[ted] Simplify embed code (#2587)
[youtube-dl] / 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
29
30 class GenericIE(InfoExtractor):
31     IE_DESC = 'Generic downloader that works on some sites'
32     _VALID_URL = r'.*'
33     IE_NAME = 'generic'
34     _TESTS = [
35         {
36             'url': 'http://www.hodiho.fr/2013/02/regis-plante-sa-jeep.html',
37             'file': '13601338388002.mp4',
38             'md5': '6e15c93721d7ec9e9ca3fdbf07982cfd',
39             'info_dict': {
40                 'uploader': 'www.hodiho.fr',
41                 'title': 'R\u00e9gis plante sa Jeep',
42             }
43         },
44         # bandcamp page with custom domain
45         {
46             'add_ie': ['Bandcamp'],
47             'url': 'http://bronyrock.com/track/the-pony-mash',
48             'file': '3235767654.mp3',
49             'info_dict': {
50                 'title': 'The Pony Mash',
51                 'uploader': 'M_Pallante',
52             },
53             'skip': 'There is a limit of 200 free downloads / month for the test song',
54         },
55         # embedded brightcove video
56         # it also tests brightcove videos that need to set the 'Referer' in the
57         # http requests
58         {
59             'add_ie': ['Brightcove'],
60             'url': 'http://www.bfmtv.com/video/bfmbusiness/cours-bourse/cours-bourse-l-analyse-technique-154522/',
61             'info_dict': {
62                 'id': '2765128793001',
63                 'ext': 'mp4',
64                 'title': 'Le cours de bourse : l’analyse technique',
65                 'description': 'md5:7e9ad046e968cb2d1114004aba466fd9',
66                 'uploader': 'BFM BUSINESS',
67             },
68             'params': {
69                 'skip_download': True,
70             },
71         },
72         {
73             # https://github.com/rg3/youtube-dl/issues/2253
74             'url': 'http://bcove.me/i6nfkrc3',
75             'file': '3101154703001.mp4',
76             'md5': '0ba9446db037002366bab3b3eb30c88c',
77             'info_dict': {
78                 'title': 'Still no power',
79                 'uploader': 'thestar.com',
80                 '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.',
81             },
82             'add_ie': ['Brightcove'],
83         },
84         # Direct link to a video
85         {
86             'url': 'http://media.w3.org/2010/05/sintel/trailer.mp4',
87             'md5': '67d406c2bcb6af27fa886f31aa934bbe',
88             'info_dict': {
89                 'id': 'trailer',
90                 'ext': 'mp4',
91                 'title': 'trailer',
92                 'upload_date': '20100513',
93             }
94         },
95         # ooyala video
96         {
97             'url': 'http://www.rollingstone.com/music/videos/norwegian-dj-cashmere-cat-goes-spartan-on-with-me-premiere-20131219',
98             'md5': '5644c6ca5d5782c1d0d350dad9bd840c',
99             'info_dict': {
100                 'id': 'BwY2RxaTrTkslxOfcan0UCf0YqyvWysJ',
101                 'ext': 'mp4',
102                 'title': '2cc213299525360.mov',  # that's what we get
103             },
104         },
105         # google redirect
106         {
107             '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',
108             'info_dict': {
109                 'id': 'cmQHVoWB5FY',
110                 'ext': 'mp4',
111                 'upload_date': '20130224',
112                 'uploader_id': 'TheVerge',
113                 '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.',
114                 'uploader': 'The Verge',
115                 'title': 'First Firefox OS phones side-by-side',
116             },
117             'params': {
118                 'skip_download': False,
119             }
120         },
121         # embed.ly video
122         {
123             'url': 'http://www.tested.com/science/weird/460206-tested-grinding-coffee-2000-frames-second/',
124             'info_dict': {
125                 'id': '9ODmcdjQcHQ',
126                 'ext': 'mp4',
127                 'title': 'Tested: Grinding Coffee at 2000 Frames Per Second',
128                 'upload_date': '20140225',
129                 'description': 'md5:06a40fbf30b220468f1e0957c0f558ff',
130                 'uploader': 'Tested',
131                 'uploader_id': 'testedcom',
132             },
133             # No need to test YoutubeIE here
134             'params': {
135                 'skip_download': True,
136             },
137         },
138         # funnyordie embed
139         {
140             'url': 'http://www.theguardian.com/world/2014/mar/11/obama-zach-galifianakis-between-two-ferns',
141             'md5': '7cf780be104d40fea7bae52eed4a470e',
142             'info_dict': {
143                 'id': '18e820ec3f',
144                 'ext': 'mp4',
145                 'title': 'Between Two Ferns with Zach Galifianakis: President Barack Obama',
146                 'description': 'Episode 18: President Barack Obama sits down with Zach Galifianakis for his most memorable interview yet.',
147             },
148         },
149         # RUTV embed
150         {
151             'url': 'http://www.rg.ru/2014/03/15/reg-dfo/anklav-anons.html',
152             'info_dict': {
153                 'id': '776940',
154                 'ext': 'mp4',
155                 'title': 'Охотское море стало целиком российским',
156                 'description': 'md5:5ed62483b14663e2a95ebbe115eb8f43',
157             },
158             'params': {
159                 # m3u8 download
160                 'skip_download': True,
161             },
162         },
163         # Embedded TED video
164         {
165             'url': 'http://en.support.wordpress.com/videos/ted-talks/',
166             'md5': 'deeeabcc1085eb2ba205474e7235a3d5',
167             'info_dict': {
168                 'id': '981',
169                 'ext': 'mp4',
170                 'title': 'My web playroom',
171                 'uploader': 'Ze Frank',
172                 'description': 'md5:ddb2a40ecd6b6a147e400e535874947b',
173             }
174         }
175     ]
176
177     def report_download_webpage(self, video_id):
178         """Report webpage download."""
179         if not self._downloader.params.get('test', False):
180             self._downloader.report_warning('Falling back on generic information extractor.')
181         super(GenericIE, self).report_download_webpage(video_id)
182
183     def report_following_redirect(self, new_url):
184         """Report information extraction."""
185         self._downloader.to_screen('[redirect] Following redirect to %s' % new_url)
186
187     def _send_head(self, url):
188         """Check if it is a redirect, like url shorteners, in case return the new url."""
189
190         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
191             """
192             Subclass the HTTPRedirectHandler to make it use our
193             HEADRequest also on the redirected URL
194             """
195             def redirect_request(self, req, fp, code, msg, headers, newurl):
196                 if code in (301, 302, 303, 307):
197                     newurl = newurl.replace(' ', '%20')
198                     newheaders = dict((k,v) for k,v in req.headers.items()
199                                       if k.lower() not in ("content-length", "content-type"))
200                     try:
201                         # This function was deprecated in python 3.3 and removed in 3.4
202                         origin_req_host = req.get_origin_req_host()
203                     except AttributeError:
204                         origin_req_host = req.origin_req_host
205                     return HEADRequest(newurl,
206                                        headers=newheaders,
207                                        origin_req_host=origin_req_host,
208                                        unverifiable=True)
209                 else:
210                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
211
212         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
213             """
214             Fallback to GET if HEAD is not allowed (405 HTTP error)
215             """
216             def http_error_405(self, req, fp, code, msg, headers):
217                 fp.read()
218                 fp.close()
219
220                 newheaders = dict((k,v) for k,v in req.headers.items()
221                                   if k.lower() not in ("content-length", "content-type"))
222                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
223                                                  headers=newheaders,
224                                                  origin_req_host=req.get_origin_req_host(),
225                                                  unverifiable=True))
226
227         # Build our opener
228         opener = compat_urllib_request.OpenerDirector()
229         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
230                         HTTPMethodFallback, HEADRedirectHandler,
231                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
232             opener.add_handler(handler())
233
234         response = opener.open(HEADRequest(url))
235         if response is None:
236             raise ExtractorError('Invalid URL protocol')
237         return response
238
239     def _extract_rss(self, url, video_id, doc):
240         playlist_title = doc.find('./channel/title').text
241         playlist_desc_el = doc.find('./channel/description')
242         playlist_desc = None if playlist_desc_el is None else playlist_desc_el.text
243
244         entries = [{
245             '_type': 'url',
246             'url': e.find('link').text,
247             'title': e.find('title').text,
248         } for e in doc.findall('./channel/item')]
249
250         return {
251             '_type': 'playlist',
252             'id': url,
253             'title': playlist_title,
254             'description': playlist_desc,
255             'entries': entries,
256         }
257
258     def _real_extract(self, url):
259         parsed_url = compat_urlparse.urlparse(url)
260         if not parsed_url.scheme:
261             default_search = self._downloader.params.get('default_search')
262             if default_search is None:
263                 default_search = 'auto'
264
265             if default_search == 'auto':
266                 if '/' in url:
267                     self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
268                     return self.url_result('http://' + url)
269                 else:
270                     return self.url_result('ytsearch:' + url)
271             else:
272                 assert ':' in default_search
273                 return self.url_result(default_search + url)
274         video_id = os.path.splitext(url.rstrip('/').split('/')[-1])[0]
275
276         self.to_screen('%s: Requesting header' % video_id)
277
278         try:
279             response = self._send_head(url)
280
281             # Check for redirect
282             new_url = response.geturl()
283             if url != new_url:
284                 self.report_following_redirect(new_url)
285                 return self.url_result(new_url)
286
287             # Check for direct link to a video
288             content_type = response.headers.get('Content-Type', '')
289             m = re.match(r'^(?P<type>audio|video|application(?=/ogg$))/(?P<format_id>.+)$', content_type)
290             if m:
291                 upload_date = response.headers.get('Last-Modified')
292                 if upload_date:
293                     upload_date = unified_strdate(upload_date)
294                 return {
295                     'id': video_id,
296                     'title': os.path.splitext(url_basename(url))[0],
297                     'formats': [{
298                         'format_id': m.group('format_id'),
299                         'url': url,
300                         'vcodec': 'none' if m.group('type') == 'audio' else None
301                     }],
302                     'upload_date': upload_date,
303                 }
304
305         except compat_urllib_error.HTTPError:
306             # This may be a stupid server that doesn't like HEAD, our UA, or so
307             pass
308
309         try:
310             webpage = self._download_webpage(url, video_id)
311         except ValueError:
312             # since this is the last-resort InfoExtractor, if
313             # this error is thrown, it'll be thrown here
314             raise ExtractorError('Failed to download URL: %s' % url)
315
316         self.report_extraction(video_id)
317
318         # Is it an RSS feed?
319         try:
320             doc = parse_xml(webpage)
321             if doc.tag == 'rss':
322                 return self._extract_rss(url, video_id, doc)
323         except compat_xml_parse_error:
324             pass
325
326         # it's tempting to parse this further, but you would
327         # have to take into account all the variations like
328         #   Video Title - Site Name
329         #   Site Name | Video Title
330         #   Video Title - Tagline | Site Name
331         # and so on and so forth; it's just not practical
332         video_title = self._html_search_regex(
333             r'(?s)<title>(.*?)</title>', webpage, 'video title',
334             default='video')
335
336         # video uploader is domain name
337         video_uploader = self._search_regex(
338             r'^(?:https?://)?([^/]*)/.*', url, 'video uploader')
339
340         # Look for BrightCove:
341         bc_urls = BrightcoveIE._extract_brightcove_urls(webpage)
342         if bc_urls:
343             self.to_screen('Brightcove video detected.')
344             entries = [{
345                 '_type': 'url',
346                 'url': smuggle_url(bc_url, {'Referer': url}),
347                 'ie_key': 'Brightcove'
348             } for bc_url in bc_urls]
349
350             return {
351                 '_type': 'playlist',
352                 'title': video_title,
353                 'id': video_id,
354                 'entries': entries,
355             }
356
357         # Look for embedded (iframe) Vimeo player
358         mobj = re.search(
359             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//player\.vimeo\.com/video/.+?)\1', webpage)
360         if mobj:
361             player_url = unescapeHTML(mobj.group('url'))
362             surl = smuggle_url(player_url, {'Referer': url})
363             return self.url_result(surl, 'Vimeo')
364
365         # Look for embedded (swf embed) Vimeo player
366         mobj = re.search(
367             r'<embed[^>]+?src="(https?://(?:www\.)?vimeo\.com/moogaloop\.swf.+?)"', webpage)
368         if mobj:
369             return self.url_result(mobj.group(1), 'Vimeo')
370
371         # Look for embedded YouTube player
372         matches = re.findall(r'''(?x)
373             (?:<iframe[^>]+?src=|embedSWF\(\s*)
374             (["\'])(?P<url>(?:https?:)?//(?:www\.)?youtube\.com/
375                 (?:embed|v)/.+?)
376             \1''', webpage)
377         if matches:
378             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Youtube')
379                      for tuppl in matches]
380             return self.playlist_result(
381                 urlrs, playlist_id=video_id, playlist_title=video_title)
382
383         # Look for embedded Dailymotion player
384         matches = re.findall(
385             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?dailymotion\.com/embed/video/.+?)\1', webpage)
386         if matches:
387             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Dailymotion')
388                      for tuppl in matches]
389             return self.playlist_result(
390                 urlrs, playlist_id=video_id, playlist_title=video_title)
391
392         # Look for embedded Wistia player
393         match = re.search(
394             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage)
395         if match:
396             return {
397                 '_type': 'url_transparent',
398                 'url': unescapeHTML(match.group('url')),
399                 'ie_key': 'Wistia',
400                 'uploader': video_uploader,
401                 'title': video_title,
402                 'id': video_id,
403             }
404
405         # Look for embedded blip.tv player
406         mobj = re.search(r'<meta\s[^>]*https?://api\.blip\.tv/\w+/redirect/\w+/(\d+)', webpage)
407         if mobj:
408             return self.url_result('http://blip.tv/a/a-'+mobj.group(1), 'BlipTV')
409         mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*(https?://(?:\w+\.)?blip\.tv/(?:play/|api\.swf#)[a-zA-Z0-9]+)', webpage)
410         if mobj:
411             return self.url_result(mobj.group(1), 'BlipTV')
412
413         # Look for Bandcamp pages with custom domain
414         mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
415         if mobj is not None:
416             burl = unescapeHTML(mobj.group(1))
417             # Don't set the extractor because it can be a track url or an album
418             return self.url_result(burl)
419
420         # Look for embedded Vevo player
421         mobj = re.search(
422             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage)
423         if mobj is not None:
424             return self.url_result(mobj.group('url'))
425
426         # Look for Ooyala videos
427         mobj = re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=([^"&]+)', webpage)
428         if mobj is not None:
429             return OoyalaIE._build_url_result(mobj.group(1))
430
431         # Look for Aparat videos
432         mobj = re.search(r'<iframe src="(http://www\.aparat\.com/video/[^"]+)"', webpage)
433         if mobj is not None:
434             return self.url_result(mobj.group(1), 'Aparat')
435
436         # Look for MPORA videos
437         mobj = re.search(r'<iframe .*?src="(http://mpora\.(?:com|de)/videos/[^"]+)"', webpage)
438         if mobj is not None:
439             return self.url_result(mobj.group(1), 'Mpora')
440
441         # Look for embedded NovaMov player
442         mobj = re.search(
443             r'<iframe[^>]+?src=(["\'])(?P<url>http://(?:(?:embed|www)\.)?novamov\.com/embed\.php.+?)\1', webpage)
444         if mobj is not None:
445             return self.url_result(mobj.group('url'), 'NovaMov')
446
447         # Look for embedded NowVideo player
448         mobj = re.search(
449             r'<iframe[^>]+?src=(["\'])(?P<url>http://(?:(?:embed|www)\.)?nowvideo\.(?:ch|sx|eu)/embed\.php.+?)\1', webpage)
450         if mobj is not None:
451             return self.url_result(mobj.group('url'), 'NowVideo')
452
453         # Look for embedded Facebook player
454         mobj = re.search(
455             r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
456         if mobj is not None:
457             return self.url_result(mobj.group('url'), 'Facebook')
458
459         # Look for embedded VK player
460         mobj = re.search(r'<iframe[^>]+?src=(["\'])(?P<url>https?://vk\.com/video_ext\.php.+?)\1', webpage)
461         if mobj is not None:
462             return self.url_result(mobj.group('url'), 'VK')
463
464         # Look for embedded Huffington Post player
465         mobj = re.search(
466             r'<iframe[^>]+?src=(["\'])(?P<url>https?://embed\.live\.huffingtonpost\.com/.+?)\1', webpage)
467         if mobj is not None:
468             return self.url_result(mobj.group('url'), 'HuffPost')
469
470         # Look for embed.ly
471         mobj = re.search(r'class=["\']embedly-card["\'][^>]href=["\'](?P<url>[^"\']+)', webpage)
472         if mobj is not None:
473             return self.url_result(mobj.group('url'))
474         mobj = re.search(r'class=["\']embedly-embed["\'][^>]src=["\'][^"\']*url=(?P<url>[^&]+)', webpage)
475         if mobj is not None:
476             return self.url_result(compat_urllib_parse.unquote(mobj.group('url')))
477
478         # Look for funnyordie embed
479         matches = re.findall(r'<iframe[^>]+?src="(https?://(?:www\.)?funnyordie\.com/embed/[^"]+)"', webpage)
480         if matches:
481             urlrs = [self.url_result(unescapeHTML(eurl), 'FunnyOrDie')
482                      for eurl in matches]
483             return self.playlist_result(
484                 urlrs, playlist_id=video_id, playlist_title=video_title)
485
486         # Look for embedded RUTV player
487         rutv_url = RUTVIE._extract_url(webpage)
488         if rutv_url:
489             return self.url_result(rutv_url, 'RUTV')
490
491         # Start with something easy: JW Player in SWFObject
492         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
493         if mobj is None:
494             # Look for gorilla-vid style embedding
495             mobj = re.search(r'(?s)(?:jw_plugins|JWPlayerOptions).*?file\s*:\s*["\'](.*?)["\']', webpage)
496         if mobj is None:
497             # Broaden the search a little bit
498             mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
499         if mobj is None:
500             # Broaden the search a little bit: JWPlayer JS loader
501             mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http(?![^\'"]+\.[0-9]+[\'"])[^\'"]+)["\']', webpage)
502
503         # Look for embedded TED player
504         mobj = re.search(
505             r'<iframe[^>]+?src=(["\'])(?P<url>http://embed\.ted\.com/.+?)\1', webpage)
506         if mobj is not None:
507             return self.url_result(mobj.group('url'), 'TED')
508
509         if mobj is None:
510             # Try to find twitter cards info
511             mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
512         if mobj is None:
513             # We look for Open Graph info:
514             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
515             m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
516             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
517             if m_video_type is not None:
518                 mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
519         if mobj is None:
520             # HTML5 video
521             mobj = re.search(r'<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage, flags=re.DOTALL)
522         if mobj is None:
523             mobj = re.search(
524                 r'(?i)<meta\s+(?=(?:[a-z-]+="[^"]+"\s+)*http-equiv="refresh")'
525                 r'(?:[a-z-]+="[^"]+"\s+)*?content="[0-9]{,2};url=\'([^\']+)\'"',
526                 webpage)
527             if mobj:
528                 new_url = mobj.group(1)
529                 self.report_following_redirect(new_url)
530                 return {
531                     '_type': 'url',
532                     'url': new_url,
533                 }
534         if mobj is None:
535             raise ExtractorError('Unsupported URL: %s' % url)
536
537         # It's possible that one of the regexes
538         # matched, but returned an empty group:
539         if mobj.group(1) is None:
540             raise ExtractorError('Did not find a valid video URL at %s' % url)
541
542         video_url = mobj.group(1)
543         video_url = compat_urlparse.urljoin(url, video_url)
544         video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
545
546         # Sometimes, jwplayer extraction will result in a YouTube URL
547         if YoutubeIE.suitable(video_url):
548             return self.url_result(video_url, 'Youtube')
549
550         # here's a fun little line of code for you:
551         video_id = os.path.splitext(video_id)[0]
552
553         return {
554             'id': video_id,
555             'url': video_url,
556             'uploader': video_uploader,
557             'title': video_title,
558         }