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

   1 # encoding: utf-8
   2
   3 import os
   4 import re
   5
   6 from .common import InfoExtractor
   7 from .youtube import YoutubeIE
   8 from ..utils import (
   9     compat_urllib_error,
  10     compat_urllib_parse,
  11     compat_urllib_request,
  12     compat_urlparse,
  13
  14     ExtractorError,
  15     HEADRequest,
  16     smuggle_url,
  17     unescapeHTML,
  18     unified_strdate,
  19     url_basename,
  20 )
  21 from .brightcove import BrightcoveIE
  22 from .ooyala import OoyalaIE
  23
  24
  25 class GenericIE(InfoExtractor):
  26     IE_DESC = u'Generic downloader that works on some sites'
  27     _VALID_URL = r'.*'
  28     IE_NAME = u'generic'
  29     _TESTS = [
  30         {
  31             u'url': u'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
  32             u'file': u'13601338388002.mp4',
  33             u'md5': u'6e15c93721d7ec9e9ca3fdbf07982cfd',
  34             u'info_dict': {
  35                 u"uploader": u"www.hodiho.fr",
  36                 u"title": u"R\u00e9gis plante sa Jeep"
  37             }
  38         },
  39         # embedded vimeo video
  40         {
  41             u'add_ie': ['Vimeo'],
  42             u'url': u'http://skillsmatter.com/podcast/home/move-semanticsperfect-forwarding-and-rvalue-references',
  43             u'file': u'22444065.mp4',
  44             u'md5': u'2903896e23df39722c33f015af0666e2',
  45             u'info_dict': {
  46                 u'title': u'ACCU 2011: Move Semantics,Perfect Forwarding, and Rvalue references- Scott Meyers- 13/04/2011',
  47                 u"uploader_id": u"skillsmatter",
  48                 u"uploader": u"Skills Matter",
  49             }
  50         },
  51         # bandcamp page with custom domain
  52         {
  53             u'add_ie': ['Bandcamp'],
  54             u'url': u'http://bronyrock.com/track/the-pony-mash',
  55             u'file': u'3235767654.mp3',
  56             u'info_dict': {
  57                 u'title': u'The Pony Mash',
  58                 u'uploader': u'M_Pallante',
  59             },
  60             u'skip': u'There is a limit of 200 free downloads / month for the test song',
  61         },
  62         # embedded brightcove video
  63         # it also tests brightcove videos that need to set the 'Referer' in the
  64         # http requests
  65         {
  66             u'add_ie': ['Brightcove'],
  67             u'url': u'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/',
  68             u'info_dict': {
  69                 u'id': u'2765128793001',
  70                 u'ext': u'mp4',
  71                 u'title': u'Le cours de bourse : l’analyse technique',
  72                 u'description': u'md5:7e9ad046e968cb2d1114004aba466fd9',
  73                 u'uploader': u'BFM BUSINESS',
  74             },
  75             u'params': {
  76                 u'skip_download': True,
  77             },
  78         },
  79         # Direct link to a video
  80         {
  81             u'url': u'http://media.w3.org/2010/05/sintel/trailer.mp4',
  82             u'file': u'trailer.mp4',
  83             u'md5': u'67d406c2bcb6af27fa886f31aa934bbe',
  84             u'info_dict': {
  85                 u'id': u'trailer',
  86                 u'title': u'trailer',
  87                 u'upload_date': u'20100513',
  88             }
  89         },
  90         # ooyala video
  91         {
  92             u'url': u'http://www.rollingstone.com/music/videos/norwegian-dj-cashmere-cat-goes-spartan-on-with-me-premiere-20131219',
  93             u'md5': u'5644c6ca5d5782c1d0d350dad9bd840c',
  94             u'info_dict': {
  95                 u'id': u'BwY2RxaTrTkslxOfcan0UCf0YqyvWysJ',
  96                 u'ext': u'mp4',
  97                 u'title': u'2cc213299525360.mov', #that's what we get
  98             },
  99         },
 100     ]
 101
 102     def report_download_webpage(self, video_id):
 103         """Report webpage download."""
 104         if not self._downloader.params.get('test', False):
 105             self._downloader.report_warning(u'Falling back on generic information extractor.')
 106         super(GenericIE, self).report_download_webpage(video_id)
 107
 108     def report_following_redirect(self, new_url):
 109         """Report information extraction."""
 110         self._downloader.to_screen(u'[redirect] Following redirect to %s' % new_url)
 111
 112     def _send_head(self, url):
 113         """Check if it is a redirect, like url shorteners, in case return the new url."""
 114
 115         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
 116             """
 117             Subclass the HTTPRedirectHandler to make it use our
 118             HEADRequest also on the redirected URL
 119             """
 120             def redirect_request(self, req, fp, code, msg, headers, newurl):
 121                 if code in (301, 302, 303, 307):
 122                     newurl = newurl.replace(' ', '%20')
 123                     newheaders = dict((k,v) for k,v in req.headers.items()
 124                                       if k.lower() not in ("content-length", "content-type"))
 125                     return HEADRequest(newurl,
 126                                        headers=newheaders,
 127                                        origin_req_host=req.get_origin_req_host(),
 128                                        unverifiable=True)
 129                 else:
 130                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
 131
 132         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
 133             """
 134             Fallback to GET if HEAD is not allowed (405 HTTP error)
 135             """
 136             def http_error_405(self, req, fp, code, msg, headers):
 137                 fp.read()
 138                 fp.close()
 139
 140                 newheaders = dict((k,v) for k,v in req.headers.items()
 141                                   if k.lower() not in ("content-length", "content-type"))
 142                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
 143                                                  headers=newheaders,
 144                                                  origin_req_host=req.get_origin_req_host(),
 145                                                  unverifiable=True))
 146
 147         # Build our opener
 148         opener = compat_urllib_request.OpenerDirector()
 149         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
 150                         HTTPMethodFallback, HEADRedirectHandler,
 151                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
 152             opener.add_handler(handler())
 153
 154         response = opener.open(HEADRequest(url))
 155         if response is None:
 156             raise ExtractorError(u'Invalid URL protocol')
 157         return response
 158
 159     def _real_extract(self, url):
 160         parsed_url = compat_urlparse.urlparse(url)
 161         if not parsed_url.scheme:
 162             self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
 163             return self.url_result('http://' + url)
 164         video_id = os.path.splitext(url.split('/')[-1])[0]
 165
 166         self.to_screen(u'%s: Requesting header' % video_id)
 167
 168         try:
 169             response = self._send_head(url)
 170
 171             # Check for redirect
 172             new_url = response.geturl()
 173             if url != new_url:
 174                 self.report_following_redirect(new_url)
 175                 return self.url_result(new_url)
 176
 177             # Check for direct link to a video
 178             content_type = response.headers.get('Content-Type', '')
 179             m = re.match(r'^(?P<type>audio|video|application(?=/ogg$))/(?P<format_id>.+)$', content_type)
 180             if m:
 181                 upload_date = response.headers.get('Last-Modified')
 182                 if upload_date:
 183                     upload_date = unified_strdate(upload_date)
 184                 return {
 185                     'id': video_id,
 186                     'title': os.path.splitext(url_basename(url))[0],
 187                     'formats': [{
 188                         'format_id': m.group('format_id'),
 189                         'url': url,
 190                         'vcodec': u'none' if m.group('type') == 'audio' else None
 191                     }],
 192                     'upload_date': upload_date,
 193                 }
 194
 195         except compat_urllib_error.HTTPError:
 196             # This may be a stupid server that doesn't like HEAD, our UA, or so
 197             pass
 198
 199         try:
 200             webpage = self._download_webpage(url, video_id)
 201         except ValueError:
 202             # since this is the last-resort InfoExtractor, if
 203             # this error is thrown, it'll be thrown here
 204             raise ExtractorError(u'Failed to download URL: %s' % url)
 205
 206         self.report_extraction(video_id)
 207
 208         # it's tempting to parse this further, but you would
 209         # have to take into account all the variations like
 210         #   Video Title - Site Name
 211         #   Site Name | Video Title
 212         #   Video Title - Tagline | Site Name
 213         # and so on and so forth; it's just not practical
 214         video_title = self._html_search_regex(
 215             r'(?s)<title>(.*?)</title>', webpage, u'video title',
 216             default=u'video')
 217
 218         # video uploader is domain name
 219         video_uploader = self._search_regex(
 220             r'^(?:https?://)?([^/]*)/.*', url, u'video uploader')
 221
 222         # Look for BrightCove:
 223         bc_url = BrightcoveIE._extract_brightcove_url(webpage)
 224         if bc_url is not None:
 225             self.to_screen(u'Brightcove video detected.')
 226             return self.url_result(bc_url, 'Brightcove')
 227
 228         # Look for embedded (iframe) Vimeo player
 229         mobj = re.search(
 230             r'<iframe[^>]+?src="(https?://player.vimeo.com/video/.+?)"', webpage)
 231         if mobj:
 232             player_url = unescapeHTML(mobj.group(1))
 233             surl = smuggle_url(player_url, {'Referer': url})
 234             return self.url_result(surl, 'Vimeo')
 235
 236         # Look for embedded (swf embed) Vimeo player
 237         mobj = re.search(
 238             r'<embed[^>]+?src="(https?://(?:www\.)?vimeo.com/moogaloop.swf.+?)"', webpage)
 239         if mobj:
 240             return self.url_result(mobj.group(1), 'Vimeo')
 241
 242         # Look for embedded YouTube player
 243         matches = re.findall(r'''(?x)
 244             (?:<iframe[^>]+?src=|embedSWF\(\s*)
 245             (["\'])(?P<url>(?:https?:)?//(?:www\.)?youtube\.com/
 246                 (?:embed|v)/.+?)
 247             \1''', webpage)
 248         if matches:
 249             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Youtube')
 250                      for tuppl in matches]
 251             return self.playlist_result(
 252                 urlrs, playlist_id=video_id, playlist_title=video_title)
 253
 254         # Look for embedded Dailymotion player
 255         matches = re.findall(
 256             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?dailymotion\.com/embed/video/.+?)\1', webpage)
 257         if matches:
 258             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Dailymotion')
 259                      for tuppl in matches]
 260             return self.playlist_result(
 261                 urlrs, playlist_id=video_id, playlist_title=video_title)
 262
 263         # Look for embedded Wistia player
 264         match = re.search(
 265             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage)
 266         if match:
 267             return {
 268                 '_type': 'url_transparent',
 269                 'url': unescapeHTML(match.group('url')),
 270                 'ie_key': 'Wistia',
 271                 'uploader': video_uploader,
 272                 'title': video_title,
 273                 'id': video_id,
 274             }
 275
 276         # Look for embedded blip.tv player
 277         mobj = re.search(r'<meta\s[^>]*https?://api\.blip\.tv/\w+/redirect/\w+/(\d+)', webpage)
 278         if mobj:
 279             return self.url_result('http://blip.tv/a/a-'+mobj.group(1), 'BlipTV')
 280         mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*(https?://(?:\w+\.)?blip\.tv/(?:play/|api\.swf#)[a-zA-Z0-9]+)', webpage)
 281         if mobj:
 282             return self.url_result(mobj.group(1), 'BlipTV')
 283
 284         # Look for Bandcamp pages with custom domain
 285         mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
 286         if mobj is not None:
 287             burl = unescapeHTML(mobj.group(1))
 288             # Don't set the extractor because it can be a track url or an album
 289             return self.url_result(burl)
 290
 291         # Look for embedded Vevo player
 292         mobj = re.search(
 293             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage)
 294         if mobj is not None:
 295             return self.url_result(mobj.group('url'))
 296
 297         # Look for Ooyala videos
 298         mobj = re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=([^"&]+)', webpage)
 299         if mobj is not None:
 300             return OoyalaIE._build_url_result(mobj.group(1))
 301
 302         # Look for Aparat videos
 303         mobj = re.search(r'<iframe src="(http://www.aparat.com/video/[^"]+)"', webpage)
 304         if mobj is not None:
 305             return self.url_result(mobj.group(1), 'Aparat')
 306
 307         # Start with something easy: JW Player in SWFObject
 308         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
 309         if mobj is None:
 310             # Look for gorilla-vid style embedding
 311             mobj = re.search(r'(?s)jw_plugins.*?file:\s*["\'](.*?)["\']', webpage)
 312         if mobj is None:
 313             # Broaden the search a little bit
 314             mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
 315         if mobj is None:
 316             # Broaden the search a little bit: JWPlayer JS loader
 317             mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http[^\'"]*)', webpage)
 318         if mobj is None:
 319             # Try to find twitter cards info
 320             mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
 321         if mobj is None:
 322             # We look for Open Graph info:
 323             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
 324             m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
 325             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
 326             if m_video_type is not None:
 327                 mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
 328         if mobj is None:
 329             # HTML5 video
 330             mobj = re.search(r'<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage, flags=re.DOTALL)
 331         if mobj is None:
 332             raise ExtractorError(u'Unsupported URL: %s' % url)
 333
 334         # It's possible that one of the regexes
 335         # matched, but returned an empty group:
 336         if mobj.group(1) is None:
 337             raise ExtractorError(u'Did not find a valid video URL at %s' % url)
 338
 339         video_url = mobj.group(1)
 340         video_url = compat_urlparse.urljoin(url, video_url)
 341         video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
 342
 343         # Sometimes, jwplayer extraction will result in a YouTube URL
 344         if YoutubeIE.suitable(video_url):
 345             return self.url_result(video_url, 'Youtube')
 346
 347         # here's a fun little line of code for you:
 348         video_id = os.path.splitext(video_id)[0]
 349
 350         return {
 351             'id': video_id,
 352             'url': video_url,
 353             'uploader': video_uploader,
 354             'title': video_title,
 355         }