40}, + '135': {'ext': 'mp4', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40}, + '136': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40}, + '137': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40}, + '138': {'ext': 'mp4', 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40}, # Height can vary (https://github.com/rg3/youtube-dl/issues/4559) + '160': {'ext': 'mp4', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40}, + '264': {'ext': 'mp4', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40}, + '298': {'ext': 'mp4', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'fps': 60, 'preference': -40}, + '299': {'ext': 'mp4', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'fps': 60, 'preference': -40}, + '266': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'h264', 'preference': -40}, # Dash mp4 audio '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 48, 'preference': -50, 'container': 'm4a_dash'}, @@ -320,25 +340,26 @@ class YoutubeIE(YoutubeBaseInfoExtractor): '170': {'ext': 'webm', 'height': 1080, 'width': 1920, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'vp8', 'preference': -40}, '218': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'vp8', 'preference': -40}, '219': {'ext': 'webm', 'height': 480, 'width': 854, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'vp8', 'preference': -40}, - '278': {'ext': 'webm', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'container': 'webm', 'vcodec': 'vp9'}, - '242': {'ext': 'webm', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '243': {'ext': 'webm', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '244': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '245': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '246': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '271': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40}, - '302': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'}, - '303': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'}, - '308': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'}, - '313': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'vp9'}, - '315': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'fps': 60, 'vcodec': 'vp9'}, + '278': {'ext': 'webm', 'height': 144, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'vp9', 'preference': -40}, + '242': {'ext': 'webm', 'height': 240, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40}, + '243': {'ext': 'webm', 'height': 360, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40}, + '244': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40}, + '245': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40}, + '246': {'ext': 'webm', 'height': 480, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40}, + '247': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40}, + '248': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40}, + '271': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40}, + # itag 272 videos are either 3840x2160 (e.g. RtoitU2A-3E) or 7680x4320 (sLprVF6d7Ug) + '272': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40}, + '302': {'ext': 'webm', 'height': 720, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'fps': 60, 'preference': -40}, + '303': {'ext': 'webm', 'height': 1080, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'fps': 60, 'preference': -40}, + '308': {'ext': 'webm', 'height': 1440, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'fps': 60, 'preference': -40}, + '313': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'preference': -40}, + '315': {'ext': 'webm', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'vcodec': 'vp9', 'fps': 60, 'preference': -40}, # Dash webm audio - '171': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 128, 'preference': -50}, - '172': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 256, 'preference': -50}, + '171': {'ext': 'webm', 'acodec': 'vorbis', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 128, 'preference': -50}, + '172': {'ext': 'webm', 'acodec': 'vorbis', 'vcodec': 'none', 'format_note': 'DASH audio', 'abr': 256, 'preference': -50}, # Dash webm audio with opus inside '249': {'ext': 'webm', 'vcodec': 'none', 'format_note': 'DASH audio', 'acodec': 'opus', 'abr': 50, 'preference': -50}, @@ -377,12 +398,14 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'ext': 'mp4', 'upload_date': '20120506', 'title': 'Icona Pop - I Love It (feat. Charli XCX) [OFFICIAL VIDEO]', + 'alt_title': 'I Love It (feat. Charli XCX)', 'description': 'md5:782e8651347686cba06e58f71ab51773', 'tags': ['Icona Pop i love it', 'sweden', 'pop music', 'big beat records', 'big beat', 'charli', 'xcx', 'charli xcx', 'girls', 'hbo', 'i love it', "i don't care", 'icona', 'pop', 'iconic ep', 'iconic', 'love', 'it'], 'uploader': 'Icona Pop', 'uploader_id': 'IconaPop', + 'creator': 'Icona Pop', } }, { @@ -393,9 +416,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'ext': 'mp4', 'upload_date': '20130703', 'title': 'Justin Timberlake - Tunnel Vision (Explicit)', + 'alt_title': 'Tunnel Vision', 'description': 'md5:64249768eec3bc4276236606ea996373', 'uploader': 'justintimberlakeVEVO', 'uploader_id': 'justintimberlakeVEVO', + 'creator': 'Justin Timberlake', 'age_limit': 18, } }, @@ -409,7 +434,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'title': 'Principal Sexually Assaults A Teacher - Episode 117 - 8th June 2012', 'description': 'md5:09b78bd971f1e3e289601dfba15ca4f7', 'uploader': 'SET India', - 'uploader_id': 'setindia' + 'uploader_id': 'setindia', + 'age_limit': 18, } }, { @@ -473,10 +499,12 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'id': 'nfWlot6h_JM', 'ext': 'm4a', 'title': 'Taylor Swift - Shake It Off', + 'alt_title': 'Shake It Off', 'description': 'md5:95f66187cd7c8b2c13eb78e1223b63c3', 'uploader': 'TaylorSwiftVEVO', 'uploader_id': 'TaylorSwiftVEVO', 'upload_date': '20140818', + 'creator': 'Taylor Swift', }, 'params': { 'youtube_include_dash_manifest': True, @@ -532,9 +560,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'ext': 'mp4', 'upload_date': '20100430', 'uploader_id': 'deadmau5', + 'creator': 'deadmau5', 'description': 'md5:12c56784b8032162bb936a5f76d55360', 'uploader': 'deadmau5', 'title': 'Deadmau5 - Some Chords (HD)', + 'alt_title': 'Some Chords', }, 'expected_warnings': [ 'DASH manifest missing', @@ -546,7 +576,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'info_dict': { 'id': 'lqQg6PlCWgI', 'ext': 'mp4', - 'upload_date': '20120724', + 'upload_date': '20150827', 'uploader_id': 'olympic', 'description': 'HO09 - Women - GER-AUS - Hockey - 31 July 2012 - London 2012 Olympic Games', 'uploader': 'Olympics', @@ -584,7 +614,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): }, 'params': { 'skip_download': 'requires avconv', - } + }, + 'skip': 'This live event has ended.', }, # Extraction from multiple DASH manifests (https://github.com/rg3/youtube-dl/pull/6097) { @@ -674,6 +705,52 @@ class YoutubeIE(YoutubeBaseInfoExtractor): { 'url': 'http://vid.plus/FlRa-iH7PGw', 'only_matching': True, + }, + { + # Title with JS-like syntax "};" (see https://github.com/rg3/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) + 'url': 'https://www.youtube.com/watch?v=lsguqyKfVQg', + 'info_dict': { + 'id': 'lsguqyKfVQg', + 'ext': 'mp4', + 'title': '{dark walk}; Loki/AC/Dishonored; collab w/Elflover21', + 'alt_title': 'Dark Walk', + 'description': 'md5:8085699c11dc3f597ce0410b0dcbb34a', + 'upload_date': '20151119', + 'uploader_id': 'IronSoulElf', + 'uploader': 'IronSoulElf', + 'creator': 'Todd Haberman, Daniel Law Heath & Aaron Kaplan', + }, + 'params': { + 'skip_download': True, + }, + }, + { + # Tags with '};' (see https://github.com/rg3/youtube-dl/issues/7468) + 'url': 'https://www.youtube.com/watch?v=Ms7iBXnlUO8', + 'only_matching': True, + }, + { + # Video with yt:stretch=17:0 + 'url': 'https://www.youtube.com/watch?v=Q39EVAstoRM', + 'info_dict': { + 'id': 'Q39EVAstoRM', + 'ext': 'mp4', + 'title': 'Clash Of Clans#14 Dicas De Ataque Para CV 4', + 'description': 'md5:ee18a25c350637c8faff806845bddee9', + 'upload_date': '20151107', + 'uploader_id': 'UCCr7TALkRbo3EtFzETQF1LA', + 'uploader': 'CH GAMER DROID', + }, + 'params': { + 'skip_download': True, + }, + }, + { + 'url': 'https://www.youtube.com/watch?feature=player_embedded&amp;v=V36LpHqtcDY', + 'only_matching': True, } ] @@ -832,7 +909,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'https://video.google.com/timedtext?hl=en&type=list&v=%s' % video_id, video_id, note=False) except ExtractorError as err: - self._downloader.report_warning('unable to download video subtitles: %s' % compat_str(err)) + self._downloader.report_warning('unable to download video subtitles: %s' % error_to_compat_str(err)) return {} sub_lang_list = {} @@ -858,19 +935,39 @@ class YoutubeIE(YoutubeBaseInfoExtractor): return {} return sub_lang_list + def _get_ytplayer_config(self, video_id, webpage): + patterns = ( + # User data may contain arbitrary character sequences that may affect + # JSON extraction with regex, e.g. when '};' is contained the second + # 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) + r';ytplayer\.config\s*=\s*({.+?});ytplayer', + r';ytplayer\.config\s*=\s*({.+?});', + ) + config = self._search_regex( + patterns, webpage, 'ytplayer.config', default=None) + if config: + return self._parse_json( + uppercase_escape(config), video_id, fatal=False) + def _get_automatic_captions(self, video_id, webpage): """We need the webpage for getting the captions url, pass it as an argument to speed up the process.""" self.to_screen('%s: Looking for automatic captions' % video_id) - mobj = re.search(r';ytplayer.config = ({.*?});', webpage) + player_config = self._get_ytplayer_config(video_id, webpage) err_msg = 'Couldn\'t find automatic captions for %s' % video_id - if mobj is None: + if not player_config: self._downloader.report_warning(err_msg) return {} - player_config = json.loads(mobj.group(1)) try: args = player_config['args'] caption_url = args['ttsurl'] + if not caption_url: + self._downloader.report_warning(err_msg) + return {} timestamp = args['timestamp'] # We get the available subtitles list_params = compat_urllib_parse.urlencode({ @@ -994,9 +1091,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): full_info.update(f) codecs = r.attrib.get('codecs') if codecs: - if full_info.get('acodec') == 'none' and 'vcodec' not in full_info: + if full_info.get('acodec') == 'none': full_info['vcodec'] = codecs - elif full_info.get('vcodec') == 'none' and 'acodec' not in full_info: + elif full_info.get('vcodec') == 'none': full_info['acodec'] = codecs formats.append(full_info) else: @@ -1074,10 +1171,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor): age_gate = False video_info = None # Try looking directly into the video webpage - mobj = re.search(r';ytplayer\.config\s*=\s*({.*?});ytplayer', video_webpage) - if mobj: - json_code = uppercase_escape(mobj.group(1)) - ytplayer_config = json.loads(json_code) + ytplayer_config = self._get_ytplayer_config(video_id, video_webpage) + if ytplayer_config: args = ytplayer_config['args'] if args.get('url_encoded_fmt_stream_map'): # Convert to the same format returned by compat_parse_qs @@ -1148,10 +1243,10 @@ class YoutubeIE(YoutubeBaseInfoExtractor): video_description = re.sub(r'''(?x) - [^<]+ + class="(?:yt-uix-redirect-link|yt-uix-sessionlink[^"]*)"[^>]*> + [^<]+\.{3}\s* ''', r'\1', video_description) video_description = clean_html(video_description) @@ -1233,6 +1328,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor): upload_date = ' '.join(re.sub(r'[/,-]', r' ', mobj.group(1)).split()) upload_date = unified_strdate(upload_date) + m_music = re.search( + r']+class="title"[^>]*>\s*Music\s*\s*]*>\s*
  • (?P.+?) by (?P<creator>.+?)(?:\(.+?\))?</li', + video_webpage) + if m_music: + video_alt_title = remove_quotes(unescapeHTML(m_music.group('title'))) + video_creator = clean_html(m_music.group('creator')) + else: + video_alt_title = video_creator = None + m_cat_container = self._search_regex( r'(?s)<h4[^>]*>\s*Category\s*</h4>\s*<ul[^>]*>(.*?)</ul>', video_webpage, 'categories', default=None) @@ -1358,15 +1462,21 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if 'ratebypass' not in url: url += '&ratebypass=yes' + dct = { + 'format_id': format_id, + 'url': url, + 'player_url': player_url, + } + if format_id in self._formats: + dct.update(self._formats[format_id]) + # Some itags are not included in DASH manifest thus corresponding formats will # lack metadata (see https://github.com/rg3/youtube-dl/pull/5993). # Trying to extract metadata from url_encoded_fmt_stream_map entry. mobj = re.search(r'^(?P<width>\d+)[xX](?P<height>\d+)$', url_data.get('size', [''])[0]) width, height = (int(mobj.group('width')), int(mobj.group('height'))) if mobj else (None, None) - dct = { - 'format_id': format_id, - 'url': url, - 'player_url': player_url, + + more_fields = { 'filesize': int_or_none(url_data.get('clen', [None])[0]), 'tbr': float_or_none(url_data.get('bitrate', [None])[0], 1000), 'width': width, @@ -1374,13 +1484,16 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'fps': int_or_none(url_data.get('fps', [None])[0]), 'format_note': url_data.get('quality_label', [None])[0] or url_data.get('quality', [None])[0], } + for key, value in more_fields.items(): + if value: + dct[key] = value type_ = url_data.get('type', [None])[0] if type_: type_split = type_.split(';') kind_ext = type_split[0].split('/') if len(kind_ext) == 2: - kind, ext = kind_ext - dct['ext'] = ext + kind, _ = kind_ext + dct['ext'] = mimetype2ext(type_split[0]) if kind in ('audio', 'video'): codecs = None for mobj in re.finditer( @@ -1391,21 +1504,27 @@ class YoutubeIE(YoutubeBaseInfoExtractor): if codecs: codecs = codecs.split(',') if len(codecs) == 2: - acodec, vcodec = codecs[0], codecs[1] + acodec, vcodec = codecs[1], codecs[0] else: acodec, vcodec = (codecs[0], 'none') if kind == 'audio' else ('none', codecs[0]) dct.update({ 'acodec': acodec, 'vcodec': vcodec, }) - if format_id in self._formats: - dct.update(self._formats[format_id]) formats.append(dct) elif video_info.get('hlsvp'): manifest_url = video_info['hlsvp'][0] url_map = self._extract_from_m3u8(manifest_url, video_id) formats = _map_to_format_list(url_map) + # Accept-Encoding header causes failures in live streams on Youtube and Youtube Gaming + for a_format in formats: + a_format.setdefault('http_headers', {})['Youtubedl-no-compression'] = 'True' else: + unavailable_message = self._html_search_regex( + r'(?s)<h1[^>]+id="unavailable-message"[^>]*>(.+?)</h1>', + video_webpage, 'unavailable message', default=None) + if unavailable_message: + raise ExtractorError(unavailable_message, expected=True) raise ExtractorError('no conn, hlsvp or url_encoded_fmt_stream_map information found in video info') # Look for the DASH manifest @@ -1442,10 +1561,15 @@ class YoutubeIE(YoutubeBaseInfoExtractor): r'<meta\s+property="og:video:tag".*?content="yt:stretch=(?P<w>[0-9]+):(?P<h>[0-9]+)">', video_webpage) if stretched_m: - ratio = float(stretched_m.group('w')) / float(stretched_m.group('h')) - for f in formats: - if f.get('vcodec') != 'none': - f['stretched_ratio'] = ratio + w = float(stretched_m.group('w')) + h = float(stretched_m.group('h')) + # yt:stretch may hold invalid ratio data (e.g. for Q39EVAstoRM ratio is 17:0). + # We will only process correct ratios. + if w > 0 and h > 0: + ratio = w / h + for f in formats: + if f.get('vcodec') != 'none': + f['stretched_ratio'] = ratio self._sort_formats(formats) @@ -1454,7 +1578,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor): 'uploader': video_uploader, 'uploader_id': video_uploader_id, 'upload_date': upload_date, + 'creator': video_creator, 'title': video_title, + 'alt_title': video_alt_title, 'thumbnail': video_thumbnail, 'description': video_description, 'categories': video_categories, @@ -1484,7 +1610,7 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor, YoutubePlaylistBaseInfoExtract youtube\.com/ (?: (?:course|view_play_list|my_playlists|artist|playlist|watch|embed/videoseries) - \? (?:.*?&)*? (?:p|a|list)= + \? (?:.*?[&;])*? (?:p|a|list)= | p/ ) ( @@ -1615,7 +1741,7 @@ class YoutubePlaylistIE(YoutubeBaseInfoExtractor, YoutubePlaylistBaseInfoExtract self.report_warning('Youtube gives an alert message: ' + match) playlist_title = self._html_search_regex( - r'(?s)<h1 class="pl-header-title[^"]*">\s*(.*?)\s*</h1>', + r'(?s)<h1 class="pl-header-title[^"]*"[^>]*>\s*(.*?)\s*</h1>', page, 'title') return self.playlist_result(self._entries(page, playlist_id), playlist_id, playlist_title) @@ -1669,6 +1795,10 @@ class YoutubeChannelIE(YoutubePlaylistBaseInfoExtractor): }, }] + @classmethod + def suitable(cls, url): + return False if YoutubePlaylistsIE.suitable(url) else super(YoutubeChannelIE, cls).suitable(url) + def _real_extract(self, url): channel_id = self._match_id(url) @@ -1742,6 +1872,36 @@ class YoutubeUserIE(YoutubeChannelIE): return super(YoutubeUserIE, cls).suitable(url) +class YoutubePlaylistsIE(YoutubePlaylistsBaseInfoExtractor): + IE_DESC = 'YouTube.com user/channel playlists' + _VALID_URL = r'https?://(?:\w+\.)?youtube\.com/(?:user|channel)/(?P<id>[^/]+)/playlists' + IE_NAME = 'youtube:playlists' + + _TESTS = [{ + 'url': 'http://www.youtube.com/user/ThirstForScience/playlists', + 'playlist_mincount': 4, + 'info_dict': { + 'id': 'ThirstForScience', + 'title': 'Thirst for Science', + }, + }, { + # with "Load more" button + 'url': 'http://www.youtube.com/user/igorkle1/playlists?view=1&sort=dd', + 'playlist_mincount': 70, + 'info_dict': { + 'id': 'igorkle1', + 'title': 'Игорь Клейнер', + }, + }, { + 'url': 'https://www.youtube.com/channel/UCiU1dHvZObB2iP6xkJ__Icw/playlists', + 'playlist_mincount': 17, + 'info_dict': { + 'id': 'UCiU1dHvZObB2iP6xkJ__Icw', + 'title': 'Chem Player', + }, + }] + + class YoutubeSearchIE(SearchInfoExtractor, YoutubePlaylistIE): IE_DESC = 'YouTube.com searches' # there doesn't appear to be a real limit, for example if you search for @@ -1837,7 +1997,7 @@ class YoutubeSearchURLIE(InfoExtractor): } -class YoutubeShowIE(InfoExtractor): +class YoutubeShowIE(YoutubePlaylistsBaseInfoExtractor): IE_DESC = 'YouTube.com (multi-season) shows' _VALID_URL = r'https?://www\.youtube\.com/show/(?P<id>[^?#]*)' IE_NAME = 'youtube:show' @@ -1851,26 +2011,9 @@ class YoutubeShowIE(InfoExtractor): }] def _real_extract(self, url): - mobj = re.match(self._VALID_URL, url) - playlist_id = mobj.group('id') - webpage = self._download_webpage( - 'https://www.youtube.com/show/%s/playlists' % playlist_id, playlist_id, 'Downloading show webpage') - # There's one playlist for each season of the show - m_seasons = list(re.finditer(r'href="(/playlist\?list=.*?)"', webpage)) - self.to_screen('%s: Found %s seasons' % (playlist_id, len(m_seasons))) - entries = [ - self.url_result( - 'https://www.youtube.com' + season.group(1), 'YoutubePlaylist') - for season in m_seasons - ] - title = self._og_search_title(webpage, fatal=False) - - return { - '_type': 'playlist', - 'id': playlist_id, - 'title': title, - 'entries': entries, - } + playlist_id = self._match_id(url) + return super(YoutubeShowIE, self)._real_extract( + 'https://www.youtube.com/show/%s/playlists' % playlist_id) class YoutubeFeedsInfoExtractor(YoutubeBaseInfoExtractor):