_ 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
  16     ExtractorError,
  17     HEADRequest,
  18     smuggle_url,
  19     unescapeHTML,
  20     unified_strdate,
  21     url_basename,
  22 )
  23 from .brightcove import BrightcoveIE
  24 from .ooyala import OoyalaIE
  25
  26
  27 class GenericIE(InfoExtractor):
  28     IE_DESC = 'Generic downloader that works on some sites'
  29     _VALID_URL = r'.*'
  30     IE_NAME = 'generic'
  31     _TESTS = [
  32         {
  33             'url': 'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
  34             'file': '13601338388002.mp4',
  35             'md5': '6e15c93721d7ec9e9ca3fdbf07982cfd',
  36             'info_dict': {
  37                 'uploader': 'www.hodiho.fr',
  38                 'title': 'R\u00e9gis plante sa Jeep',
  39             }
  40         },
  41         # embedded vimeo video
  42         {
  43             'add_ie': ['Vimeo'],
  44             'url': 'http://skillsmatter.com/podcast/home/move-semanticsperfect-forwarding-and-rvalue-references',
  45             'file': '22444065.mp4',
  46             'md5': '2903896e23df39722c33f015af0666e2',
  47             'info_dict': {
  48                 'title': 'ACCU 2011: Move Semantics,Perfect Forwarding, and Rvalue references- Scott Meyers- 13/04/2011',
  49                 'uploader_id': 'skillsmatter',
  50                 'uploader': 'Skills Matter',
  51             }
  52         },
  53         # bandcamp page with custom domain
  54         {
  55             'add_ie': ['Bandcamp'],
  56             'url': 'http://bronyrock.com/track/the-pony-mash',
  57             'file': '3235767654.mp3',
  58             'info_dict': {
  59                 'title': 'The Pony Mash',
  60                 'uploader': 'M_Pallante',
  61             },
  62             'skip': 'There is a limit of 200 free downloads / month for the test song',
  63         },
  64         # embedded brightcove video
  65         # it also tests brightcove videos that need to set the 'Referer' in the
  66         # http requests
  67         {
  68             'add_ie': ['Brightcove'],
  69             'url': 'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/',
  70             'info_dict': {
  71                 'id': '2765128793001',
  72                 'ext': 'mp4',
  73                 'title': 'Le cours de bourse : l’analyse technique',
  74                 'description': 'md5:7e9ad046e968cb2d1114004aba466fd9',
  75                 'uploader': 'BFM BUSINESS',
  76             },
  77             'params': {
  78                 'skip_download': True,
  79             },
  80         },
  81         {
  82             # https://github.com/rg3/youtube-dl/issues/2253
  83             'url': 'http://bcove.me/i6nfkrc3',
  84             'file': '3101154703001.mp4',
  85             'md5': '0ba9446db037002366bab3b3eb30c88c',
  86             'info_dict': {
  87                 'title': 'Still no power',
  88                 'uploader': 'thestar.com',
  89                 '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.',
  90             },
  91             'add_ie': ['Brightcove'],
  92         },
  93         # Direct link to a video
  94         {
  95             'url': 'http://media.w3.org/2010/05/sintel/trailer.mp4',
  96             'file': 'trailer.mp4',
  97             'md5': '67d406c2bcb6af27fa886f31aa934bbe',
  98             'info_dict': {
  99                 'id': 'trailer',
 100                 'title': 'trailer',
 101                 'upload_date': '20100513',
 102             }
 103         },
 104         # ooyala video
 105         {
 106             'url': 'http://www.rollingstone.com/music/videos/norwegian-dj-cashmere-cat-goes-spartan-on-with-me-premiere-20131219',
 107             'file': 'BwY2RxaTrTkslxOfcan0UCf0YqyvWysJ.mp4',
 108             'md5': '5644c6ca5d5782c1d0d350dad9bd840c',
 109             'info_dict': {
 110                 'id': 'BwY2RxaTrTkslxOfcan0UCf0YqyvWysJ',
 111                 'ext': 'mp4',
 112                 'title': '2cc213299525360.mov',  # that's what we get
 113             },
 114         },
 115     ]
 116
 117     def report_download_webpage(self, video_id):
 118         """Report webpage download."""
 119         if not self._downloader.params.get('test', False):
 120             self._downloader.report_warning('Falling back on generic information extractor.')
 121         super(GenericIE, self).report_download_webpage(video_id)
 122
 123     def report_following_redirect(self, new_url):
 124         """Report information extraction."""
 125         self._downloader.to_screen('[redirect] Following redirect to %s' % new_url)
 126
 127     def _send_head(self, url):
 128         """Check if it is a redirect, like url shorteners, in case return the new url."""
 129
 130         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
 131             """
 132             Subclass the HTTPRedirectHandler to make it use our
 133             HEADRequest also on the redirected URL
 134             """
 135             def redirect_request(self, req, fp, code, msg, headers, newurl):
 136                 if code in (301, 302, 303, 307):
 137                     newurl = newurl.replace(' ', '%20')
 138                     newheaders = dict((k,v) for k,v in req.headers.items()
 139                                       if k.lower() not in ("content-length", "content-type"))
 140                     return HEADRequest(newurl,
 141                                        headers=newheaders,
 142                                        origin_req_host=req.get_origin_req_host(),
 143                                        unverifiable=True)
 144                 else:
 145                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
 146
 147         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
 148             """
 149             Fallback to GET if HEAD is not allowed (405 HTTP error)
 150             """
 151             def http_error_405(self, req, fp, code, msg, headers):
 152                 fp.read()
 153                 fp.close()
 154
 155                 newheaders = dict((k,v) for k,v in req.headers.items()
 156                                   if k.lower() not in ("content-length", "content-type"))
 157                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
 158                                                  headers=newheaders,
 159                                                  origin_req_host=req.get_origin_req_host(),
 160                                                  unverifiable=True))
 161
 162         # Build our opener
 163         opener = compat_urllib_request.OpenerDirector()
 164         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
 165                         HTTPMethodFallback, HEADRedirectHandler,
 166                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
 167             opener.add_handler(handler())
 168
 169         response = opener.open(HEADRequest(url))
 170         if response is None:
 171             raise ExtractorError('Invalid URL protocol')
 172         return response
 173
 174     def _real_extract(self, url):
 175         parsed_url = compat_urlparse.urlparse(url)
 176         if not parsed_url.scheme:
 177             default_search = self._downloader.params.get('default_search')
 178             if default_search is None:
 179                 default_search = 'auto'
 180
 181             if default_search == 'auto':
 182                 if '/' in url:
 183                     self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
 184                     return self.url_result('http://' + url)
 185                 else:
 186                     return self.url_result('ytsearch:' + url)
 187             else:
 188                 assert ':' in default_search
 189                 return self.url_result(default_search + url)
 190         video_id = os.path.splitext(url.split('/')[-1])[0]
 191
 192         self.to_screen('%s: Requesting header' % video_id)
 193
 194         try:
 195             response = self._send_head(url)
 196
 197             # Check for redirect
 198             new_url = response.geturl()
 199             if url != new_url:
 200                 self.report_following_redirect(new_url)
 201                 return self.url_result(new_url)
 202
 203             # Check for direct link to a video
 204             content_type = response.headers.get('Content-Type', '')
 205             m = re.match(r'^(?P<type>audio|video|application(?=/ogg$))/(?P<format_id>.+)$', content_type)
 206             if m:
 207                 upload_date = response.headers.get('Last-Modified')
 208                 if upload_date:
 209                     upload_date = unified_strdate(upload_date)
 210                 return {
 211                     'id': video_id,
 212                     'title': os.path.splitext(url_basename(url))[0],
 213                     'formats': [{
 214                         'format_id': m.group('format_id'),
 215                         'url': url,
 216                         'vcodec': 'none' if m.group('type') == 'audio' else None
 217                     }],
 218                     'upload_date': upload_date,
 219                 }
 220
 221         except compat_urllib_error.HTTPError:
 222             # This may be a stupid server that doesn't like HEAD, our UA, or so
 223             pass
 224
 225         try:
 226             webpage = self._download_webpage(url, video_id)
 227         except ValueError:
 228             # since this is the last-resort InfoExtractor, if
 229             # this error is thrown, it'll be thrown here
 230             raise ExtractorError('Failed to download URL: %s' % url)
 231
 232         self.report_extraction(video_id)
 233
 234         # it's tempting to parse this further, but you would
 235         # have to take into account all the variations like
 236         #   Video Title - Site Name
 237         #   Site Name | Video Title
 238         #   Video Title - Tagline | Site Name
 239         # and so on and so forth; it's just not practical
 240         video_title = self._html_search_regex(
 241             r'(?s)<title>(.*?)</title>', webpage, 'video title',
 242             default='video')
 243
 244         # video uploader is domain name
 245         video_uploader = self._search_regex(
 246             r'^(?:https?://)?([^/]*)/.*', url, 'video uploader')
 247
 248         # Look for BrightCove:
 249         bc_url = BrightcoveIE._extract_brightcove_url(webpage)
 250         if bc_url is not None:
 251             self.to_screen('Brightcove video detected.')
 252             surl = smuggle_url(bc_url, {'Referer': url})
 253             return self.url_result(surl, 'Brightcove')
 254
 255         # Look for embedded (iframe) Vimeo player
 256         mobj = re.search(
 257             r'<iframe[^>]+?src="((?:https?:)?//player\.vimeo\.com/video/.+?)"', webpage)
 258         if mobj:
 259             player_url = unescapeHTML(mobj.group(1))
 260             surl = smuggle_url(player_url, {'Referer': url})
 261             return self.url_result(surl, 'Vimeo')
 262
 263         # Look for embedded (swf embed) Vimeo player
 264         mobj = re.search(
 265             r'<embed[^>]+?src="(https?://(?:www\.)?vimeo\.com/moogaloop\.swf.+?)"', webpage)
 266         if mobj:
 267             return self.url_result(mobj.group(1), 'Vimeo')
 268
 269         # Look for embedded YouTube player
 270         matches = re.findall(r'''(?x)
 271             (?:<iframe[^>]+?src=|embedSWF\(\s*)
 272             (["\'])(?P<url>(?:https?:)?//(?:www\.)?youtube\.com/
 273                 (?:embed|v)/.+?)
 274             \1''', webpage)
 275         if matches:
 276             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Youtube')
 277                      for tuppl in matches]
 278             return self.playlist_result(
 279                 urlrs, playlist_id=video_id, playlist_title=video_title)
 280
 281         # Look for embedded Dailymotion player
 282         matches = re.findall(
 283             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?dailymotion\.com/embed/video/.+?)\1', webpage)
 284         if matches:
 285             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Dailymotion')
 286                      for tuppl in matches]
 287             return self.playlist_result(
 288                 urlrs, playlist_id=video_id, playlist_title=video_title)
 289
 290         # Look for embedded Wistia player
 291         match = re.search(
 292             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage)
 293         if match:
 294             return {
 295                 '_type': 'url_transparent',
 296                 'url': unescapeHTML(match.group('url')),
 297                 'ie_key': 'Wistia',
 298                 'uploader': video_uploader,
 299                 'title': video_title,
 300                 'id': video_id,
 301             }
 302
 303         # Look for embedded blip.tv player
 304         mobj = re.search(r'<meta\s[^>]*https?://api\.blip\.tv/\w+/redirect/\w+/(\d+)', webpage)
 305         if mobj:
 306             return self.url_result('http://blip.tv/a/a-'+mobj.group(1), 'BlipTV')
 307         mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*(https?://(?:\w+\.)?blip\.tv/(?:play/|api\.swf#)[a-zA-Z0-9]+)', webpage)
 308         if mobj:
 309             return self.url_result(mobj.group(1), 'BlipTV')
 310
 311         # Look for Bandcamp pages with custom domain
 312         mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
 313         if mobj is not None:
 314             burl = unescapeHTML(mobj.group(1))
 315             # Don't set the extractor because it can be a track url or an album
 316             return self.url_result(burl)
 317
 318         # Look for embedded Vevo player
 319         mobj = re.search(
 320             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage)
 321         if mobj is not None:
 322             return self.url_result(mobj.group('url'))
 323
 324         # Look for Ooyala videos
 325         mobj = re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=([^"&]+)', webpage)
 326         if mobj is not None:
 327             return OoyalaIE._build_url_result(mobj.group(1))
 328
 329         # Look for Aparat videos
 330         mobj = re.search(r'<iframe src="(http://www\.aparat\.com/video/[^"]+)"', webpage)
 331         if mobj is not None:
 332             return self.url_result(mobj.group(1), 'Aparat')
 333
 334         # Look for MPORA videos
 335         mobj = re.search(r'<iframe .*?src="(http://mpora\.(?:com|de)/videos/[^"]+)"', webpage)
 336         if mobj is not None:
 337             return self.url_result(mobj.group(1), 'Mpora')
 338
 339         # Look for embedded Novamov player
 340         mobj = re.search(
 341             r'<iframe[^>]+?src=(["\'])(?P<url>http://(?:(?:embed|www)\.)?novamov\.com/embed\.php.+?)\1', webpage)
 342         if mobj is not None:
 343             return self.url_result(mobj.group('url'), 'Novamov')
 344
 345         # Look for embedded Facebook player
 346         mobj = re.search(
 347             r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
 348         if mobj is not None:
 349             return self.url_result(mobj.group('url'), 'Facebook')
 350
 351         # Look for embedded Huffington Post player
 352         mobj = re.search(
 353             r'<iframe[^>]+?src=(["\'])(?P<url>https?://embed\.live\.huffingtonpost\.com/.+?)\1', webpage)
 354         if mobj is not None:
 355             return self.url_result(mobj.group('url'), 'HuffPost')
 356
 357         # Start with something easy: JW Player in SWFObject
 358         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
 359         if mobj is None:
 360             # Look for gorilla-vid style embedding
 361             mobj = re.search(r'(?s)jw_plugins.*?file:\s*["\'](.*?)["\']', webpage)
 362         if mobj is None:
 363             # Broaden the search a little bit
 364             mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
 365         if mobj is None:
 366             # Broaden the search a little bit: JWPlayer JS loader
 367             mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http(?![^\'"]+\.[0-9]+[\'"])[^\'"]+)["\']', webpage)
 368         if mobj is None:
 369             # Try to find twitter cards info
 370             mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
 371         if mobj is None:
 372             # We look for Open Graph info:
 373             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
 374             m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
 375             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
 376             if m_video_type is not None:
 377                 mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
 378         if mobj is None:
 379             # HTML5 video
 380             mobj = re.search(r'<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage, flags=re.DOTALL)
 381         if mobj is None:
 382             raise ExtractorError('Unsupported URL: %s' % url)
 383
 384         # It's possible that one of the regexes
 385         # matched, but returned an empty group:
 386         if mobj.group(1) is None:
 387             raise ExtractorError('Did not find a valid video URL at %s' % url)
 388
 389         video_url = mobj.group(1)
 390         video_url = compat_urlparse.urljoin(url, video_url)
 391         video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
 392
 393         # Sometimes, jwplayer extraction will result in a YouTube URL
 394         if YoutubeIE.suitable(video_url):
 395             return self.url_result(video_url, 'Youtube')
 396
 397         # here's a fun little line of code for you:
 398         video_id = os.path.splitext(video_id)[0]
 399
 400         return {
 401             'id': video_id,
 402             'url': video_url,
 403             'uploader': video_uploader,
 404             'title': video_title,
 405         }