[ooyala] Modernize
[ytdl] / 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         # second style of embedded ooyala videos
106         {
107             'url': 'http://www.smh.com.au/tv/business/show/financial-review-sunday/behind-the-scenes-financial-review-sunday--4350201.html',
108             'info_dict': {
109                 'id': '13djJjYjptA1XpPx8r9kuzPyj3UZH0Uk',
110                 'ext': 'mp4',
111                 'title': 'Behind-the-scenes: Financial Review Sunday ',
112                 'description': 'Step inside Channel Nine studios for an exclusive tour of its upcoming financial business show.',
113             },
114             'params': {
115                 # m3u8 download
116                 'skip_download': True,
117             },
118         },
119         # google redirect
120         {
121             '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',
122             'info_dict': {
123                 'id': 'cmQHVoWB5FY',
124                 'ext': 'mp4',
125                 'upload_date': '20130224',
126                 'uploader_id': 'TheVerge',
127                 '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.',
128                 'uploader': 'The Verge',
129                 'title': 'First Firefox OS phones side-by-side',
130             },
131             'params': {
132                 'skip_download': False,
133             }
134         },
135         # embed.ly video
136         {
137             'url': 'http://www.tested.com/science/weird/460206-tested-grinding-coffee-2000-frames-second/',
138             'info_dict': {
139                 'id': '9ODmcdjQcHQ',
140                 'ext': 'mp4',
141                 'title': 'Tested: Grinding Coffee at 2000 Frames Per Second',
142                 'upload_date': '20140225',
143                 'description': 'md5:06a40fbf30b220468f1e0957c0f558ff',
144                 'uploader': 'Tested',
145                 'uploader_id': 'testedcom',
146             },
147             # No need to test YoutubeIE here
148             'params': {
149                 'skip_download': True,
150             },
151         },
152         # funnyordie embed
153         {
154             'url': 'http://www.theguardian.com/world/2014/mar/11/obama-zach-galifianakis-between-two-ferns',
155             'md5': '7cf780be104d40fea7bae52eed4a470e',
156             'info_dict': {
157                 'id': '18e820ec3f',
158                 'ext': 'mp4',
159                 'title': 'Between Two Ferns with Zach Galifianakis: President Barack Obama',
160                 'description': 'Episode 18: President Barack Obama sits down with Zach Galifianakis for his most memorable interview yet.',
161             },
162         },
163         # RUTV embed
164         {
165             'url': 'http://www.rg.ru/2014/03/15/reg-dfo/anklav-anons.html',
166             'info_dict': {
167                 'id': '776940',
168                 'ext': 'mp4',
169                 'title': 'Охотское море стало целиком российским',
170                 'description': 'md5:5ed62483b14663e2a95ebbe115eb8f43',
171             },
172             'params': {
173                 # m3u8 download
174                 'skip_download': True,
175             },
176         },
177         # Embedded TED video
178         {
179             'url': 'http://en.support.wordpress.com/videos/ted-talks/',
180             'md5': 'deeeabcc1085eb2ba205474e7235a3d5',
181             'info_dict': {
182                 'id': '981',
183                 'ext': 'mp4',
184                 'title': 'My web playroom',
185                 'uploader': 'Ze Frank',
186                 'description': 'md5:ddb2a40ecd6b6a147e400e535874947b',
187             }
188         }
189     ]
190
191     def report_download_webpage(self, video_id):
192         """Report webpage download."""
193         if not self._downloader.params.get('test', False):
194             self._downloader.report_warning('Falling back on generic information extractor.')
195         super(GenericIE, self).report_download_webpage(video_id)
196
197     def report_following_redirect(self, new_url):
198         """Report information extraction."""
199         self._downloader.to_screen('[redirect] Following redirect to %s' % new_url)
200
201     def _send_head(self, url):
202         """Check if it is a redirect, like url shorteners, in case return the new url."""
203
204         class HEADRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
205             """
206             Subclass the HTTPRedirectHandler to make it use our
207             HEADRequest also on the redirected URL
208             """
209             def redirect_request(self, req, fp, code, msg, headers, newurl):
210                 if code in (301, 302, 303, 307):
211                     newurl = newurl.replace(' ', '%20')
212                     newheaders = dict((k,v) for k,v in req.headers.items()
213                                       if k.lower() not in ("content-length", "content-type"))
214                     try:
215                         # This function was deprecated in python 3.3 and removed in 3.4
216                         origin_req_host = req.get_origin_req_host()
217                     except AttributeError:
218                         origin_req_host = req.origin_req_host
219                     return HEADRequest(newurl,
220                                        headers=newheaders,
221                                        origin_req_host=origin_req_host,
222                                        unverifiable=True)
223                 else:
224                     raise compat_urllib_error.HTTPError(req.get_full_url(), code, msg, headers, fp)
225
226         class HTTPMethodFallback(compat_urllib_request.BaseHandler):
227             """
228             Fallback to GET if HEAD is not allowed (405 HTTP error)
229             """
230             def http_error_405(self, req, fp, code, msg, headers):
231                 fp.read()
232                 fp.close()
233
234                 newheaders = dict((k,v) for k,v in req.headers.items()
235                                   if k.lower() not in ("content-length", "content-type"))
236                 return self.parent.open(compat_urllib_request.Request(req.get_full_url(),
237                                                  headers=newheaders,
238                                                  origin_req_host=req.get_origin_req_host(),
239                                                  unverifiable=True))
240
241         # Build our opener
242         opener = compat_urllib_request.OpenerDirector()
243         for handler in [compat_urllib_request.HTTPHandler, compat_urllib_request.HTTPDefaultErrorHandler,
244                         HTTPMethodFallback, HEADRedirectHandler,
245                         compat_urllib_request.HTTPErrorProcessor, compat_urllib_request.HTTPSHandler]:
246             opener.add_handler(handler())
247
248         response = opener.open(HEADRequest(url))
249         if response is None:
250             raise ExtractorError('Invalid URL protocol')
251         return response
252
253     def _extract_rss(self, url, video_id, doc):
254         playlist_title = doc.find('./channel/title').text
255         playlist_desc_el = doc.find('./channel/description')
256         playlist_desc = None if playlist_desc_el is None else playlist_desc_el.text
257
258         entries = [{
259             '_type': 'url',
260             'url': e.find('link').text,
261             'title': e.find('title').text,
262         } for e in doc.findall('./channel/item')]
263
264         return {
265             '_type': 'playlist',
266             'id': url,
267             'title': playlist_title,
268             'description': playlist_desc,
269             'entries': entries,
270         }
271
272     def _real_extract(self, url):
273         parsed_url = compat_urlparse.urlparse(url)
274         if not parsed_url.scheme:
275             default_search = self._downloader.params.get('default_search')
276             if default_search is None:
277                 default_search = 'auto'
278
279             if default_search == 'auto':
280                 if '/' in url:
281                     self._downloader.report_warning('The url doesn\'t specify the protocol, trying with http')
282                     return self.url_result('http://' + url)
283                 else:
284                     return self.url_result('ytsearch:' + url)
285             else:
286                 assert ':' in default_search
287                 return self.url_result(default_search + url)
288         video_id = os.path.splitext(url.rstrip('/').split('/')[-1])[0]
289
290         self.to_screen('%s: Requesting header' % video_id)
291
292         try:
293             response = self._send_head(url)
294
295             # Check for redirect
296             new_url = response.geturl()
297             if url != new_url:
298                 self.report_following_redirect(new_url)
299                 return self.url_result(new_url)
300
301             # Check for direct link to a video
302             content_type = response.headers.get('Content-Type', '')
303             m = re.match(r'^(?P<type>audio|video|application(?=/ogg$))/(?P<format_id>.+)$', content_type)
304             if m:
305                 upload_date = response.headers.get('Last-Modified')
306                 if upload_date:
307                     upload_date = unified_strdate(upload_date)
308                 return {
309                     'id': video_id,
310                     'title': os.path.splitext(url_basename(url))[0],
311                     'formats': [{
312                         'format_id': m.group('format_id'),
313                         'url': url,
314                         'vcodec': 'none' if m.group('type') == 'audio' else None
315                     }],
316                     'upload_date': upload_date,
317                 }
318
319         except compat_urllib_error.HTTPError:
320             # This may be a stupid server that doesn't like HEAD, our UA, or so
321             pass
322
323         try:
324             webpage = self._download_webpage(url, video_id)
325         except ValueError:
326             # since this is the last-resort InfoExtractor, if
327             # this error is thrown, it'll be thrown here
328             raise ExtractorError('Failed to download URL: %s' % url)
329
330         self.report_extraction(video_id)
331
332         # Is it an RSS feed?
333         try:
334             doc = parse_xml(webpage)
335             if doc.tag == 'rss':
336                 return self._extract_rss(url, video_id, doc)
337         except compat_xml_parse_error:
338             pass
339
340         # it's tempting to parse this further, but you would
341         # have to take into account all the variations like
342         #   Video Title - Site Name
343         #   Site Name | Video Title
344         #   Video Title - Tagline | Site Name
345         # and so on and so forth; it's just not practical
346         video_title = self._html_search_regex(
347             r'(?s)<title>(.*?)</title>', webpage, 'video title',
348             default='video')
349
350         # video uploader is domain name
351         video_uploader = self._search_regex(
352             r'^(?:https?://)?([^/]*)/.*', url, 'video uploader')
353
354         # Look for BrightCove:
355         bc_urls = BrightcoveIE._extract_brightcove_urls(webpage)
356         if bc_urls:
357             self.to_screen('Brightcove video detected.')
358             entries = [{
359                 '_type': 'url',
360                 'url': smuggle_url(bc_url, {'Referer': url}),
361                 'ie_key': 'Brightcove'
362             } for bc_url in bc_urls]
363
364             return {
365                 '_type': 'playlist',
366                 'title': video_title,
367                 'id': video_id,
368                 'entries': entries,
369             }
370
371         # Look for embedded (iframe) Vimeo player
372         mobj = re.search(
373             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//player\.vimeo\.com/video/.+?)\1', webpage)
374         if mobj:
375             player_url = unescapeHTML(mobj.group('url'))
376             surl = smuggle_url(player_url, {'Referer': url})
377             return self.url_result(surl, 'Vimeo')
378
379         # Look for embedded (swf embed) Vimeo player
380         mobj = re.search(
381             r'<embed[^>]+?src="(https?://(?:www\.)?vimeo\.com/moogaloop\.swf.+?)"', webpage)
382         if mobj:
383             return self.url_result(mobj.group(1), 'Vimeo')
384
385         # Look for embedded YouTube player
386         matches = re.findall(r'''(?x)
387             (?:<iframe[^>]+?src=|embedSWF\(\s*)
388             (["\'])(?P<url>(?:https?:)?//(?:www\.)?youtube\.com/
389                 (?:embed|v)/.+?)
390             \1''', webpage)
391         if matches:
392             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Youtube')
393                      for tuppl in matches]
394             return self.playlist_result(
395                 urlrs, playlist_id=video_id, playlist_title=video_title)
396
397         # Look for embedded Dailymotion player
398         matches = re.findall(
399             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:www\.)?dailymotion\.com/embed/video/.+?)\1', webpage)
400         if matches:
401             urlrs = [self.url_result(unescapeHTML(tuppl[1]), 'Dailymotion')
402                      for tuppl in matches]
403             return self.playlist_result(
404                 urlrs, playlist_id=video_id, playlist_title=video_title)
405
406         # Look for embedded Wistia player
407         match = re.search(
408             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:fast\.)?wistia\.net/embed/iframe/.+?)\1', webpage)
409         if match:
410             return {
411                 '_type': 'url_transparent',
412                 'url': unescapeHTML(match.group('url')),
413                 'ie_key': 'Wistia',
414                 'uploader': video_uploader,
415                 'title': video_title,
416                 'id': video_id,
417             }
418
419         # Look for embedded blip.tv player
420         mobj = re.search(r'<meta\s[^>]*https?://api\.blip\.tv/\w+/redirect/\w+/(\d+)', webpage)
421         if mobj:
422             return self.url_result('http://blip.tv/a/a-'+mobj.group(1), 'BlipTV')
423         mobj = re.search(r'<(?:iframe|embed|object)\s[^>]*(https?://(?:\w+\.)?blip\.tv/(?:play/|api\.swf#)[a-zA-Z0-9]+)', webpage)
424         if mobj:
425             return self.url_result(mobj.group(1), 'BlipTV')
426
427         # Look for Bandcamp pages with custom domain
428         mobj = re.search(r'<meta property="og:url"[^>]*?content="(.*?bandcamp\.com.*?)"', webpage)
429         if mobj is not None:
430             burl = unescapeHTML(mobj.group(1))
431             # Don't set the extractor because it can be a track url or an album
432             return self.url_result(burl)
433
434         # Look for embedded Vevo player
435         mobj = re.search(
436             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//(?:cache\.)?vevo\.com/.+?)\1', webpage)
437         if mobj is not None:
438             return self.url_result(mobj.group('url'))
439
440         # Look for Ooyala videos
441         mobj = (re.search(r'player.ooyala.com/[^"?]+\?[^"]*?(?:embedCode|ec)=(?P<ec>[^"&]+)', webpage) or
442              re.search(r'OO.Player.create\([\'"].*?[\'"],\s*[\'"](?P<ec>.{32})[\'"]', webpage))
443         if mobj is not None:
444             return OoyalaIE._build_url_result(mobj.group('ec'))
445
446         # Look for Aparat videos
447         mobj = re.search(r'<iframe src="(http://www\.aparat\.com/video/[^"]+)"', webpage)
448         if mobj is not None:
449             return self.url_result(mobj.group(1), 'Aparat')
450
451         # Look for MPORA videos
452         mobj = re.search(r'<iframe .*?src="(http://mpora\.(?:com|de)/videos/[^"]+)"', webpage)
453         if mobj is not None:
454             return self.url_result(mobj.group(1), 'Mpora')
455
456         # Look for embedded NovaMov player
457         mobj = re.search(
458             r'<iframe[^>]+?src=(["\'])(?P<url>http://(?:(?:embed|www)\.)?novamov\.com/embed\.php.+?)\1', webpage)
459         if mobj is not None:
460             return self.url_result(mobj.group('url'), 'NovaMov')
461
462         # Look for embedded NowVideo player
463         mobj = re.search(
464             r'<iframe[^>]+?src=(["\'])(?P<url>http://(?:(?:embed|www)\.)?nowvideo\.(?:ch|sx|eu)/embed\.php.+?)\1', webpage)
465         if mobj is not None:
466             return self.url_result(mobj.group('url'), 'NowVideo')
467
468         # Look for embedded Facebook player
469         mobj = re.search(
470             r'<iframe[^>]+?src=(["\'])(?P<url>https://www\.facebook\.com/video/embed.+?)\1', webpage)
471         if mobj is not None:
472             return self.url_result(mobj.group('url'), 'Facebook')
473
474         # Look for embedded VK player
475         mobj = re.search(r'<iframe[^>]+?src=(["\'])(?P<url>https?://vk\.com/video_ext\.php.+?)\1', webpage)
476         if mobj is not None:
477             return self.url_result(mobj.group('url'), 'VK')
478
479         # Look for embedded Huffington Post player
480         mobj = re.search(
481             r'<iframe[^>]+?src=(["\'])(?P<url>https?://embed\.live\.huffingtonpost\.com/.+?)\1', webpage)
482         if mobj is not None:
483             return self.url_result(mobj.group('url'), 'HuffPost')
484
485         # Look for embed.ly
486         mobj = re.search(r'class=["\']embedly-card["\'][^>]href=["\'](?P<url>[^"\']+)', webpage)
487         if mobj is not None:
488             return self.url_result(mobj.group('url'))
489         mobj = re.search(r'class=["\']embedly-embed["\'][^>]src=["\'][^"\']*url=(?P<url>[^&]+)', webpage)
490         if mobj is not None:
491             return self.url_result(compat_urllib_parse.unquote(mobj.group('url')))
492
493         # Look for funnyordie embed
494         matches = re.findall(r'<iframe[^>]+?src="(https?://(?:www\.)?funnyordie\.com/embed/[^"]+)"', webpage)
495         if matches:
496             urlrs = [self.url_result(unescapeHTML(eurl), 'FunnyOrDie')
497                      for eurl in matches]
498             return self.playlist_result(
499                 urlrs, playlist_id=video_id, playlist_title=video_title)
500
501         # Look for embedded RUTV player
502         rutv_url = RUTVIE._extract_url(webpage)
503         if rutv_url:
504             return self.url_result(rutv_url, 'RUTV')
505
506         # Start with something easy: JW Player in SWFObject
507         mobj = re.search(r'flashvars: [\'"](?:.*&)?file=(http[^\'"&]*)', webpage)
508         if mobj is None:
509             # Look for gorilla-vid style embedding
510             mobj = re.search(r'(?s)(?:jw_plugins|JWPlayerOptions).*?file\s*:\s*["\'](.*?)["\']', webpage)
511         if mobj is None:
512             # Broaden the search a little bit
513             mobj = re.search(r'[^A-Za-z0-9]?(?:file|source)=(http[^\'"&]*)', webpage)
514         if mobj is None:
515             # Broaden the search a little bit: JWPlayer JS loader
516             mobj = re.search(r'[^A-Za-z0-9]?file["\']?:\s*["\'](http(?![^\'"]+\.[0-9]+[\'"])[^\'"]+)["\']', webpage)
517
518         # Look for embedded TED player
519         mobj = re.search(
520             r'<iframe[^>]+?src=(["\'])(?P<url>http://embed\.ted\.com/.+?)\1', webpage)
521         if mobj is not None:
522             return self.url_result(mobj.group('url'), 'TED')
523
524         if mobj is None:
525             # Try to find twitter cards info
526             mobj = re.search(r'<meta (?:property|name)="twitter:player:stream" (?:content|value)="(.+?)"', webpage)
527         if mobj is None:
528             # We look for Open Graph info:
529             # We have to match any number spaces between elements, some sites try to align them (eg.: statigr.am)
530             m_video_type = re.search(r'<meta.*?property="og:video:type".*?content="video/(.*?)"', webpage)
531             # We only look in og:video if the MIME type is a video, don't try if it's a Flash player:
532             if m_video_type is not None:
533                 mobj = re.search(r'<meta.*?property="og:video".*?content="(.*?)"', webpage)
534         if mobj is None:
535             # HTML5 video
536             mobj = re.search(r'<video[^<]*(?:>.*?<source.*?)? src="([^"]+)"', webpage, flags=re.DOTALL)
537         if mobj is None:
538             mobj = re.search(
539                 r'(?i)<meta\s+(?=(?:[a-z-]+="[^"]+"\s+)*http-equiv="refresh")'
540                 r'(?:[a-z-]+="[^"]+"\s+)*?content="[0-9]{,2};url=\'([^\']+)\'"',
541                 webpage)
542             if mobj:
543                 new_url = mobj.group(1)
544                 self.report_following_redirect(new_url)
545                 return {
546                     '_type': 'url',
547                     'url': new_url,
548                 }
549         if mobj is None:
550             raise ExtractorError('Unsupported URL: %s' % url)
551
552         # It's possible that one of the regexes
553         # matched, but returned an empty group:
554         if mobj.group(1) is None:
555             raise ExtractorError('Did not find a valid video URL at %s' % url)
556
557         video_url = mobj.group(1)
558         video_url = compat_urlparse.urljoin(url, video_url)
559         video_id = compat_urllib_parse.unquote(os.path.basename(video_url))
560
561         # Sometimes, jwplayer extraction will result in a YouTube URL
562         if YoutubeIE.suitable(video_url):
563             return self.url_result(video_url, 'Youtube')
564
565         # here's a fun little line of code for you:
566         video_id = os.path.splitext(video_id)[0]
567
568         return {
569             'id': video_id,
570             'url': video_url,
571             'uploader': video_uploader,
572             'title': video_title,
573         }