Raphaël G. Git Repositories - youtubedl/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     HEADRequest,
  19     parse_xml,
  20     smuggle_url,
  21     unescapeHTML,
  22     unified_strdate,
  23     url_basename,
  24 )
  25 from .brightcove import BrightcoveIE
  26 from .ooyala import OoyalaIE
  27 from .rutv import RUTVIE
  28 from .smotri import SmotriIE
  29
  30
  31 class GenericIE(InfoExtractor):
  32     IE_DESC = 'Generic downloader that works on some sites'
  33     _VALID_URL = r'.*'
  34     IE_NAME = 'generic'
  35     _TESTS = [
  36         {
  37             'url': 'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
  38             'md5': '85b90ccc9d73b4acd9138d3af4c27f89',
  39             'info_dict': {
  40                 'id': '13601338388002',
  41                 'ext': 'mp4',
  42                 'uploader': 'www.hodiho.fr',
  43                 'title': 'R\u00e9gis plante sa Jeep',
  44             }
  45         },
  46         # bandcamp page with custom domain
  47         {
  48             'add_ie': ['Bandcamp'],
  49             'url': 'http://bronyrock.com/track/the-pony-mash',
  50             'info_dict': {
  51                 'id': '3235767654',
  52                 'ext': 'mp3',
  53                 'title': 'The Pony Mash',
  54                 'uploader': 'M_Pallante',
  55             },
  56             'skip': 'There is a limit of 200 free downloads / month for the test song',
  57         },
  58         # embedded brightcove video
  59         # it also tests brightcove videos that need to set the 'Referer' in the
  60         # http requests
  61         {
  62             'add_ie': ['Brightcove'],
  63             'url': 'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/',
  64             'info_dict': {
  65                 'id': '2765128793001',
  66                 'ext': 'mp4',
  67                 'title': 'Le cours de bourse : l’analyse technique',
  68                 'description': 'md5:7e9ad046e968cb2d1114004aba466fd9',
  69                 'uploader': 'BFM BUSINESS',
  70             },
  71             'params': {
  72                 'skip_download': True,
  73             },
  74         },
  75         {
  76             # https://github.com/rg3/youtube-dl/issues/2253
  77             'url': 'http://bcove.me/i6nfkrc3',
  78             'md5': '0ba9446db037002366bab3b3eb30c88c',
  79             'info_dict': {
  80                 'id': '3101154703001',
  81                 'ext': 'mp4',
  82                 'title': 'Still no power',
  83                 'uploader': 'thestar.com',
  84                 '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.',
  85             },
  86             'add_ie': ['Brightcove'],
  87         },
  88         {
  89             'url': 'http://www.championat.com/video/football/v/87/87499.html',
  90             'md5': 'fb973ecf6e4a78a67453647444222983',
  91             'info_dict': {
  92                 'id': '3414141473001',
  93                 'ext': 'mp4',
  94                 'title': 'Видео. Удаление Дзагоева (ЦСКА)',
  95                 'description': 'Онлайн-трансляция матча ЦСКА - "Волга"',
  96                 'uploader': 'Championat',
  97             },
  98         },
  99         # Direct link to a video
 100         {
 101             'url': 'http://media.w3.org/2010/05/sintel/trailer.mp4',
 102             'md5': '67d406c2bcb6af27fa886f31aa934bbe',
 103             'info_dict': {
 104                 'id': 'trailer',
 105                 'ext': 'mp4',
 106                 'title': 'trailer',
 107                 'upload_date': '20100513',
 108             }
 109         },
 110         # ooyala video
 111         {
 112             'url': 'http://www.rollingstone.com/music/videos/norwegian-dj-cashmere-cat-goes-spartan-on-with-me-premiere-20131219',
 113             'md5': '5644c6ca5d5782c1d0d350dad9bd840c',
 114             'info_dict': {
 115                 'id': 'BwY2RxaTrTkslxOfcan0UCf0YqyvWysJ',
 116                 'ext': 'mp4',
 117                 'title': '2cc213299525360.mov',  # that's what we get
 118             },
 119         },
 120         # google redirect
 121         {
 122             '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',
 123             'info_dict': {
 124                 'id': 'cmQHVoWB5FY',
 125                 'ext': 'mp4',
 126                 'upload_date': '20130224',
 127                 'uploader_id': 'TheVerge',
 128                 '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.',
 129                 'uploader': 'The Verge',
 130                 'title': 'First Firefox OS phones side-by-side',
 131             },
 132             'params': {
 133                 'skip_download': False,
 134             }
 135         },
 136         # embed.ly video
 137         {
 138             'url': 'http://www.tested.com/science/weird/460206-tested-grinding-coffee-2000-frames-second/',
 139             'info_dict': {
 140                 'id': '9ODmcdjQcHQ',
 141                 'ext': 'mp4',
 142                 'title': 'Tested: Grinding Coffee at 2000 Frames Per Second',
 143                 'upload_date': '20140225',
 144                 'description': 'md5:06a40fbf30b220468f1e0957c0f558ff',
 145                 'uploader': 'Tested',
 146                 'uploader_id': 'testedcom',
 147             },
 148             # No need to test YoutubeIE here
 149             'params': {
 150                 'skip_download': True,
 151             },
 152         },
 153         # funnyordie embed
 154         {
 155             'url': 'http://www.theguardian.com/world/2014/mar/11/obama-zach-galifianakis-between-two-ferns',
 156             'md5': '7cf780be104d40fea7bae52eed4a470e',
 157             'info_dict': {
 158                 'id': '18e820ec3f',
 159                 'ext': 'mp4',
 160                 'title': 'Between Two Ferns with Zach Galifianakis: President Barack Obama',
 161                 'description': 'Episode 18: President Barack Obama sits down with Zach Galifianakis for his most memorable interview yet.',
 162             },
 163         },
 164         # RUTV embed
 165         {
 166             'url': 'http://www.rg.ru/2014/03/15/reg-dfo/anklav-anons.html',
 167             'info_dict': {
 168                 'id': '776940',
 169                 'ext': 'mp4',
 170                 'title': 'Охотское море стало целиком российским',
 171                 'description': 'md5:5ed62483b14663e2a95ebbe115eb8f43',
 172             },
 173             'params': {
 174                 # m3u8 download
 175                 'skip_download': True,
 176             },
 177         },
 178         # Embedded TED video
 179         {
 180             'url': 'http://en.support.wordpress.com/videos/ted-talks/',
 181             'md5': 'deeeabcc1085eb2ba205474e7235a3d5',
 182             'info_dict': {
 183                 'id': '981',
 184                 'ext': 'mp4',
 185                 'title': 'My web playroom',
 186                 'uploader': 'Ze Frank',
 187                 'description': 'md5:ddb2a40ecd6b6a147e400e535874947b',
 188             }
 189         },
 190         # Embeded Ustream video
 191         {
 192             'url': 'http://www.american.edu/spa/pti/nsa-privacy-janus-2014.cfm',
 193             'md5': '27b99cdb639c9b12a79bca876a073417',
 194             'info_dict': {
 195                 'id': '45734260',
 196                 'ext': 'flv',
 197                 'uploader': 'AU SPA:  The NSA and Privacy',
 198                 'title': 'NSA and Privacy Forum Debate featuring General Hayden and Barton Gellman'
 199             }
 200         },
 201         # nowvideo embed hidden behind percent encoding
 202         {
 203             'url': 'http://www.waoanime.tv/the-super-dimension-fortress-macross-episode-1/',
 204             'md5': '2baf4ddd70f697d94b1c18cf796d5107',
 205             'info_dict': {
 206                 'id': '06e53103ca9aa',
 207                 'ext': 'flv',
 208                 'title': 'Macross Episode 001  Watch Macross Episode 001 onl',
 209                 'description': 'No description',
 210             },
 211         },
 212         # arte embed
 213         {
 214             'url': 'http://www.tv-replay.fr/redirection/20-03-14/x-enius-arte-10753389.html',
 215             'md5': '7653032cbb25bf6c80d80f217055fa43',
 216             'info_dict': {
 217                 'id': '048195-004_PLUS7-F',
 218                 'ext': 'flv',
 219                 'title': 'X:enius',
 220                 'description': 'md5:d5fdf32ef6613cdbfd516ae658abf168',
 221                 'upload_date': '20140320',
 222             },
 223             'params': {
 224                 'skip_download': 'Requires rtmpdump'
 225             }
 226         },
 227         # smotri embed
 228         {
 229             'url': 'http://rbctv.rbc.ru/archive/news/562949990879132.shtml',
 230             'md5': 'ec40048448e9284c9a1de77bb188108b',
 231             'info_dict': {
 232                 'id': 'v27008541fad',
 233                 'ext': 'mp4',
 234                 'title': 'Крым и Севастополь вошли в состав России',
 235                 'description': 'md5:fae01b61f68984c7bd2fa741e11c3175',
 236                 'duration': 900,
 237                 'upload_date': '20140318',
 238                 'uploader': 'rbctv_2012_4',
 239                 'uploader_id': 'rbctv_2012_4',
 240             },
 241         },
 242         # Condé Nast embed
 243         {
 244             'url': 'http://www.wired.com/2014/04/honda-asimo/',
 245             'md5': 'ba0dfe966fa007657bd1443ee672db0f',
 246             'info_dict': {
 247                 'id': '53501be369702d3275860000',
 248                 'ext': 'mp4',
 249                 'title': 'Honda’s  New Asimo Robot Is More Human Than Ever',
 250             }
 251         },
 252         # Dailymotion embed
 253         {
 254             'url': 'http://www.spi0n.com/zap-spi0n-com-n216/',
 255             'md5': '441aeeb82eb72c422c7f14ec533999cd',
 256             'info_dict': {
 257                 'id': 'k2mm4bCdJ6CQ2i7c8o2',
 258                 'ext': 'mp4',
 259                 'title': 'Le Zap de Spi0n n°216 - Zapping du Web',
 260                 'uploader': 'Spi0n',
 261             },
 262             'add_ie': ['Dailymotion'],
 263         },
 264         # YouTube embed
 265         {
 266             'url': 'http://www.badzine.de/ansicht/datum/2014/06/09/so-funktioniert-die-neue-englische-badminton-liga.html',
 267             'info_dict': {
 268                 'id': 'FXRb4ykk4S0',
 269                 'ext': 'mp4',
 270                 'title': 'The NBL Auction 2014',
 271                 'uploader': 'BADMINTON England',
 272                 'uploader_id': 'BADMINTONEvents',
 273                 'upload_date': '20140603',
 274                 'description': 'md5:9ef128a69f1e262a700ed83edb163a73',
 275             },
 276             'add_ie': ['Youtube'],
 277             'params': {
 278                 'skip_download': True,
 279             }
 280         },
 281     ]
 282
 283     def report_download_webpage(self, video_id):
 284         """Report webpage download."""
 285         if not self._downloader.params.get('test', False):
 286             self._downloader.report_warning('Falling back on generic information extractor.')
 287         super(GenericIE, self).report_download_webpage(video_id)
 288
 289     def report_following_redirect(self, new_url):
 290         """Report information extraction."""
 291         self._downloader.to_screen('[redirect] Following redirect to %s' % new_url)
 292
 293     def _send_head(self, url):
 294         """Check if it is a redirect, like url shorteners, in case return the new url."""
 295
 296         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
 297             """
 298             Subclass the HTTPRedirectHandler to make it use our
 299             HEADRequest also on the redirected URL
 300             """
 301             def redirect_request(self, req, fp, code, msg, headers, newurl):
 302                 if code in (301, 302, 303, 307):
 303                     newurl = newurl.replace(' ', '%20')
 304                     newheaders = dict((k,v) for k,v in req.headers.items()
 305                                       if k.lower() not in ("content-length", "content-type"))
 306                     try:
 307                         # This function was deprecated in python 3.3 and removed in 3.4
 308                         origin_req_host = req.get_origin_req_host()
 309                     except AttributeError:
 310                         origin_req_host = req.origin_req_host
 311                     return HEADRequest(newurl,
 312                                        headers=newheaders,
 313                                        origin_req_host=origin_req_host,
 314                                        unverifiable=True)
 315                 else:
 316                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
 317
 318         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
 319             """
 320             Fallback to GET if HEAD is not allowed (405 HTTP error)
 321             """
 322             def http_error_405(self, req, fp, code, msg, headers):
 323                 fp.read()
 324                 fp.close()
 325
 326                 newheaders = dict((k,v) for k,v in req.headers.items()
 327                                   if k.lower() not in ("content-length", "content-type"))
 328                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
 329                                                  headers=newheaders,
 330                                                  origin_req_host=req.get_origin_req_host(),
 331                                                  unverifiable=True))
 332
 333         # Build our opener
 334         opener = compat_urllib_request.OpenerDirector()
 335         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
 336                         HTTPMethodFallback, HEADRedirectHandler,
 337                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
 338             opener.add_handler(handler())
 339
 340         response = opener.open(HEADRequest(url))
 341         if response is None:
 342             raise ExtractorError('Invalid URL protocol')
 343         return response
 344
 345     def _extract_rss(self, url, video_id, doc):
 346         playlist_title = doc.find('./channel/title').text
 347         playlist_desc_el = doc.find('./channel/description')
 348         playlist_desc = None if playlist_desc_el is None else playlist_desc_el.text
 349
 350         entries = [{
 351             '_type': 'url',
 352             'url': e.find('link').text,
 353             'title': e.find('title').text,
 354         } for e in doc.findall('./channel/item')]
 355
 356         return {
 357             '_type': 'playlist',
 358             'id': url,
 359             'title': playlist_title,
 360             'description': playlist_desc,
 361             'entries': entries,
 362         }
 363
 364     def _real_extract(self, url):
 365         if url.startswith('//'):
 366             return {
 367                 '_type': 'url',
 368                 'url': self.http_scheme() + url,
 369             }
 370
 371         parsed_url = compat_urlparse.urlparse(url)
 372         if not parsed_url.scheme:
 373             default_search = self._downloader.params.get('default_search')
 374             if default_search is None:
 375                 default_search = 'auto_warning'
 376
 377             if default_search in ('auto', 'auto_warning'):
 378                 if '/' in url:
 379                     self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
 380                     return self.url_result('http://' + url)
 381                 else:
 382                     if default_search == 'auto_warning':
 383                         if re.match(r'^(?:url|URL)$', url):
 384                             raise ExtractorError(
 385                                 'Invalid URL:  %r . Call youtube-dl like this:  youtube-dl -v "https://www.youtube.com/watch?v=BaW_jenozKc"  ' % url,
 386                                 expected=True)
 387                         else:
 388                             self._downloader.report_warning(
 389                                 'Falling back to youtube search for  %s . Set --default-search to "auto" to suppress this warning.' % url)
 390                     return self.url_result('ytsearch:' + url)
 391             else:
 392                 assert ':' in default_search
 393                 return self.url_result(default_search + url)
 394         video_id = os.path.splitext(url.rstrip('/').split('/')[-1])[0]
 395
 396         self.to_screen('%s: Requesting header' % video_id)
 397
 398         try:
 399             response = self._send_head(url)
 400
 401             # Check for redirect
 402             new_url = response.geturl()
 403             if url != new_url:
 404                 self.report_following_redirect(new_url)
 405                 return self.url_result(new_url)
 406
 407             # Check for direct link to a video
 408             content_type = response.headers.get('Content-Type', '')
 409             m = re.match(r'^(?P<type>audio|video|application(?=/ogg$))/(?P<format_id>.+)$', content_type)
 410             if m:
 411                 upload_date = response.headers.get('Last-Modified')
 412                 if upload_date:
 413                     upload_date = unified_strdate(upload_date)
 414                 return {
 415                     'id': video_id,
 416                     'title': os.path.splitext(url_basename(url))[0],
 417                     'formats': [{
 418                         'format_id': m.group('format_id'),
 419                         'url': url,
 420                         'vcodec': 'none' if m.group('type') == 'audio' else None
 421                     }],
 422                     'upload_date': upload_date,
 423                 }
 424
 425         except compat_urllib_error.HTTPError:
 426             # This may be a stupid server that doesn't like HEAD, our UA, or so
 427             pass
 428
 429         try:
 430             webpage = self._download_webpage(url, video_id)
 431         except ValueError:
 432             # since this is the last-resort InfoExtractor, if
 433             # this error is thrown, it'll be thrown here
 434             raise ExtractorError('Failed to download URL: %s' % url)
 435
 436         self.report_extraction(video_id)
 437
 438         # Is it an RSS feed?
 439         try:
 440             doc = parse_xml(webpage)
 441             if doc.tag == 'rss':
 442                 return self._extract_rss(url, video_id, doc)
 443         except compat_xml_parse_error:
 444             pass
 445
 446         # Sometimes embedded video player is hidden behind percent encoding
 447         # (e.g. https://github.com/rg3/youtube-dl/issues/2448)
 448         # Unescaping the whole page allows to handle those cases in a generic way
 449         webpage = compat_urllib_parse.unquote(webpage)
 450
 451         # it's tempting to parse this further, but you would
 452         # have to take into account all the variations like
 453         #   Video Title - Site Name
 454         #   Site Name | Video Title
 455         #   Video Title - Tagline | Site Name
 456         # and so on and so forth; it's just not practical
 457         video_title = self._html_search_regex(
 458             r'(?s)<title>(.*?)</title>', webpage, 'video title',
 459             default='video')
 460
 461         # video uploader is domain name
 462         video_uploader = self._search_regex(
 463             r'^(?:https?://)?([^/]*)/.*', url, 'video uploader')
 464
 465         # Look for BrightCove:
 466         bc_urls = BrightcoveIE._extract_brightcove_urls(webpage)
 467         if bc_urls:
 468             self.to_screen('Brightcove video detected.')
 469             entries = [{
 470                 '_type': 'url',
 471                 'url': smuggle_url(bc_url, {'Referer': url}),
 472                 'ie_key': 'Brightcove'
 473             } for bc_url in bc_urls]
 474
 475             return {
 476                 '_type': 'playlist',
 477                 'title': video_title,
 478                 'id': video_id,
 479                 'entries': entries,
 480             }
 481
 482         # Look for embedded (iframe) Vimeo player
 483         mobj = re.search(
 484             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//player\.vimeo\.com/video/.+?)\1', webpage)
 485         if mobj:
 486             player_url = unescapeHTML(mobj.group('url'))
 487             surl = smuggle_url(player_url, {'Referer': url})
 488             return self.url_result(surl, 'Vimeo')
 489
 490         # Look for embedded (swf embed) Vimeo player
 491         mobj = re.search(
 492             r'<embed[^>]+?src="(https?://(?:www\.)?vimeo\.com/moogaloop\.swf.+?)"', webpage)
 493         if mobj:
 494             return self.url_result(mobj.group(1), 'Vimeo')
 495
 496         # Look for embedded YouTube player
 497         matches = re.findall(r'''(?x)
 498             (?:
 499                 <iframe[^>]+?src=|
 500                 <embed[^>]+?src=|
 501                 embedSWF\(?:\s*
 502             )
 503             (["\'])
 504                 (?P<url>(?:https?:)?//(?:www\.)?youtube\.com/
 505                 (?:embed|v)/.+?)
 506             \1''', webpage)
 507         if matches:
 508             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Youtube')
 509                      for tuppl in matches]
 510             return self.playlist_result(
 511                 urlrs, playlist_id=video_id, playlist_title=video_title)
 512
 513         # Look for embedded Dailymotion player
 514         matches = re.findall(
 515             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?dailymotion\.com/embed/video/.+?)\1', webpage)
 516         if matches:
 517             urlrs = [self.url_result(unescapeHTML(tuppl[1]))
 518                      for tuppl in matches]
 519             return self.playlist_result(
 520                 urlrs, playlist_id=video_id, playlist_title=video_title)
 521
 522         # Look for embedded Wistia player
 523         match = re.search(
 524             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage)
 525         if match:
 526             return {
 527                 '_type': 'url_transparent',
 528                 'url': unescapeHTML(match.group('url')),
 529                 'ie_key': 'Wistia',
 530                 'uploader': video_uploader,
 531                 'title': video_title,
 532                 'id': video_id,
 533             }
 534
 535         # Look for embedded blip.tv player
 536         mobj = re.search(r'<meta\s[^>]*https?://api\.blip\.tv/\w+/redirect/\w+/(\d+)', webpage)
 537         if mobj:
 538             return self.url_result('http://blip.tv/a/a-'+mobj.group(1), 'BlipTV')
 539         mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*(https?://(?:\w+\.)?blip\.tv/(?:play/|api\.swf#)[a-zA-Z0-9]+)', webpage)
 540         if mobj:
 541             return self.url_result(mobj.group(1), 'BlipTV')
 542
 543         # Look for embedded condenast player
 544         matches = re.findall(
 545             r'<iframe\s+(?:[a-zA-Z-]+="[^"]+"\s+)*?src="(https?://player\.cnevids\.com/embed/[^"]+")',
 546             webpage)
 547         if matches:
 548             return {
 549                 '_type': 'playlist',
 550                 'entries': [{
 551                     '_type': 'url',
 552                     'ie_key': 'CondeNast',
 553                     'url': ma,
 554                 } for ma in matches],
 555                 'title': video_title,
 556                 'id': video_id,
 557             }
 558
 559         # Look for Bandcamp pages with custom domain
 560         mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
 561         if mobj is not None:
 562             burl = unescapeHTML(mobj.group(1))
 563             # Don't set the extractor because it can be a track url or an album
 564             return self.url_result(burl)
 565
 566         # Look for embedded Vevo player
 567         mobj = re.search(
 568             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage)
 569         if mobj is not None:
 570             return self.url_result(mobj.group('url'))
 571
 572         # Look for Ooyala videos
 573         mobj = (re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=(?P<ec>[^"&]+)', webpage) or
 574              re.search(r'OO.Player.create\([\'"].*?[\'"],\s*[\'"](?P<ec>.{32})[\'"]', webpage))
 575         if mobj is not None:
 576             return OoyalaIE._build_url_result(mobj.group('ec'))
 577
 578         # Look for Aparat videos
 579         mobj = re.search(r'<iframe .*?src="(http://www\.aparat\.com/video/[^"]+)"', webpage)
 580         if mobj is not None:
 581             return self.url_result(mobj.group(1), 'Aparat')
 582
 583         # Look for MPORA videos
 584         mobj = re.search(r'<iframe .*?src="(http://mpora\.(?:com|de)/videos/[^"]+)"', webpage)
 585         if mobj is not None:
 586             return self.url_result(mobj.group(1), 'Mpora')
 587
 588         # Look for embedded NovaMov-based player
 589         mobj = re.search(
 590             r'''(?x)<(?:pagespeed_)?iframe[^>]+?src=(["\'])
 591                     (?P<url>http://(?:(?:embed|www)\.)?
 592                         (?:novamov\.com|
 593                            nowvideo\.(?:ch|sx|eu|at|ag|co)|
 594                            videoweed\.(?:es|com)|
 595                            movshare\.(?:net|sx|ag)|
 596                            divxstage\.(?:eu|net|ch|co|at|ag))
 597                         /embed\.php.+?)\1''', webpage)
 598         if mobj is not None:
 599             return self.url_result(mobj.group('url'))
 600
 601         # Look for embedded Facebook player
 602         mobj = re.search(
 603             r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
 604         if mobj is not None:
 605             return self.url_result(mobj.group('url'), 'Facebook')
 606
 607         # Look for embedded VK player
 608         mobj = re.search(r'<iframe[^>]+?src=(["\'])(?P<url>https?://vk\.com/video_ext\.php.+?)\1', webpage)
 609         if mobj is not None:
 610             return self.url_result(mobj.group('url'), 'VK')
 611
 612         # Look for embedded Huffington Post player
 613         mobj = re.search(
 614             r'<iframe[^>]+?src=(["\'])(?P<url>https?://embed\.live\.huffingtonpost\.com/.+?)\1', webpage)
 615         if mobj is not None:
 616             return self.url_result(mobj.group('url'), 'HuffPost')
 617
 618         # Look for embed.ly
 619         mobj = re.search(r'class=["\']embedly-card["\'][^>]href=["\'](?P<url>[^"\']+)', webpage)
 620         if mobj is not None:
 621             return self.url_result(mobj.group('url'))
 622         mobj = re.search(r'class=["\']embedly-embed["\'][^>]src=["\'][^"\']*url=(?P<url>[^&]+)', webpage)
 623         if mobj is not None:
 624             return self.url_result(compat_urllib_parse.unquote(mobj.group('url')))
 625
 626         # Look for funnyordie embed
 627         matches = re.findall(r'<iframe[^>]+?src="(https?://(?:www\.)?funnyordie\.com/embed/[^"]+)"', webpage)
 628         if matches:
 629             urlrs = [self.url_result(unescapeHTML(eurl), 'FunnyOrDie')
 630                      for eurl in matches]
 631             return self.playlist_result(
 632                 urlrs, playlist_id=video_id, playlist_title=video_title)
 633
 634         # Look for embedded RUTV player
 635         rutv_url = RUTVIE._extract_url(webpage)
 636         if rutv_url:
 637             return self.url_result(rutv_url, 'RUTV')
 638
 639         # Look for embedded TED player
 640         mobj = re.search(
 641             r'<iframe[^>]+?src=(["\'])(?P<url>http://embed\.ted\.com/.+?)\1', webpage)
 642         if mobj is not None:
 643             return self.url_result(mobj.group('url'), 'TED')
 644
 645         # Look for embedded Ustream videos
 646         mobj = re.search(
 647             r'<iframe[^>]+?src=(["\'])(?P<url>http://www\.ustream\.tv/embed/.+?)\1', webpage)
 648         if mobj is not None:
 649             return self.url_result(mobj.group('url'), 'Ustream')
 650
 651         # Look for embedded arte.tv player
 652         mobj = re.search(
 653             r'<script [^>]*?src="(?P<url>http://www\.arte\.tv/playerv2/embed[^"]+)"',
 654             webpage)
 655         if mobj is not None:
 656             return self.url_result(mobj.group('url'), 'ArteTVEmbed')
 657
 658         # Look for embedded smotri.com player
 659         smotri_url = SmotriIE._extract_url(webpage)
 660         if smotri_url:
 661             return self.url_result(smotri_url, 'Smotri')
 662
 663         # Look for embeded soundcloud player
 664         mobj = re.search(
 665             r'<iframe src="(?P<url>https?://(?:w\.)?soundcloud\.com/player[^"]+)"',
 666             webpage)
 667         if mobj is not None:
 668             url = unescapeHTML(mobj.group('url'))
 669             return self.url_result(url)
 670
 671         # Look for embedded vulture.com player
 672         mobj = re.search(
 673             r'<iframe src="(?P<url>https?://video\.vulture\.com/[^"]+)"',
 674             webpage)
 675         if mobj is not None:
 676             url = unescapeHTML(mobj.group('url'))
 677             return self.url_result(url, ie='Vulture')
 678
 679         # Start with something easy: JW Player in SWFObject
 680         found = re.findall(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
 681         if not found:
 682             # Look for gorilla-vid style embedding
 683             found = re.findall(r'''(?sx)
 684                 (?:
 685                     jw_plugins|
 686                     JWPlayerOptions|
 687                     jwplayer\s*\(\s*["'][^'"]+["']\s*\)\s*\.setup
 688                 )
 689                 .*?file\s*:\s*["\'](.*?)["\']''', webpage)
 690         if not found:
 691             # Broaden the search a little bit
 692             found = re.findall(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
 693         if not found:
 694             # Broaden the findall a little bit: JWPlayer JS loader
 695             found = re.findall(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http(?![^\'"]+\.[0-9]+[\'"])[^\'"]+)["\']', webpage)
 696         if not found:
 697             # Try to find twitter cards info
 698             found = re.findall(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
 699         if not found:
 700             # We look for Open Graph info:
 701             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
 702             m_video_type = re.findall(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
 703             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
 704             if m_video_type is not None:
 705                 found = re.findall(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
 706         if not found:
 707             # HTML5 video
 708             found = re.findall(r'(?s)<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage)
 709         if not found:
 710             found = re.search(
 711                 r'(?i)<meta\s+(?=(?:[a-z-]+="[^"]+"\s+)*http-equiv="refresh")'
 712                 r'(?:[a-z-]+="[^"]+"\s+)*?content="[0-9]{,2};url=\'([^\']+)\'"',
 713                 webpage)
 714             if found:
 715                 new_url = found.group(1)
 716                 self.report_following_redirect(new_url)
 717                 return {
 718                     '_type': 'url',
 719                     'url': new_url,
 720                 }
 721         if not found:
 722             raise ExtractorError('Unsupported URL: %s' % url)
 723
 724         entries = []
 725         for video_url in found:
 726             video_url = compat_urlparse.urljoin(url, video_url)
 727             video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
 728
 729             # Sometimes, jwplayer extraction will result in a YouTube URL
 730             if YoutubeIE.suitable(video_url):
 731                 entries.append(self.url_result(video_url, 'Youtube'))
 732                 continue
 733
 734             # here's a fun little line of code for you:
 735             video_id = os.path.splitext(video_id)[0]
 736
 737             entries.append({
 738                 'id': video_id,
 739                 'url': video_url,
 740                 'uploader': video_uploader,
 741                 'title': video_title,
 742             })
 743
 744         if len(entries) == 1:
 745             return entries[0]
 746         else:
 747             for num, e in enumerate(entries, start=1):
 748                 e['title'] = '%s (%d)' % (e['title'], num)
 749             return {
 750                 '_type': 'playlist',
 751                 'entries': entries,
 752             }
 753