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