_ Git - youtube-dl/blob - youtube_dl/extractor/generic.py

   1 # encoding: utf-8
   2
   3 from __future__ import unicode_literals
   4
   5 import os
   6 import re
   7
   8 from .common import InfoExtractor
   9 from .youtube import YoutubeIE
  10 from ..utils import (
  11     compat_urllib_error,
  12     compat_urllib_parse,
  13     compat_urllib_request,
  14     compat_urlparse,
  15     compat_xml_parse_error,
  16
  17     ExtractorError,
  18     float_or_none,
  19     HEADRequest,
  20     orderedSet,
  21     parse_xml,
  22     smuggle_url,
  23     unescapeHTML,
  24     unified_strdate,
  25     unsmuggle_url,
  26     url_basename,
  27 )
  28 from .brightcove import BrightcoveIE
  29 from .ooyala import OoyalaIE
  30 from .rutv import RUTVIE
  31 from .smotri import SmotriIE
  32
  33
  34 class GenericIE(InfoExtractor):
  35     IE_DESC = 'Generic downloader that works on some sites'
  36     _VALID_URL = r'.*'
  37     IE_NAME = 'generic'
  38     _TESTS = [
  39         {
  40             'url': 'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
  41             'md5': '85b90ccc9d73b4acd9138d3af4c27f89',
  42             'info_dict': {
  43                 'id': '13601338388002',
  44                 'ext': 'mp4',
  45                 'uploader': 'www.hodiho.fr',
  46                 'title': 'R\u00e9gis plante sa Jeep',
  47             }
  48         },
  49         # bandcamp page with custom domain
  50         {
  51             'add_ie': ['Bandcamp'],
  52             'url': 'http://bronyrock.com/track/the-pony-mash',
  53             'info_dict': {
  54                 'id': '3235767654',
  55                 'ext': 'mp3',
  56                 'title': 'The Pony Mash',
  57                 'uploader': 'M_Pallante',
  58             },
  59             'skip': 'There is a limit of 200 free downloads / month for the test song',
  60         },
  61         # embedded brightcove video
  62         # it also tests brightcove videos that need to set the 'Referer' in the
  63         # http requests
  64         {
  65             'add_ie': ['Brightcove'],
  66             'url': 'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/',
  67             'info_dict': {
  68                 'id': '2765128793001',
  69                 'ext': 'mp4',
  70                 'title': 'Le cours de bourse : l’analyse technique',
  71                 'description': 'md5:7e9ad046e968cb2d1114004aba466fd9',
  72                 'uploader': 'BFM BUSINESS',
  73             },
  74             'params': {
  75                 'skip_download': True,
  76             },
  77         },
  78         {
  79             # https://github.com/rg3/youtube-dl/issues/2253
  80             'url': 'http://bcove.me/i6nfkrc3',
  81             'md5': '0ba9446db037002366bab3b3eb30c88c',
  82             'info_dict': {
  83                 'id': '3101154703001',
  84                 'ext': 'mp4',
  85                 'title': 'Still no power',
  86                 'uploader': 'thestar.com',
  87                 'description': 'Mississauga resident David Farmer is still out of power as a result of the ice storm a month ago. To keep the house warm, Farmer cuts wood from his property for a wood burning stove downstairs.',
  88             },
  89             'add_ie': ['Brightcove'],
  90         },
  91         {
  92             'url': 'http://www.championat.com/video/football/v/87/87499.html',
  93             'md5': 'fb973ecf6e4a78a67453647444222983',
  94             'info_dict': {
  95                 'id': '3414141473001',
  96                 'ext': 'mp4',
  97                 'title': 'Видео. Удаление Дзагоева (ЦСКА)',
  98                 'description': 'Онлайн-трансляция матча ЦСКА - "Волга"',
  99                 'uploader': 'Championat',
 100             },
 101         },
 102         # Direct link to a video
 103         {
 104             'url': 'http://media.w3.org/2010/05/sintel/trailer.mp4',
 105             'md5': '67d406c2bcb6af27fa886f31aa934bbe',
 106             'info_dict': {
 107                 'id': 'trailer',
 108                 'ext': 'mp4',
 109                 'title': 'trailer',
 110                 'upload_date': '20100513',
 111             }
 112         },
 113         # ooyala video
 114         {
 115             'url': 'http://www.rollingstone.com/music/videos/norwegian-dj-cashmere-cat-goes-spartan-on-with-me-premiere-20131219',
 116             'md5': '5644c6ca5d5782c1d0d350dad9bd840c',
 117             'info_dict': {
 118                 'id': 'BwY2RxaTrTkslxOfcan0UCf0YqyvWysJ',
 119                 'ext': 'mp4',
 120                 'title': '2cc213299525360.mov',  # that's what we get
 121             },
 122         },
 123         # google redirect
 124         {
 125             'url': 'http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&ved=0CCUQtwIwAA&url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DcmQHVoWB5FY&ei=F-sNU-LLCaXk4QT52ICQBQ&usg=AFQjCNEw4hL29zgOohLXvpJ-Bdh2bils1Q&bvm=bv.61965928,d.bGE',
 126             'info_dict': {
 127                 'id': 'cmQHVoWB5FY',
 128                 'ext': 'mp4',
 129                 'upload_date': '20130224',
 130                 'uploader_id': 'TheVerge',
 131                 'description': 'Chris Ziegler takes a look at the Alcatel OneTouch Fire and the ZTE Open; two of the first Firefox OS handsets to be officially announced.',
 132                 'uploader': 'The Verge',
 133                 'title': 'First Firefox OS phones side-by-side',
 134             },
 135             'params': {
 136                 'skip_download': False,
 137             }
 138         },
 139         # embed.ly video
 140         {
 141             'url': 'http://www.tested.com/science/weird/460206-tested-grinding-coffee-2000-frames-second/',
 142             'info_dict': {
 143                 'id': '9ODmcdjQcHQ',
 144                 'ext': 'mp4',
 145                 'title': 'Tested: Grinding Coffee at 2000 Frames Per Second',
 146                 'upload_date': '20140225',
 147                 'description': 'md5:06a40fbf30b220468f1e0957c0f558ff',
 148                 'uploader': 'Tested',
 149                 'uploader_id': 'testedcom',
 150             },
 151             # No need to test YoutubeIE here
 152             'params': {
 153                 'skip_download': True,
 154             },
 155         },
 156         # funnyordie embed
 157         {
 158             'url': 'http://www.theguardian.com/world/2014/mar/11/obama-zach-galifianakis-between-two-ferns',
 159             'md5': '7cf780be104d40fea7bae52eed4a470e',
 160             'info_dict': {
 161                 'id': '18e820ec3f',
 162                 'ext': 'mp4',
 163                 'title': 'Between Two Ferns with Zach Galifianakis: President Barack Obama',
 164                 'description': 'Episode 18: President Barack Obama sits down with Zach Galifianakis for his most memorable interview yet.',
 165             },
 166         },
 167         # RUTV embed
 168         {
 169             'url': 'http://www.rg.ru/2014/03/15/reg-dfo/anklav-anons.html',
 170             'info_dict': {
 171                 'id': '776940',
 172                 'ext': 'mp4',
 173                 'title': 'Охотское море стало целиком российским',
 174                 'description': 'md5:5ed62483b14663e2a95ebbe115eb8f43',
 175             },
 176             'params': {
 177                 # m3u8 download
 178                 'skip_download': True,
 179             },
 180         },
 181         # Embedded TED video
 182         {
 183             'url': 'http://en.support.wordpress.com/videos/ted-talks/',
 184             'md5': 'deeeabcc1085eb2ba205474e7235a3d5',
 185             'info_dict': {
 186                 'id': '981',
 187                 'ext': 'mp4',
 188                 'title': 'My web playroom',
 189                 'uploader': 'Ze Frank',
 190                 'description': 'md5:ddb2a40ecd6b6a147e400e535874947b',
 191             }
 192         },
 193         # Embeded Ustream video
 194         {
 195             'url': 'http://www.american.edu/spa/pti/nsa-privacy-janus-2014.cfm',
 196             'md5': '27b99cdb639c9b12a79bca876a073417',
 197             'info_dict': {
 198                 'id': '45734260',
 199                 'ext': 'flv',
 200                 'uploader': 'AU SPA:  The NSA and Privacy',
 201                 'title': 'NSA and Privacy Forum Debate featuring General Hayden and Barton Gellman'
 202             }
 203         },
 204         # nowvideo embed hidden behind percent encoding
 205         {
 206             'url': 'http://www.waoanime.tv/the-super-dimension-fortress-macross-episode-1/',
 207             'md5': '2baf4ddd70f697d94b1c18cf796d5107',
 208             'info_dict': {
 209                 'id': '06e53103ca9aa',
 210                 'ext': 'flv',
 211                 'title': 'Macross Episode 001  Watch Macross Episode 001 onl',
 212                 'description': 'No description',
 213             },
 214         },
 215         # arte embed
 216         {
 217             'url': 'http://www.tv-replay.fr/redirection/20-03-14/x-enius-arte-10753389.html',
 218             'md5': '7653032cbb25bf6c80d80f217055fa43',
 219             'info_dict': {
 220                 'id': '048195-004_PLUS7-F',
 221                 'ext': 'flv',
 222                 'title': 'X:enius',
 223                 'description': 'md5:d5fdf32ef6613cdbfd516ae658abf168',
 224                 'upload_date': '20140320',
 225             },
 226             'params': {
 227                 'skip_download': 'Requires rtmpdump'
 228             }
 229         },
 230         # smotri embed
 231         {
 232             'url': 'http://rbctv.rbc.ru/archive/news/562949990879132.shtml',
 233             'md5': 'ec40048448e9284c9a1de77bb188108b',
 234             'info_dict': {
 235                 'id': 'v27008541fad',
 236                 'ext': 'mp4',
 237                 'title': 'Крым и Севастополь вошли в состав России',
 238                 'description': 'md5:fae01b61f68984c7bd2fa741e11c3175',
 239                 'duration': 900,
 240                 'upload_date': '20140318',
 241                 'uploader': 'rbctv_2012_4',
 242                 'uploader_id': 'rbctv_2012_4',
 243             },
 244         },
 245         # Condé Nast embed
 246         {
 247             'url': 'http://www.wired.com/2014/04/honda-asimo/',
 248             'md5': 'ba0dfe966fa007657bd1443ee672db0f',
 249             'info_dict': {
 250                 'id': '53501be369702d3275860000',
 251                 'ext': 'mp4',
 252                 'title': 'Honda’s  New Asimo Robot Is More Human Than Ever',
 253             }
 254         },
 255         # Dailymotion embed
 256         {
 257             'url': 'http://www.spi0n.com/zap-spi0n-com-n216/',
 258             'md5': '441aeeb82eb72c422c7f14ec533999cd',
 259             'info_dict': {
 260                 'id': 'k2mm4bCdJ6CQ2i7c8o2',
 261                 'ext': 'mp4',
 262                 'title': 'Le Zap de Spi0n n°216 - Zapping du Web',
 263                 'uploader': 'Spi0n',
 264             },
 265             'add_ie': ['Dailymotion'],
 266         },
 267         # YouTube embed
 268         {
 269             'url': 'http://www.badzine.de/ansicht/datum/2014/06/09/so-funktioniert-die-neue-englische-badminton-liga.html',
 270             'info_dict': {
 271                 'id': 'FXRb4ykk4S0',
 272                 'ext': 'mp4',
 273                 'title': 'The NBL Auction 2014',
 274                 'uploader': 'BADMINTON England',
 275                 'uploader_id': 'BADMINTONEvents',
 276                 'upload_date': '20140603',
 277                 'description': 'md5:9ef128a69f1e262a700ed83edb163a73',
 278             },
 279             'add_ie': ['Youtube'],
 280             'params': {
 281                 'skip_download': True,
 282             }
 283         },
 284         # MTVSercices embed
 285         {
 286             'url': 'http://www.gametrailers.com/news-post/76093/north-america-europe-is-getting-that-mario-kart-8-mercedes-dlc-too',
 287             'md5': '35727f82f58c76d996fc188f9755b0d5',
 288             'info_dict': {
 289                 'id': '0306a69b-8adf-4fb5-aace-75f8e8cbfca9',
 290                 'ext': 'mp4',
 291                 'title': 'Review',
 292                 'description': 'Mario\'s life in the fast lane has never looked so good.',
 293             },
 294         },
 295         # YouTube embed via <data-embed-url="">
 296         {
 297             'url': 'https://play.google.com/store/apps/details?id=com.gameloft.android.ANMP.GloftA8HM',
 298             'info_dict': {
 299                 'id': 'jpSGZsgga_I',
 300                 'ext': 'mp4',
 301                 'title': 'Asphalt 8: Airborne - Launch Trailer',
 302                 'uploader': 'Gameloft',
 303                 'uploader_id': 'gameloft',
 304                 'upload_date': '20130821',
 305                 'description': 'md5:87bd95f13d8be3e7da87a5f2c443106a',
 306             },
 307             'params': {
 308                 'skip_download': True,
 309             }
 310         },
 311         # Camtasia studio
 312         {
 313             'url': 'http://www.ll.mit.edu/workshops/education/videocourses/antennas/lecture1/video/',
 314             'playlist': [{
 315                 'md5': '0c5e352edabf715d762b0ad4e6d9ee67',
 316                 'info_dict': {
 317                     'id': 'Fenn-AA_PA_Radar_Course_Lecture_1c_Final',
 318                     'title': 'Fenn-AA_PA_Radar_Course_Lecture_1c_Final - video1',
 319                     'ext': 'flv',
 320                     'duration': 2235.90,
 321                 }
 322             }, {
 323                 'md5': '10e4bb3aaca9fd630e273ff92d9f3c63',
 324                 'info_dict': {
 325                     'id': 'Fenn-AA_PA_Radar_Course_Lecture_1c_Final_PIP',
 326                     'title': 'Fenn-AA_PA_Radar_Course_Lecture_1c_Final - pip',
 327                     'ext': 'flv',
 328                     'duration': 2235.93,
 329                 }
 330             }],
 331             'info_dict': {
 332                 'title': 'Fenn-AA_PA_Radar_Course_Lecture_1c_Final',
 333             }
 334         },
 335         # Flowplayer
 336         {
 337             'url': 'http://www.handjobhub.com/video/busty-blonde-siri-tit-fuck-while-wank-6313.html',
 338             'md5': '9d65602bf31c6e20014319c7d07fba27',
 339             'info_dict': {
 340                 'id': '5123ea6d5e5a7',
 341                 'ext': 'mp4',
 342                 'age_limit': 18,
 343                 'uploader': 'www.handjobhub.com',
 344                 'title': 'Busty Blonde Siri Tit Fuck While Wank at Handjob Hub',
 345             }
 346         }
 347     ]
 348
 349     def report_download_webpage(self, video_id):
 350         """Report webpage download."""
 351         if not self._downloader.params.get('test', False):
 352             self._downloader.report_warning('Falling back on generic information extractor.')
 353         super(GenericIE, self).report_download_webpage(video_id)
 354
 355     def report_following_redirect(self, new_url):
 356         """Report information extraction."""
 357         self._downloader.to_screen('[redirect] Following redirect to %s' % new_url)
 358
 359     def _send_head(self, url):
 360         """Check if it is a redirect, like url shorteners, in case return the new url."""
 361
 362         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
 363             """
 364             Subclass the HTTPRedirectHandler to make it use our
 365             HEADRequest also on the redirected URL
 366             """
 367             def redirect_request(self, req, fp, code, msg, headers, newurl):
 368                 if code in (301, 302, 303, 307):
 369                     newurl = newurl.replace(' ', '%20')
 370                     newheaders = dict((k,v) for k,v in req.headers.items()
 371                                       if k.lower() not in ("content-length", "content-type"))
 372                     try:
 373                         # This function was deprecated in python 3.3 and removed in 3.4
 374                         origin_req_host = req.get_origin_req_host()
 375                     except AttributeError:
 376                         origin_req_host = req.origin_req_host
 377                     return HEADRequest(newurl,
 378                                        headers=newheaders,
 379                                        origin_req_host=origin_req_host,
 380                                        unverifiable=True)
 381                 else:
 382                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
 383
 384         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
 385             """
 386             Fallback to GET if HEAD is not allowed (405 HTTP error)
 387             """
 388             def http_error_405(self, req, fp, code, msg, headers):
 389                 fp.read()
 390                 fp.close()
 391
 392                 newheaders = dict((k,v) for k,v in req.headers.items()
 393                                   if k.lower() not in ("content-length", "content-type"))
 394                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
 395                                                  headers=newheaders,
 396                                                  origin_req_host=req.get_origin_req_host(),
 397                                                  unverifiable=True))
 398
 399         # Build our opener
 400         opener = compat_urllib_request.OpenerDirector()
 401         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
 402                         HTTPMethodFallback, HEADRedirectHandler,
 403                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
 404             opener.add_handler(handler())
 405
 406         response = opener.open(HEADRequest(url))
 407         if response is None:
 408             raise ExtractorError('Invalid URL protocol')
 409         return response
 410
 411     def _extract_rss(self, url, video_id, doc):
 412         playlist_title = doc.find('./channel/title').text
 413         playlist_desc_el = doc.find('./channel/description')
 414         playlist_desc = None if playlist_desc_el is None else playlist_desc_el.text
 415
 416         entries = [{
 417             '_type': 'url',
 418             'url': e.find('link').text,
 419             'title': e.find('title').text,
 420         } for e in doc.findall('./channel/item')]
 421
 422         return {
 423             '_type': 'playlist',
 424             'id': url,
 425             'title': playlist_title,
 426             'description': playlist_desc,
 427             'entries': entries,
 428         }
 429
 430     def _extract_camtasia(self, url, video_id, webpage):
 431         """ Returns None if no camtasia video can be found. """
 432
 433         camtasia_cfg = self._search_regex(
 434             r'fo\.addVariable\(\s*"csConfigFile",\s*"([^"]+)"\s*\);',
 435             webpage, 'camtasia configuration file', default=None)
 436         if camtasia_cfg is None:
 437             return None
 438
 439         title = self._html_search_meta('DC.title', webpage, fatal=True)
 440
 441         camtasia_url = compat_urlparse.urljoin(url, camtasia_cfg)
 442         camtasia_cfg = self._download_xml(
 443             camtasia_url, video_id,
 444             note='Downloading camtasia configuration',
 445             errnote='Failed to download camtasia configuration')
 446         fileset_node = camtasia_cfg.find('./playlist/array/fileset')
 447
 448         entries = []
 449         for n in fileset_node.getchildren():
 450             url_n = n.find('./uri')
 451             if url_n is None:
 452                 continue
 453
 454             entries.append({
 455                 'id': os.path.splitext(url_n.text.rpartition('/')[2])[0],
 456                 'title': '%s - %s' % (title, n.tag),
 457                 'url': compat_urlparse.urljoin(url, url_n.text),
 458                 'duration': float_or_none(n.find('./duration').text),
 459             })
 460
 461         return {
 462             '_type': 'playlist',
 463             'entries': entries,
 464             'title': title,
 465         }
 466
 467     def _real_extract(self, url):
 468         if url.startswith('//'):
 469             return {
 470                 '_type': 'url',
 471                 'url': self.http_scheme() + url,
 472             }
 473
 474         parsed_url = compat_urlparse.urlparse(url)
 475         if not parsed_url.scheme:
 476             default_search = self._downloader.params.get('default_search')
 477             if default_search is None:
 478                 default_search = 'fixup_error'
 479
 480             if default_search in ('auto', 'auto_warning', 'fixup_error'):
 481                 if '/' in url:
 482                     self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
 483                     return self.url_result('http://' + url)
 484                 elif default_search != 'fixup_error':
 485                     if default_search == 'auto_warning':
 486                         if re.match(r'^(?:url|URL)$', url):
 487                             raise ExtractorError(
 488                                 'Invalid URL:  %r . Call youtube-dl like this:  youtube-dl -v "https://www.youtube.com/watch?v=BaW_jenozKc"  ' % url,
 489                                 expected=True)
 490                         else:
 491                             self._downloader.report_warning(
 492                                 'Falling back to youtube search for  %s . Set --default-search "auto" to suppress this warning.' % url)
 493                     return self.url_result('ytsearch:' + url)
 494
 495             if default_search in ('error', 'fixup_error'):
 496                 raise ExtractorError(
 497                     ('%r is not a valid URL. '
 498                      'Set --default-search "ytsearch" (or run  youtube-dl "ytsearch:%s" ) to search YouTube'
 499                     ) % (url, url), expected=True)
 500             else:
 501                 assert ':' in default_search
 502                 return self.url_result(default_search + url)
 503
 504         url, smuggled_data = unsmuggle_url(url)
 505         force_videoid = None
 506         if smuggled_data and 'force_videoid' in smuggled_data:
 507             force_videoid = smuggled_data['force_videoid']
 508             video_id = force_videoid
 509         else:
 510             video_id = os.path.splitext(url.rstrip('/').split('/')[-1])[0]
 511
 512         self.to_screen('%s: Requesting header' % video_id)
 513
 514         try:
 515             response = self._send_head(url)
 516
 517             # Check for redirect
 518             new_url = response.geturl()
 519             if url != new_url:
 520                 self.report_following_redirect(new_url)
 521                 if force_videoid:
 522                     new_url = smuggle_url(
 523                         new_url, {'force_videoid': force_videoid})
 524                 return self.url_result(new_url)
 525
 526             # Check for direct link to a video
 527             content_type = response.headers.get('Content-Type', '')
 528             m = re.match(r'^(?P<type>audio|video|application(?=/ogg$))/(?P<format_id>.+)$', content_type)
 529             if m:
 530                 upload_date = response.headers.get('Last-Modified')
 531                 if upload_date:
 532                     upload_date = unified_strdate(upload_date)
 533                 return {
 534                     'id': video_id,
 535                     'title': os.path.splitext(url_basename(url))[0],
 536                     'formats': [{
 537                         'format_id': m.group('format_id'),
 538                         'url': url,
 539                         'vcodec': 'none' if m.group('type') == 'audio' else None
 540                     }],
 541                     'upload_date': upload_date,
 542                 }
 543
 544         except compat_urllib_error.HTTPError:
 545             # This may be a stupid server that doesn't like HEAD, our UA, or so
 546             pass
 547
 548         try:
 549             webpage = self._download_webpage(url, video_id)
 550         except ValueError:
 551             # since this is the last-resort InfoExtractor, if
 552             # this error is thrown, it'll be thrown here
 553             raise ExtractorError('Failed to download URL: %s' % url)
 554
 555         self.report_extraction(video_id)
 556
 557         # Is it an RSS feed?
 558         try:
 559             doc = parse_xml(webpage)
 560             if doc.tag == 'rss':
 561                 return self._extract_rss(url, video_id, doc)
 562         except compat_xml_parse_error:
 563             pass
 564
 565         # Is it a Camtasia project?
 566         camtasia_res = self._extract_camtasia(url, video_id, webpage)
 567         if camtasia_res is not None:
 568             return camtasia_res
 569
 570         # Sometimes embedded video player is hidden behind percent encoding
 571         # (e.g. https://github.com/rg3/youtube-dl/issues/2448)
 572         # Unescaping the whole page allows to handle those cases in a generic way
 573         webpage = compat_urllib_parse.unquote(webpage)
 574
 575         # it's tempting to parse this further, but you would
 576         # have to take into account all the variations like
 577         #   Video Title - Site Name
 578         #   Site Name | Video Title
 579         #   Video Title - Tagline | Site Name
 580         # and so on and so forth; it's just not practical
 581         video_title = self._html_search_regex(
 582             r'(?s)<title>(.*?)</title>', webpage, 'video title',
 583             default='video')
 584
 585         # Try to detect age limit automatically
 586         age_limit = self._rta_search(webpage)
 587         # And then there are the jokers who advertise that they use RTA,
 588         # but actually don't.
 589         AGE_LIMIT_MARKERS = [
 590             r'Proudly Labeled <a href="http://www.rtalabel.org/" title="Restricted to Adults">RTA</a>',
 591         ]
 592         if any(re.search(marker, webpage) for marker in AGE_LIMIT_MARKERS):
 593             age_limit = 18
 594
 595         # video uploader is domain name
 596         video_uploader = self._search_regex(
 597             r'^(?:https?://)?([^/]*)/.*', url, 'video uploader')
 598
 599         # Helper method
 600         def _playlist_from_matches(matches, getter, ie=None):
 601             urlrs = orderedSet(self.url_result(getter(m), ie) for m in matches)
 602             return self.playlist_result(
 603                 urlrs, playlist_id=video_id, playlist_title=video_title)
 604
 605         # Look for BrightCove:
 606         bc_urls = BrightcoveIE._extract_brightcove_urls(webpage)
 607         if bc_urls:
 608             self.to_screen('Brightcove video detected.')
 609             entries = [{
 610                 '_type': 'url',
 611                 'url': smuggle_url(bc_url, {'Referer': url}),
 612                 'ie_key': 'Brightcove'
 613             } for bc_url in bc_urls]
 614
 615             return {
 616                 '_type': 'playlist',
 617                 'title': video_title,
 618                 'id': video_id,
 619                 'entries': entries,
 620             }
 621
 622         # Look for embedded (iframe) Vimeo player
 623         mobj = re.search(
 624             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//player\.vimeo\.com/video/.+?)\1', webpage)
 625         if mobj:
 626             player_url = unescapeHTML(mobj.group('url'))
 627             surl = smuggle_url(player_url, {'Referer': url})
 628             return self.url_result(surl, 'Vimeo')
 629
 630         # Look for embedded (swf embed) Vimeo player
 631         mobj = re.search(
 632             r'<embed[^>]+?src="(https?://(?:www\.)?vimeo\.com/moogaloop\.swf.+?)"', webpage)
 633         if mobj:
 634             return self.url_result(mobj.group(1), 'Vimeo')
 635
 636         # Look for embedded YouTube player
 637         matches = re.findall(r'''(?x)
 638             (?:
 639                 <iframe[^>]+?src=|
 640                 data-video-url=|
 641                 <embed[^>]+?src=|
 642                 embedSWF\(?:\s*
 643             )
 644             (["\'])
 645                 (?P<url>(?:https?:)?//(?:www\.)?youtube\.com/
 646                 (?:embed|v)/.+?)
 647             \1''', webpage)
 648         if matches:
 649             return _playlist_from_matches(
 650                 matches, lambda m: unescapeHTML(m[1]), ie='Youtube')
 651
 652         # Look for embedded Dailymotion player
 653         matches = re.findall(
 654             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?dailymotion\.com/embed/video/.+?)\1', webpage)
 655         if matches:
 656             return _playlist_from_matches(
 657                 matches, lambda m: unescapeHTML(m[1]))
 658
 659         # Look for embedded Wistia player
 660         match = re.search(
 661             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage)
 662         if match:
 663             return {
 664                 '_type': 'url_transparent',
 665                 'url': unescapeHTML(match.group('url')),
 666                 'ie_key': 'Wistia',
 667                 'uploader': video_uploader,
 668                 'title': video_title,
 669                 'id': video_id,
 670             }
 671
 672         # Look for embedded blip.tv player
 673         mobj = re.search(r'<meta\s[^>]*https?://api\.blip\.tv/\w+/redirect/\w+/(\d+)', webpage)
 674         if mobj:
 675             return self.url_result('http://blip.tv/a/a-'+mobj.group(1), 'BlipTV')
 676         mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*(https?://(?:\w+\.)?blip\.tv/(?:play/|api\.swf#)[a-zA-Z0-9_]+)', webpage)
 677         if mobj:
 678             return self.url_result(mobj.group(1), 'BlipTV')
 679
 680         # Look for embedded condenast player
 681         matches = re.findall(
 682             r'<iframe\s+(?:[a-zA-Z-]+="[^"]+"\s+)*?src="(https?://player\.cnevids\.com/embed/[^"]+")',
 683             webpage)
 684         if matches:
 685             return {
 686                 '_type': 'playlist',
 687                 'entries': [{
 688                     '_type': 'url',
 689                     'ie_key': 'CondeNast',
 690                     'url': ma,
 691                 } for ma in matches],
 692                 'title': video_title,
 693                 'id': video_id,
 694             }
 695
 696         # Look for Bandcamp pages with custom domain
 697         mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
 698         if mobj is not None:
 699             burl = unescapeHTML(mobj.group(1))
 700             # Don't set the extractor because it can be a track url or an album
 701             return self.url_result(burl)
 702
 703         # Look for embedded Vevo player
 704         mobj = re.search(
 705             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage)
 706         if mobj is not None:
 707             return self.url_result(mobj.group('url'))
 708
 709         # Look for Ooyala videos
 710         mobj = (re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=(?P<ec>[^"&]+)', webpage) or
 711              re.search(r'OO.Player.create\([\'"].*?[\'"],\s*[\'"](?P<ec>.{32})[\'"]', webpage))
 712         if mobj is not None:
 713             return OoyalaIE._build_url_result(mobj.group('ec'))
 714
 715         # Look for Aparat videos
 716         mobj = re.search(r'<iframe .*?src="(http://www\.aparat\.com/video/[^"]+)"', webpage)
 717         if mobj is not None:
 718             return self.url_result(mobj.group(1), 'Aparat')
 719
 720         # Look for MPORA videos
 721         mobj = re.search(r'<iframe .*?src="(http://mpora\.(?:com|de)/videos/[^"]+)"', webpage)
 722         if mobj is not None:
 723             return self.url_result(mobj.group(1), 'Mpora')
 724
 725         # Look for embedded NovaMov-based player
 726         mobj = re.search(
 727             r'''(?x)<(?:pagespeed_)?iframe[^>]+?src=(["\'])
 728                     (?P<url>http://(?:(?:embed|www)\.)?
 729                         (?:novamov\.com|
 730                            nowvideo\.(?:ch|sx|eu|at|ag|co)|
 731                            videoweed\.(?:es|com)|
 732                            movshare\.(?:net|sx|ag)|
 733                            divxstage\.(?:eu|net|ch|co|at|ag))
 734                         /embed\.php.+?)\1''', webpage)
 735         if mobj is not None:
 736             return self.url_result(mobj.group('url'))
 737
 738         # Look for embedded Facebook player
 739         mobj = re.search(
 740             r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
 741         if mobj is not None:
 742             return self.url_result(mobj.group('url'), 'Facebook')
 743
 744         # Look for embedded VK player
 745         mobj = re.search(r'<iframe[^>]+?src=(["\'])(?P<url>https?://vk\.com/video_ext\.php.+?)\1', webpage)
 746         if mobj is not None:
 747             return self.url_result(mobj.group('url'), 'VK')
 748
 749         # Look for embedded ivi player
 750         mobj = re.search(r'<embed[^>]+?src=(["\'])(?P<url>https?://(?:www\.)?ivi\.ru/video/player.+?)\1', webpage)
 751         if mobj is not None:
 752             return self.url_result(mobj.group('url'), 'Ivi')
 753
 754         # Look for embedded Huffington Post player
 755         mobj = re.search(
 756             r'<iframe[^>]+?src=(["\'])(?P<url>https?://embed\.live\.huffingtonpost\.com/.+?)\1', webpage)
 757         if mobj is not None:
 758             return self.url_result(mobj.group('url'), 'HuffPost')
 759
 760         # Look for embed.ly
 761         mobj = re.search(r'class=["\']embedly-card["\'][^>]href=["\'](?P<url>[^"\']+)', webpage)
 762         if mobj is not None:
 763             return self.url_result(mobj.group('url'))
 764         mobj = re.search(r'class=["\']embedly-embed["\'][^>]src=["\'][^"\']*url=(?P<url>[^&]+)', webpage)
 765         if mobj is not None:
 766             return self.url_result(compat_urllib_parse.unquote(mobj.group('url')))
 767
 768         # Look for funnyordie embed
 769         matches = re.findall(r'<iframe[^>]+?src="(https?://(?:www\.)?funnyordie\.com/embed/[^"]+)"', webpage)
 770         if matches:
 771             return _playlist_from_matches(
 772                 matches, getter=unescapeHTML, ie='FunnyOrDie')
 773
 774         # Look for embedded RUTV player
 775         rutv_url = RUTVIE._extract_url(webpage)
 776         if rutv_url:
 777             return self.url_result(rutv_url, 'RUTV')
 778
 779         # Look for embedded TED player
 780         mobj = re.search(
 781             r'<iframe[^>]+?src=(["\'])(?P<url>http://embed\.ted\.com/.+?)\1', webpage)
 782         if mobj is not None:
 783             return self.url_result(mobj.group('url'), 'TED')
 784
 785         # Look for embedded Ustream videos
 786         mobj = re.search(
 787             r'<iframe[^>]+?src=(["\'])(?P<url>http://www\.ustream\.tv/embed/.+?)\1', webpage)
 788         if mobj is not None:
 789             return self.url_result(mobj.group('url'), 'Ustream')
 790
 791         # Look for embedded arte.tv player
 792         mobj = re.search(
 793             r'<script [^>]*?src="(?P<url>http://www\.arte\.tv/playerv2/embed[^"]+)"',
 794             webpage)
 795         if mobj is not None:
 796             return self.url_result(mobj.group('url'), 'ArteTVEmbed')
 797
 798         # Look for embedded smotri.com player
 799         smotri_url = SmotriIE._extract_url(webpage)
 800         if smotri_url:
 801             return self.url_result(smotri_url, 'Smotri')
 802
 803         # Look for embeded soundcloud player
 804         mobj = re.search(
 805             r'<iframe src="(?P<url>https?://(?:w\.)?soundcloud\.com/player[^"]+)"',
 806             webpage)
 807         if mobj is not None:
 808             url = unescapeHTML(mobj.group('url'))
 809             return self.url_result(url)
 810
 811         # Look for embedded vulture.com player
 812         mobj = re.search(
 813             r'<iframe src="(?P<url>https?://video\.vulture\.com/[^"]+)"',
 814             webpage)
 815         if mobj is not None:
 816             url = unescapeHTML(mobj.group('url'))
 817             return self.url_result(url, ie='Vulture')
 818
 819         # Look for embedded mtvservices player
 820         mobj = re.search(
 821             r'<iframe src="(?P<url>https?://media\.mtvnservices\.com/embed/[^"]+)"',
 822             webpage)
 823         if mobj is not None:
 824             url = unescapeHTML(mobj.group('url'))
 825             return self.url_result(url, ie='MTVServicesEmbedded')
 826
 827         # Look for embedded yahoo player
 828         mobj = re.search(
 829             r'<iframe[^>]+?src=(["\'])(?P<url>https?://(?:screen|movies)\.yahoo\.com/.+?\.html\?format=embed)\1',
 830             webpage)
 831         if mobj is not None:
 832             return self.url_result(mobj.group('url'), 'Yahoo')
 833
 834         # Look for embedded sbs.com.au player
 835         mobj = re.search(
 836             r'<iframe[^>]+?src=(["\'])(?P<url>https?://(?:www\.)sbs\.com\.au/ondemand/video/single/.+?)\1',
 837             webpage)
 838         if mobj is not None:
 839             return self.url_result(mobj.group('url'), 'SBS')
 840
 841         # Start with something easy: JW Player in SWFObject
 842         found = re.findall(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
 843         if not found:
 844             # Look for gorilla-vid style embedding
 845             found = re.findall(r'''(?sx)
 846                 (?:
 847                     jw_plugins|
 848                     JWPlayerOptions|
 849                     jwplayer\s*\(\s*["'][^'"]+["']\s*\)\s*\.setup
 850                 )
 851                 .*?file\s*:\s*["\'](.*?)["\']''', webpage)
 852         if not found:
 853             # Broaden the search a little bit
 854             found = re.findall(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
 855         if not found:
 856             # Broaden the findall a little bit: JWPlayer JS loader
 857             found = re.findall(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http(?![^\'"]+\.[0-9]+[\'"])[^\'"]+)["\']', webpage)
 858         if not found:
 859             # Flow player
 860             found = re.findall(r'''(?xs)
 861                 flowplayer\("[^"]+",\s*
 862                     \{[^}]+?\}\s*,
 863                     \s*{[^}]+? ["']?clip["']?\s*:\s*\{\s*
 864                         ["']?url["']?\s*:\s*["']([^"']+)["']
 865             ''', webpage)
 866             assert found
 867         if not found:
 868             # Try to find twitter cards info
 869             found = re.findall(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
 870         if not found:
 871             # We look for Open Graph info:
 872             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
 873             m_video_type = re.findall(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
 874             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
 875             if m_video_type is not None:
 876                 def check_video(vurl):
 877                     vpath = compat_urlparse.urlparse(vurl).path
 878                     return '.' in vpath and not vpath.endswith('.swf')
 879                 found = list(filter(
 880                     check_video,
 881                     re.findall(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)))
 882         if not found:
 883             # HTML5 video
 884             found = re.findall(r'(?s)<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage)
 885         if not found:
 886             found = re.search(
 887                 r'(?i)<meta\s+(?=(?:[a-z-]+="[^"]+"\s+)*http-equiv="refresh")'
 888                 r'(?:[a-z-]+="[^"]+"\s+)*?content="[0-9]{,2};url=\'([^\']+)\'"',
 889                 webpage)
 890             if found:
 891                 new_url = found.group(1)
 892                 self.report_following_redirect(new_url)
 893                 return {
 894                     '_type': 'url',
 895                     'url': new_url,
 896                 }
 897         if not found:
 898             raise ExtractorError('Unsupported URL: %s' % url)
 899
 900         entries = []
 901         for video_url in found:
 902             video_url = compat_urlparse.urljoin(url, video_url)
 903             video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
 904
 905             # Sometimes, jwplayer extraction will result in a YouTube URL
 906             if YoutubeIE.suitable(video_url):
 907                 entries.append(self.url_result(video_url, 'Youtube'))
 908                 continue
 909
 910             # here's a fun little line of code for you:
 911             video_id = os.path.splitext(video_id)[0]
 912
 913             entries.append({
 914                 'id': video_id,
 915                 'url': video_url,
 916                 'uploader': video_uploader,
 917                 'title': video_title,
 918                 'age_limit': age_limit,
 919             })
 920
 921         if len(entries) == 1:
 922             return entries[0]
 923         else:
 924             for num, e in enumerate(entries, start=1):
 925                 e['title'] = '%s (%d)' % (e['title'], num)
 926             return {
 927                 '_type': 'playlist',
 928                 'entries': entries,
 929             }
 930