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