github/yt-dlp
1
0
Fork 0
mirror of https://github.com/yt-dlp/yt-dlp.git synced 2024-11-21 20:46:36 -05:00
Code Issues Projects Releases Packages Wiki Activity

[youtube] Add fallback for duration extraction (closes #11841)

Browse source
This commit is contained in:
Sergey M․ 2017-01-26 21:43:14 +07:00
parent 2417d41535
commit 556dbe7fe3
No known key found for this signature in database
GPG key ID: 2C393E0F18A9236D
1 changed files with 29 additions and 7 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

36
youtube_dl/extractor/youtube.py
View file

@ -40,6 +40,7 @@
sanitized_Request, sanitized_Request,
smuggle_url, smuggle_url,
str_to_int, str_to_int,
try_get,
unescapeHTML, unescapeHTML,
unified_strdate, unified_strdate,
unsmuggle_url, unsmuggle_url,
@ -383,6 +384,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'description': 'test chars: "\'/\\ä↭𝕐\ntest URL: https://github.com/rg3/youtube-dl/issues/1892\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de .', 'description': 'test chars: "\'/\\ä↭𝕐\ntest URL: https://github.com/rg3/youtube-dl/issues/1892\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de .',
'categories': ['Science & Technology'], 'categories': ['Science & Technology'],
'tags': ['youtube-dl'], 'tags': ['youtube-dl'],
'duration': 10,
'like_count': int, 'like_count': int,
'dislike_count': int, 'dislike_count': int,
'start_time': 1, 'start_time': 1,
@ -402,6 +404,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'tags': ['Icona Pop i love it', 'sweden', 'pop music', 'big beat records', 'big beat', 'charli', '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', 'xcx', 'charli xcx', 'girls', 'hbo', 'i love it', "i don't care", 'icona', 'pop',
'iconic ep', 'iconic', 'love', 'it'], 'iconic ep', 'iconic', 'love', 'it'],
'duration': 180,
'uploader': 'Icona Pop', 'uploader': 'Icona Pop',
'uploader_id': 'IconaPop', 'uploader_id': 'IconaPop',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/IconaPop', 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/IconaPop',
@ -419,6 +422,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'title': 'Justin Timberlake - Tunnel Vision (Explicit)', 'title': 'Justin Timberlake - Tunnel Vision (Explicit)',
'alt_title': 'Tunnel Vision', 'alt_title': 'Tunnel Vision',
'description': 'md5:64249768eec3bc4276236606ea996373', 'description': 'md5:64249768eec3bc4276236606ea996373',
'duration': 419,
'uploader': 'justintimberlakeVEVO', 'uploader': 'justintimberlakeVEVO',
'uploader_id': 'justintimberlakeVEVO', 'uploader_id': 'justintimberlakeVEVO',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/justintimberlakeVEVO', 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/justintimberlakeVEVO',
@ -458,6 +462,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'description': 'test chars: "\'/\\ä↭𝕐\ntest URL: https://github.com/rg3/youtube-dl/issues/1892\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de .', 'description': 'test chars: "\'/\\ä↭𝕐\ntest URL: https://github.com/rg3/youtube-dl/issues/1892\n\nThis is a test video for youtube-dl.\n\nFor more information, contact phihag@phihag.de .',
'categories': ['Science & Technology'], 'categories': ['Science & Technology'],
'tags': ['youtube-dl'], 'tags': ['youtube-dl'],
'duration': 10,
'like_count': int, 'like_count': int,
'dislike_count': int, 'dislike_count': int,
}, },
@ -493,6 +498,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'ext': 'm4a', 'ext': 'm4a',
'title': 'Afrojack, Spree Wilson - The Spark ft. Spree Wilson', 'title': 'Afrojack, Spree Wilson - The Spark ft. Spree Wilson',
'description': 'md5:12e7067fa6735a77bdcbb58cb1187d2d', 'description': 'md5:12e7067fa6735a77bdcbb58cb1187d2d',
'duration': 244,
'uploader': 'AfrojackVEVO', 'uploader': 'AfrojackVEVO',
'uploader_id': 'AfrojackVEVO', 'uploader_id': 'AfrojackVEVO',
'upload_date': '20131011', 'upload_date': '20131011',
@ -512,6 +518,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'title': 'Taylor Swift - Shake It Off', 'title': 'Taylor Swift - Shake It Off',
'alt_title': 'Shake It Off', 'alt_title': 'Shake It Off',
'description': 'md5:95f66187cd7c8b2c13eb78e1223b63c3', 'description': 'md5:95f66187cd7c8b2c13eb78e1223b63c3',
'duration': 242,
'uploader': 'TaylorSwiftVEVO', 'uploader': 'TaylorSwiftVEVO',
'uploader_id': 'TaylorSwiftVEVO', 'uploader_id': 'TaylorSwiftVEVO',
'upload_date': '20140818', 'upload_date': '20140818',
@ -529,6 +536,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'info_dict': { 'info_dict': {
'id': 'T4XJQO3qol8', 'id': 'T4XJQO3qol8',
'ext': 'mp4', 'ext': 'mp4',
'duration': 219,
'upload_date': '20100909', 'upload_date': '20100909',
'uploader': 'The Amazing Atheist', 'uploader': 'The Amazing Atheist',
'uploader_id': 'TheAmazingAtheist', 'uploader_id': 'TheAmazingAtheist',
@ -546,6 +554,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'The Witcher 3: Wild Hunt - The Sword Of Destiny Trailer', 'title': 'The Witcher 3: Wild Hunt - The Sword Of Destiny Trailer',
'description': r're:(?s).{100,}About the Game\n.*?The Witcher 3: Wild Hunt.{100,}', 'description': r're:(?s).{100,}About the Game\n.*?The Witcher 3: Wild Hunt.{100,}',
'duration': 142,
'uploader': 'The Witcher', 'uploader': 'The Witcher',
'uploader_id': 'WitcherGame', 'uploader_id': 'WitcherGame',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/WitcherGame', 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/WitcherGame',
@ -562,6 +571,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'Dedication To My Ex (Miss That) (Lyric Video)', 'title': 'Dedication To My Ex (Miss That) (Lyric Video)',
'description': 'md5:33765bb339e1b47e7e72b5490139bb41', 'description': 'md5:33765bb339e1b47e7e72b5490139bb41',
'duration': 247,
'uploader': 'LloydVEVO', 'uploader': 'LloydVEVO',
'uploader_id': 'LloydVEVO', 'uploader_id': 'LloydVEVO',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/LloydVEVO', 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/LloydVEVO',
@ -576,6 +586,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'info_dict': { 'info_dict': {
'id': '__2ABJjxzNo', 'id': '__2ABJjxzNo',
'ext': 'mp4', 'ext': 'mp4',
'duration': 266,
'upload_date': '20100430', 'upload_date': '20100430',
'uploader_id': 'deadmau5', 'uploader_id': 'deadmau5',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/deadmau5', 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/deadmau5',
@ -596,6 +607,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'info_dict': { 'info_dict': {
'id': 'lqQg6PlCWgI', 'id': 'lqQg6PlCWgI',
'ext': 'mp4', 'ext': 'mp4',
'duration': 6085,
'upload_date': '20150827', 'upload_date': '20150827',
'uploader_id': 'olympic', 'uploader_id': 'olympic',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/olympic', 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/olympic',
@ -615,6 +627,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'id': '_b-2C3KPAM0', 'id': '_b-2C3KPAM0',
'ext': 'mp4', 'ext': 'mp4',
'stretched_ratio': 16 / 9., 'stretched_ratio': 16 / 9.,
'duration': 85,
'upload_date': '20110310', 'upload_date': '20110310',
'uploader_id': 'AllenMeow', 'uploader_id': 'AllenMeow',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/AllenMeow', 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/AllenMeow',
@ -649,6 +662,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'md5:7b81415841e02ecd4313668cde88737a', 'title': 'md5:7b81415841e02ecd4313668cde88737a',
'description': 'md5:116377fd2963b81ec4ce64b542173306', 'description': 'md5:116377fd2963b81ec4ce64b542173306',
'duration': 220,
'upload_date': '20150625', 'upload_date': '20150625',
'uploader_id': 'dorappi2000', 'uploader_id': 'dorappi2000',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/dorappi2000', 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/dorappi2000',
@ -691,6 +705,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'teamPGP: Rocket League Noob Stream (Main Camera)', 'title': 'teamPGP: Rocket League Noob Stream (Main Camera)',
'description': 'md5:dc7872fb300e143831327f1bae3af010', 'description': 'md5:dc7872fb300e143831327f1bae3af010',
'duration': 7335,
'upload_date': '20150721', 'upload_date': '20150721',
'uploader': 'Beer Games Beer', 'uploader': 'Beer Games Beer',
'uploader_id': 'beergamesbeer', 'uploader_id': 'beergamesbeer',
@ -703,6 +718,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'teamPGP: Rocket League Noob Stream (kreestuh)', 'title': 'teamPGP: Rocket League Noob Stream (kreestuh)',
'description': 'md5:dc7872fb300e143831327f1bae3af010', 'description': 'md5:dc7872fb300e143831327f1bae3af010',
'duration': 7337,
'upload_date': '20150721', 'upload_date': '20150721',
'uploader': 'Beer Games Beer', 'uploader': 'Beer Games Beer',
'uploader_id': 'beergamesbeer', 'uploader_id': 'beergamesbeer',
@ -715,6 +731,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'teamPGP: Rocket League Noob Stream (grizzle)', 'title': 'teamPGP: Rocket League Noob Stream (grizzle)',
'description': 'md5:dc7872fb300e143831327f1bae3af010', 'description': 'md5:dc7872fb300e143831327f1bae3af010',
'duration': 7337,
'upload_date': '20150721', 'upload_date': '20150721',
'uploader': 'Beer Games Beer', 'uploader': 'Beer Games Beer',
'uploader_id': 'beergamesbeer', 'uploader_id': 'beergamesbeer',
@ -727,6 +744,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'teamPGP: Rocket League Noob Stream (zim)', 'title': 'teamPGP: Rocket League Noob Stream (zim)',
'description': 'md5:dc7872fb300e143831327f1bae3af010', 'description': 'md5:dc7872fb300e143831327f1bae3af010',
'duration': 7334,
'upload_date': '20150721', 'upload_date': '20150721',
'uploader': 'Beer Games Beer', 'uploader': 'Beer Games Beer',
'uploader_id': 'beergamesbeer', 'uploader_id': 'beergamesbeer',
@ -768,6 +786,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'title': '{dark walk}; Loki/AC/Dishonored; collab w/Elflover21', 'title': '{dark walk}; Loki/AC/Dishonored; collab w/Elflover21',
'alt_title': 'Dark Walk', 'alt_title': 'Dark Walk',
'description': 'md5:8085699c11dc3f597ce0410b0dcbb34a', 'description': 'md5:8085699c11dc3f597ce0410b0dcbb34a',
'duration': 133,
'upload_date': '20151119', 'upload_date': '20151119',
'uploader_id': 'IronSoulElf', 'uploader_id': 'IronSoulElf',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/IronSoulElf', 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/IronSoulElf',
@ -809,10 +828,11 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'md5:e41008789470fc2533a3252216f1c1d1', 'title': 'md5:e41008789470fc2533a3252216f1c1d1',
'description': 'md5:a677553cf0840649b731a3024aeff4cc', 'description': 'md5:a677553cf0840649b731a3024aeff4cc',
'duration': 721,
'upload_date': '20150127', 'upload_date': '20150127',
'uploader_id': 'BerkmanCenter', 'uploader_id': 'BerkmanCenter',
'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/BerkmanCenter', 'uploader_url': r're:https?://(?:www\.)?youtube\.com/user/BerkmanCenter',
'uploader': 'BerkmanCenter', 'uploader': 'The Berkman Klein Center for Internet & Society',
'license': 'Creative Commons Attribution license (reuse allowed)', 'license': 'Creative Commons Attribution license (reuse allowed)',
}, },
'params': { 'params': {
@ -827,6 +847,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'ext': 'mp4', 'ext': 'mp4',
'title': 'Democratic Socialism and Foreign Policy | Bernie Sanders', 'title': 'Democratic Socialism and Foreign Policy | Bernie Sanders',
'description': 'md5:dda0d780d5a6e120758d1711d062a867', 'description': 'md5:dda0d780d5a6e120758d1711d062a867',
'duration': 4060,
'upload_date': '20151119', 'upload_date': '20151119',
'uploader': 'Bernie 2016', 'uploader': 'Bernie 2016',
'uploader_id': 'UCH1dpzjCEiGAt8CXkryhkZg', 'uploader_id': 'UCH1dpzjCEiGAt8CXkryhkZg',
@ -871,7 +892,8 @@ class YoutubeIE(YoutubeBaseInfoExtractor):
'id': 'iqKdEhx-dD4', 'id': 'iqKdEhx-dD4',
'ext': 'mp4', 'ext': 'mp4',
'title': 'Isolation - Mind Field (Ep 1)', 'title': 'Isolation - Mind Field (Ep 1)',
'description': 'md5:3a72f23c086a1496c9e2c54a25fa0822', 'description': 'md5:8013b7ddea787342608f63a13ddc9492',
'duration': 2085,
'upload_date': '20170118', 'upload_date': '20170118',
'uploader': 'Vsauce', 'uploader': 'Vsauce',
'uploader_id': 'Vsauce', 'uploader_id': 'Vsauce',
@ -1516,11 +1538,11 @@ def _extract_count(count_name):
video_subtitles = self.extract_subtitles(video_id, video_webpage) video_subtitles = self.extract_subtitles(video_id, video_webpage)
automatic_captions = self.extract_automatic_captions(video_id, video_webpage) automatic_captions = self.extract_automatic_captions(video_id, video_webpage)
if 'length_seconds' not in video_info: video_duration = try_get(
self._downloader.report_warning('unable to extract video duration') video_info, lambda x: int_or_none(x['length_seconds'][0]))
video_duration = None if not video_duration:
else: video_duration = parse_duration(self._html_search_meta(
video_duration = int(compat_urllib_parse_unquote_plus(video_info['length_seconds'][0])) 'duration', video_webpage, 'video duration'))
# annotations # annotations
video_annotations = None video_annotations = None