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,
   8     compat_urllib_request,
   9     compat_str,
  10 )
  11 from ..utils import (
  12     ExtractorError,
  13     find_xpath_attr,
  14     fix_xml_ampersands,
  15     HEADRequest,
  16     unescapeHTML,
  17     url_basename,
  18     RegexNotFoundError,
  19 )
  20
  21
  22 def _media_xml_tag(tag):
  23     return '{http://search.yahoo.com/mrss/}%s' % tag
  24
  25
  26 class MTVServicesInfoExtractor(InfoExtractor):
  27     _MOBILE_TEMPLATE = None
  28
  29     @staticmethod
  30     def _id_from_uri(uri):
  31         return uri.split(':')[-1]
  32
  33     # This was originally implemented for ComedyCentral, but it also works here
  34     @staticmethod
  35     def _transform_rtmp_url(rtmp_video_url):
  36         m = re.match(r'^rtmpe?://.*?/(?P<finalid>gsp\..+?/.*)$', rtmp_video_url)
  37         if not m:
  38             return rtmp_video_url
  39         base = 'http://viacommtvstrmfs.fplive.net/'
  40         return base + m.group('finalid')
  41
  42     def _get_feed_url(self, uri):
  43         return self._FEED_URL
  44
  45     def _get_thumbnail_url(self, uri, itemdoc):
  46         search_path = '%s/%s' % (_media_xml_tag('group'), _media_xml_tag('thumbnail'))
  47         thumb_node = itemdoc.find(search_path)
  48         if thumb_node is None:
  49             return None
  50         else:
  51             return thumb_node.attrib['url']
  52
  53     def _extract_mobile_video_formats(self, mtvn_id):
  54         webpage_url = self._MOBILE_TEMPLATE % mtvn_id
  55         req = compat_urllib_request.Request(webpage_url)
  56         # Otherwise we get a webpage that would execute some javascript
  57         req.add_header('User-Agent', 'curl/7')
  58         webpage = self._download_webpage(req, mtvn_id,
  59                                          'Downloading mobile page')
  60         metrics_url = unescapeHTML(self._search_regex(r'<a href="(http://metrics.+?)"', webpage, 'url'))
  61         req = HEADRequest(metrics_url)
  62         response = self._request_webpage(req, mtvn_id, 'Resolving url')
  63         url = response.geturl()
  64         # Transform the url to get the best quality:
  65         url = re.sub(r'.+pxE=mp4', 'http://mtvnmobile.vo.llnwd.net/kip0/_pxn=0+_pxK=18639+_pxE=mp4', url, 1)
  66         return [{'url': url, 'ext': 'mp4'}]
  67
  68     def _extract_video_formats(self, mdoc, mtvn_id):
  69         if re.match(r'.*/(error_country_block\.swf|geoblock\.mp4)$', mdoc.find('.//src').text) is not None:
  70             if mtvn_id is not None and self._MOBILE_TEMPLATE is not None:
  71                 self.to_screen('The normal version is not available from your '
  72                                'country, trying with the mobile version')
  73                 return self._extract_mobile_video_formats(mtvn_id)
  74             raise ExtractorError('This video is not available from your country.',
  75                                  expected=True)
  76
  77         formats = []
  78         for rendition in mdoc.findall('.//rendition'):
  79             try:
  80                 _, _, ext = rendition.attrib['type'].partition('/')
  81                 rtmp_video_url = rendition.find('./src').text
  82                 if rtmp_video_url.endswith('siteunavail.png'):
  83                     continue
  84                 formats.append({
  85                     'ext': ext,
  86                     'url': self._transform_rtmp_url(rtmp_video_url),
  87                     'format_id': rendition.get('bitrate'),
  88                     'width': int(rendition.get('width')),
  89                     'height': int(rendition.get('height')),
  90                 })
  91             except (KeyError, TypeError):
  92                 raise ExtractorError('Invalid rendition field.')
  93         self._sort_formats(formats)
  94         return formats
  95
  96     def _extract_subtitles(self, mdoc, mtvn_id):
  97         subtitles = {}
  98         for transcript in mdoc.findall('.//transcript'):
  99             if transcript.get('kind') != 'captions':
 100                 continue
 101             lang = transcript.get('srclang')
 102             subtitles[lang] = [{
 103                 'url': compat_str(typographic.get('src')),
 104                 'ext': typographic.get('format')
 105             } for typographic in transcript.findall('./typographic')]
 106         return subtitles
 107
 108     def _get_video_info(self, itemdoc):
 109         uri = itemdoc.find('guid').text
 110         video_id = self._id_from_uri(uri)
 111         self.report_extraction(video_id)
 112         mediagen_url = itemdoc.find('%s/%s' % (_media_xml_tag('group'), _media_xml_tag('content'))).attrib['url']
 113         # Remove the templates, like &device={device}
 114         mediagen_url = re.sub(r'&[^=]*?={.*?}(?=(&|$))', '', mediagen_url)
 115         if 'acceptMethods' not in mediagen_url:
 116             mediagen_url += '&acceptMethods=fms'
 117
 118         mediagen_doc = self._download_xml(mediagen_url, video_id,
 119                                           'Downloading video urls')
 120
 121         item = mediagen_doc.find('./video/item')
 122         if item is not None and item.get('type') == 'text':
 123             message = '%s returned error: ' % self.IE_NAME
 124             if item.get('code') is not None:
 125                 message += '%s - ' % item.get('code')
 126             message += item.text
 127             raise ExtractorError(message, expected=True)
 128
 129         description_node = itemdoc.find('description')
 130         if description_node is not None:
 131             description = description_node.text.strip()
 132         else:
 133             description = None
 134
 135         title_el = None
 136         if title_el is None:
 137             title_el = find_xpath_attr(
 138                 itemdoc, './/{http://search.yahoo.com/mrss/}category',
 139                 'scheme', 'urn:mtvn:video_title')
 140         if title_el is None:
 141             title_el = itemdoc.find('.//{http://search.yahoo.com/mrss/}title')
 142         if title_el is None:
 143             title_el = itemdoc.find('.//title')
 144             if title_el.text is None:
 145                 title_el = None
 146
 147         title = title_el.text
 148         if title is None:
 149             raise ExtractorError('Could not find video title')
 150         title = title.strip()
 151
 152         # This a short id that's used in the webpage urls
 153         mtvn_id = None
 154         mtvn_id_node = find_xpath_attr(itemdoc, './/{http://search.yahoo.com/mrss/}category',
 155                                        'scheme', 'urn:mtvn:id')
 156         if mtvn_id_node is not None:
 157             mtvn_id = mtvn_id_node.text
 158
 159         return {
 160             'title': title,
 161             'formats': self._extract_video_formats(mediagen_doc, mtvn_id),
 162             'subtitles': self._extract_subtitles(mediagen_doc, mtvn_id),
 163             'id': video_id,
 164             'thumbnail': self._get_thumbnail_url(uri, itemdoc),
 165             'description': description,
 166         }
 167
 168     def _get_videos_info(self, uri):
 169         video_id = self._id_from_uri(uri)
 170         feed_url = self._get_feed_url(uri)
 171         data = compat_urllib_parse.urlencode({'uri': uri})
 172         idoc = self._download_xml(
 173             feed_url + '?' + data, video_id,
 174             'Downloading info', transform_source=fix_xml_ampersands)
 175         return self.playlist_result(
 176             [self._get_video_info(item) for item in idoc.findall('.//item')])
 177
 178     def _real_extract(self, url):
 179         title = url_basename(url)
 180         webpage = self._download_webpage(url, title)
 181         try:
 182             # the url can be http://media.mtvnservices.com/fb/{mgid}.swf
 183             # or http://media.mtvnservices.com/{mgid}
 184             og_url = self._og_search_video_url(webpage)
 185             mgid = url_basename(og_url)
 186             if mgid.endswith('.swf'):
 187                 mgid = mgid[:-4]
 188         except RegexNotFoundError:
 189             mgid = None
 190
 191         if mgid is None or ':' not in mgid:
 192             mgid = self._search_regex(
 193                 [r'data-mgid="(.*?)"', r'swfobject.embedSWF\(".*?(mgid:.*?)"'],
 194                 webpage, 'mgid')
 195
 196         videos_info = self._get_videos_info(mgid)
 197         return videos_info
 198
 199
 200 class MTVServicesEmbeddedIE(MTVServicesInfoExtractor):
 201     IE_NAME = 'mtvservices:embedded'
 202     _VALID_URL = r'https?://media\.mtvnservices\.com/embed/(?P<mgid>.+?)(\?|/|$)'
 203
 204     _TEST = {
 205         # From http://www.thewrap.com/peter-dinklage-sums-up-game-of-thrones-in-45-seconds-video/
 206         'url': 'http://media.mtvnservices.com/embed/mgid:uma:video:mtv.com:1043906/cp~vid%3D1043906%26uri%3Dmgid%3Auma%3Avideo%3Amtv.com%3A1043906',
 207         'md5': 'cb349b21a7897164cede95bd7bf3fbb9',
 208         'info_dict': {
 209             'id': '1043906',
 210             'ext': 'mp4',
 211             'title': 'Peter Dinklage Sums Up \'Game Of Thrones\' In 45 Seconds',
 212             'description': '"Sexy sexy sexy, stabby stabby stabby, beautiful language," says Peter Dinklage as he tries summarizing "Game of Thrones" in under a minute.',
 213         },
 214     }
 215
 216     def _get_feed_url(self, uri):
 217         video_id = self._id_from_uri(uri)
 218         site_id = uri.replace(video_id, '')
 219         config_url = ('http://media.mtvnservices.com/pmt/e1/players/{0}/'
 220                       'context4/context5/config.xml'.format(site_id))
 221         config_doc = self._download_xml(config_url, video_id)
 222         feed_node = config_doc.find('.//feed')
 223         feed_url = feed_node.text.strip().split('?')[0]
 224         return feed_url
 225
 226     def _real_extract(self, url):
 227         mobj = re.match(self._VALID_URL, url)
 228         mgid = mobj.group('mgid')
 229         return self._get_videos_info(mgid)
 230
 231
 232 class MTVIE(MTVServicesInfoExtractor):
 233     _VALID_URL = r'''(?x)^https?://
 234         (?:(?:www\.)?mtv\.com/videos/.+?/(?P<videoid>[0-9]+)/[^/]+$|
 235            m\.mtv\.com/videos/video\.rbml\?.*?id=(?P<mgid>[^&]+))'''
 236
 237     _FEED_URL = 'http://www.mtv.com/player/embed/AS3/rss/'
 238
 239     _TESTS = [
 240         {
 241             'url': 'http://www.mtv.com/videos/misc/853555/ours-vh1-storytellers.jhtml',
 242             'md5': '850f3f143316b1e71fa56a4edfd6e0f8',
 243             'info_dict': {
 244                 'id': '853555',
 245                 'ext': 'mp4',
 246                 'title': 'Taylor Swift - "Ours (VH1 Storytellers)"',
 247                 'description': 'Album: Taylor Swift performs "Ours" for VH1 Storytellers at Harvey Mudd College.',
 248             },
 249         },
 250     ]
 251
 252     def _get_thumbnail_url(self, uri, itemdoc):
 253         return 'http://mtv.mtvnimages.com/uri/' + uri
 254
 255     def _real_extract(self, url):
 256         mobj = re.match(self._VALID_URL, url)
 257         video_id = mobj.group('videoid')
 258         uri = mobj.groupdict().get('mgid')
 259         if uri is None:
 260             webpage = self._download_webpage(url, video_id)
 261
 262             # Some videos come from Vevo.com
 263             m_vevo = re.search(
 264                 r'(?s)isVevoVideo = true;.*?vevoVideoId = "(.*?)";', webpage)
 265             if m_vevo:
 266                 vevo_id = m_vevo.group(1)
 267                 self.to_screen('Vevo video detected: %s' % vevo_id)
 268                 return self.url_result('vevo:%s' % vevo_id, ie='Vevo')
 269
 270             uri = self._html_search_regex(r'/uri/(.*?)\?', webpage, 'uri')
 271         return self._get_videos_info(uri)
 272
 273
 274 class MTVIggyIE(MTVServicesInfoExtractor):
 275     IE_NAME = 'mtviggy.com'
 276     _VALID_URL = r'https?://www\.mtviggy\.com/videos/.+'
 277     _TEST = {
 278         'url': 'http://www.mtviggy.com/videos/arcade-fire-behind-the-scenes-at-the-biggest-music-experiment-yet/',
 279         'info_dict': {
 280             'id': '984696',
 281             'ext': 'mp4',
 282             'title': 'Arcade Fire: Behind the Scenes at the Biggest Music Experiment Yet',
 283         }
 284     }
 285     _FEED_URL = 'http://all.mtvworldverticals.com/feed-xml/'