Merge pull request #9288 from reyyed/issue#9063fix
[youtube-dl] / youtube_dl / extractor / mtv.py
1 from __future__ import unicode_literals
2
3 import re
4
5 from .common import InfoExtractor
6 from ..compat import (
7     compat_urllib_parse_urlencode,
8     compat_str,
9     compat_xpath,
10 )
11 from ..utils import (
12     ExtractorError,
13     find_xpath_attr,
14     fix_xml_ampersands,
15     float_or_none,
16     HEADRequest,
17     sanitized_Request,
18     unescapeHTML,
19     url_basename,
20     RegexNotFoundError,
21     xpath_text,
22 )
23
24
25 def _media_xml_tag(tag):
26     return '{http://search.yahoo.com/mrss/}%s' % tag
27
28
29 class MTVServicesInfoExtractor(InfoExtractor):
30     _MOBILE_TEMPLATE = None
31     _LANG = None
32
33     @staticmethod
34     def _id_from_uri(uri):
35         return uri.split(':')[-1]
36
37     # This was originally implemented for ComedyCentral, but it also works here
38     @staticmethod
39     def _transform_rtmp_url(rtmp_video_url):
40         m = re.match(r'^rtmpe?://.*?/(?P<finalid>gsp\..+?/.*)$', rtmp_video_url)
41         if not m:
42             return rtmp_video_url
43         base = 'http://viacommtvstrmfs.fplive.net/'
44         return base + m.group('finalid')
45
46     def _get_feed_url(self, uri):
47         return self._FEED_URL
48
49     def _get_thumbnail_url(self, uri, itemdoc):
50         search_path = '%s/%s' % (_media_xml_tag('group'), _media_xml_tag('thumbnail'))
51         thumb_node = itemdoc.find(search_path)
52         if thumb_node is None:
53             return None
54         else:
55             return thumb_node.attrib['url']
56
57     def _extract_mobile_video_formats(self, mtvn_id):
58         webpage_url = self._MOBILE_TEMPLATE % mtvn_id
59         req = sanitized_Request(webpage_url)
60         # Otherwise we get a webpage that would execute some javascript
61         req.add_header('User-Agent', 'curl/7')
62         webpage = self._download_webpage(req, mtvn_id,
63                                          'Downloading mobile page')
64         metrics_url = unescapeHTML(self._search_regex(r'<a href="(http://metrics.+?)"', webpage, 'url'))
65         req = HEADRequest(metrics_url)
66         response = self._request_webpage(req, mtvn_id, 'Resolving url')
67         url = response.geturl()
68         # Transform the url to get the best quality:
69         url = re.sub(r'.+pxE=mp4', 'http://mtvnmobile.vo.llnwd.net/kip0/_pxn=0+_pxK=18639+_pxE=mp4', url, 1)
70         return [{'url': url, 'ext': 'mp4'}]
71
72     def _extract_video_formats(self, mdoc, mtvn_id):
73         if re.match(r'.*/(error_country_block\.swf|geoblock\.mp4|copyright_error\.flv(?:\?geo\b.+?)?)$', mdoc.find('.//src').text) is not None:
74             if mtvn_id is not None and self._MOBILE_TEMPLATE is not None:
75                 self.to_screen('The normal version is not available from your '
76                                'country, trying with the mobile version')
77                 return self._extract_mobile_video_formats(mtvn_id)
78             raise ExtractorError('This video is not available from your country.',
79                                  expected=True)
80
81         formats = []
82         for rendition in mdoc.findall('.//rendition'):
83             try:
84                 _, _, ext = rendition.attrib['type'].partition('/')
85                 rtmp_video_url = rendition.find('./src').text
86                 if rtmp_video_url.endswith('siteunavail.png'):
87                     continue
88                 new_url = self._transform_rtmp_url(rtmp_video_url)
89                 formats.append({
90                     'ext': 'flv' if new_url.startswith('rtmp') else ext,
91                     'url': new_url,
92                     'format_id': rendition.get('bitrate'),
93                     'width': int(rendition.get('width')),
94                     'height': int(rendition.get('height')),
95                 })
96             except (KeyError, TypeError):
97                 raise ExtractorError('Invalid rendition field.')
98         self._sort_formats(formats)
99         return formats
100
101     def _extract_subtitles(self, mdoc, mtvn_id):
102         subtitles = {}
103         for transcript in mdoc.findall('.//transcript'):
104             if transcript.get('kind') != 'captions':
105                 continue
106             lang = transcript.get('srclang')
107             subtitles[lang] = [{
108                 'url': compat_str(typographic.get('src')),
109                 'ext': typographic.get('format')
110             } for typographic in transcript.findall('./typographic')]
111         return subtitles
112
113     def _get_video_info(self, itemdoc):
114         uri = itemdoc.find('guid').text
115         video_id = self._id_from_uri(uri)
116         self.report_extraction(video_id)
117         content_el = itemdoc.find('%s/%s' % (_media_xml_tag('group'), _media_xml_tag('content')))
118         mediagen_url = content_el.attrib['url']
119         # Remove the templates, like &device={device}
120         mediagen_url = re.sub(r'&[^=]*?={.*?}(?=(&|$))', '', mediagen_url)
121         if 'acceptMethods' not in mediagen_url:
122             mediagen_url += '&' if '?' in mediagen_url else '?'
123             mediagen_url += 'acceptMethods=fms'
124
125         mediagen_doc = self._download_xml(mediagen_url, video_id,
126                                           'Downloading video urls')
127
128         item = mediagen_doc.find('./video/item')
129         if item is not None and item.get('type') == 'text':
130             message = '%s returned error: ' % self.IE_NAME
131             if item.get('code') is not None:
132                 message += '%s - ' % item.get('code')
133             message += item.text
134             raise ExtractorError(message, expected=True)
135
136         description = xpath_text(itemdoc, 'description')
137
138         title_el = None
139         if title_el is None:
140             title_el = find_xpath_attr(
141                 itemdoc, './/{http://search.yahoo.com/mrss/}category',
142                 'scheme', 'urn:mtvn:video_title')
143         if title_el is None:
144             title_el = itemdoc.find(compat_xpath('.//{http://search.yahoo.com/mrss/}title'))
145         if title_el is None:
146             title_el = itemdoc.find(compat_xpath('.//title'))
147             if title_el.text is None:
148                 title_el = None
149
150         title = title_el.text
151         if title is None:
152             raise ExtractorError('Could not find video title')
153         title = title.strip()
154
155         # This a short id that's used in the webpage urls
156         mtvn_id = None
157         mtvn_id_node = find_xpath_attr(itemdoc, './/{http://search.yahoo.com/mrss/}category',
158                                        'scheme', 'urn:mtvn:id')
159         if mtvn_id_node is not None:
160             mtvn_id = mtvn_id_node.text
161
162         return {
163             'title': title,
164             'formats': self._extract_video_formats(mediagen_doc, mtvn_id),
165             'subtitles': self._extract_subtitles(mediagen_doc, mtvn_id),
166             'id': video_id,
167             'thumbnail': self._get_thumbnail_url(uri, itemdoc),
168             'description': description,
169             'duration': float_or_none(content_el.attrib.get('duration')),
170         }
171
172     def _get_feed_query(self, uri):
173         data = {'uri': uri}
174         if self._LANG:
175             data['lang'] = self._LANG
176         return compat_urllib_parse_urlencode(data)
177
178     def _get_videos_info(self, uri):
179         video_id = self._id_from_uri(uri)
180         feed_url = self._get_feed_url(uri)
181         info_url = feed_url + '?' + self._get_feed_query(uri)
182         return self._get_videos_info_from_url(info_url, video_id)
183
184     def _get_videos_info_from_url(self, url, video_id):
185         idoc = self._download_xml(
186             url, video_id,
187             'Downloading info', transform_source=fix_xml_ampersands)
188         return self.playlist_result(
189             [self._get_video_info(item) for item in idoc.findall('.//item')])
190
191     def _extract_mgid(self, webpage):
192         try:
193             # the url can be http://media.mtvnservices.com/fb/{mgid}.swf
194             # or http://media.mtvnservices.com/{mgid}
195             og_url = self._og_search_video_url(webpage)
196             mgid = url_basename(og_url)
197             if mgid.endswith('.swf'):
198                 mgid = mgid[:-4]
199         except RegexNotFoundError:
200             mgid = None
201
202         if mgid is None or ':' not in mgid:
203             mgid = self._search_regex(
204                 [r'data-mgid="(.*?)"', r'swfobject.embedSWF\(".*?(mgid:.*?)"'],
205                 webpage, 'mgid', default=None)
206
207         if not mgid:
208             sm4_embed = self._html_search_meta(
209                 'sm4:video:embed', webpage, 'sm4 embed', default='')
210             mgid = self._search_regex(
211                 r'embed/(mgid:.+?)["\'&?/]', sm4_embed, 'mgid')
212         return mgid
213
214     def _real_extract(self, url):
215         title = url_basename(url)
216         webpage = self._download_webpage(url, title)
217         mgid = self._extract_mgid(webpage)
218         videos_info = self._get_videos_info(mgid)
219         return videos_info
220
221
222 class MTVServicesEmbeddedIE(MTVServicesInfoExtractor):
223     IE_NAME = 'mtvservices:embedded'
224     _VALID_URL = r'https?://media\.mtvnservices\.com/embed/(?P<mgid>.+?)(\?|/|$)'
225
226     _TEST = {
227         # From http://www.thewrap.com/peter-dinklage-sums-up-game-of-thrones-in-45-seconds-video/
228         'url': 'http://media.mtvnservices.com/embed/mgid:uma:video:mtv.com:1043906/cp~vid%3D1043906%26uri%3Dmgid%3Auma%3Avideo%3Amtv.com%3A1043906',
229         'md5': 'cb349b21a7897164cede95bd7bf3fbb9',
230         'info_dict': {
231             'id': '1043906',
232             'ext': 'mp4',
233             'title': 'Peter Dinklage Sums Up \'Game Of Thrones\' In 45 Seconds',
234             'description': '"Sexy sexy sexy, stabby stabby stabby, beautiful language," says Peter Dinklage as he tries summarizing "Game of Thrones" in under a minute.',
235         },
236     }
237
238     @staticmethod
239     def _extract_url(webpage):
240         mobj = re.search(
241             r'<iframe[^>]+?src=(["\'])(?P<url>(?:https?:)?//media.mtvnservices.com/embed/.+?)\1', webpage)
242         if mobj:
243             return mobj.group('url')
244
245     def _get_feed_url(self, uri):
246         video_id = self._id_from_uri(uri)
247         site_id = uri.replace(video_id, '')
248         config_url = ('http://media.mtvnservices.com/pmt/e1/players/{0}/'
249                       'context4/context5/config.xml'.format(site_id))
250         config_doc = self._download_xml(config_url, video_id)
251         feed_node = config_doc.find('.//feed')
252         feed_url = feed_node.text.strip().split('?')[0]
253         return feed_url
254
255     def _real_extract(self, url):
256         mobj = re.match(self._VALID_URL, url)
257         mgid = mobj.group('mgid')
258         return self._get_videos_info(mgid)
259
260
261 class MTVIE(MTVServicesInfoExtractor):
262     _VALID_URL = r'''(?x)^https?://
263         (?:(?:www\.)?mtv\.com/videos/.+?/(?P<videoid>[0-9]+)/[^/]+$|
264            m\.mtv\.com/videos/video\.rbml\?.*?id=(?P<mgid>[^&]+))'''
265
266     _FEED_URL = 'http://www.mtv.com/player/embed/AS3/rss/'
267
268     _TESTS = [
269         {
270             'url': 'http://www.mtv.com/videos/misc/853555/ours-vh1-storytellers.jhtml',
271             'md5': '850f3f143316b1e71fa56a4edfd6e0f8',
272             'info_dict': {
273                 'id': '853555',
274                 'ext': 'mp4',
275                 'title': 'Taylor Swift - "Ours (VH1 Storytellers)"',
276                 'description': 'Album: Taylor Swift performs "Ours" for VH1 Storytellers at Harvey Mudd College.',
277             },
278         },
279     ]
280
281     def _get_thumbnail_url(self, uri, itemdoc):
282         return 'http://mtv.mtvnimages.com/uri/' + uri
283
284     def _real_extract(self, url):
285         mobj = re.match(self._VALID_URL, url)
286         video_id = mobj.group('videoid')
287         uri = mobj.groupdict().get('mgid')
288         if uri is None:
289             webpage = self._download_webpage(url, video_id)
290
291             # Some videos come from Vevo.com
292             m_vevo = re.search(
293                 r'(?s)isVevoVideo = true;.*?vevoVideoId = "(.*?)";', webpage)
294             if m_vevo:
295                 vevo_id = m_vevo.group(1)
296                 self.to_screen('Vevo video detected: %s' % vevo_id)
297                 return self.url_result('vevo:%s' % vevo_id, ie='Vevo')
298
299             uri = self._html_search_regex(r'/uri/(.*?)\?', webpage, 'uri')
300         return self._get_videos_info(uri)
301
302
303 class MTVIggyIE(MTVServicesInfoExtractor):
304     IE_NAME = 'mtviggy.com'
305     _VALID_URL = r'https?://www\.mtviggy\.com/videos/.+'
306     _TEST = {
307         'url': 'http://www.mtviggy.com/videos/arcade-fire-behind-the-scenes-at-the-biggest-music-experiment-yet/',
308         'info_dict': {
309             'id': '984696',
310             'ext': 'mp4',
311             'title': 'Arcade Fire: Behind the Scenes at the Biggest Music Experiment Yet',
312         }
313     }
314     _FEED_URL = 'http://all.mtvworldverticals.com/feed-xml/'
315
316
317 class MTVDEIE(MTVServicesInfoExtractor):
318     IE_NAME = 'mtv.de'
319     _VALID_URL = r'https?://(?:www\.)?mtv\.de/(?:artists|shows|news)/(?:[^/]+/)*(?P<id>\d+)-[^/#?]+/*(?:[#?].*)?$'
320     _TESTS = [{
321         'url': 'http://www.mtv.de/artists/10571-cro/videos/61131-traum',
322         'info_dict': {
323             'id': 'music_video-a50bc5f0b3aa4b3190aa',
324             'ext': 'mp4',
325             'title': 'MusicVideo_cro-traum',
326             'description': 'Cro - Traum',
327         },
328         'params': {
329             # rtmp download
330             'skip_download': True,
331         },
332     }, {
333         # mediagen URL without query (e.g. http://videos.mtvnn.com/mediagen/e865da714c166d18d6f80893195fcb97)
334         'url': 'http://www.mtv.de/shows/933-teen-mom-2/staffeln/5353/folgen/63565-enthullungen',
335         'info_dict': {
336             'id': 'local_playlist-f5ae778b9832cc837189',
337             'ext': 'mp4',
338             'title': 'Episode_teen-mom-2_shows_season-5_episode-1_full-episode_part1',
339         },
340         'params': {
341             # rtmp download
342             'skip_download': True,
343         },
344     }, {
345         # single video in pagePlaylist with different id
346         'url': 'http://www.mtv.de/news/77491-mtv-movies-spotlight-pixels-teil-3',
347         'info_dict': {
348             'id': 'local_playlist-4e760566473c4c8c5344',
349             'ext': 'mp4',
350             'title': 'Article_mtv-movies-spotlight-pixels-teil-3_short-clips_part1',
351             'description': 'MTV Movies Supercut',
352         },
353         'params': {
354             # rtmp download
355             'skip_download': True,
356         },
357     }]
358
359     def _real_extract(self, url):
360         video_id = self._match_id(url)
361
362         webpage = self._download_webpage(url, video_id)
363
364         playlist = self._parse_json(
365             self._search_regex(
366                 r'window\.pagePlaylist\s*=\s*(\[.+?\]);\n', webpage, 'page playlist'),
367             video_id)
368
369         # news pages contain single video in playlist with different id
370         if len(playlist) == 1:
371             return self._get_videos_info_from_url(playlist[0]['mrss'], video_id)
372
373         for item in playlist:
374             item_id = item.get('id')
375             if item_id and compat_str(item_id) == video_id:
376                 return self._get_videos_info_from_url(item['mrss'], video_id)