youtube_dl/extractor/condenast.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5 import json
   6
   7 from .common import InfoExtractor
   8 from ..utils import (
   9     compat_urllib_parse,
  10     orderedSet,
  11     compat_urllib_parse_urlparse,
  12     compat_urlparse,
  13 )
  14
  15
  16 class CondeNastIE(InfoExtractor):
  17     """
  18     Condé Nast is a media group, some of its sites use a custom HTML5 player
  19     that works the same in all of them.
  20     """
  21
  22     # The keys are the supported sites and the values are the name to be shown
  23     # to the user and in the extractor description.
  24     _SITES = {
  25         'wired': 'WIRED',
  26         'gq': 'GQ',
  27         'vogue': 'Vogue',
  28         'glamour': 'Glamour',
  29         'wmagazine': 'W Magazine',
  30         'vanityfair': 'Vanity Fair',
  31     }
  32
  33     _VALID_URL = r'http://(video|www).(?P<site>%s).com/(?P<type>watch|series|video)/(?P<id>.+)' % '|'.join(_SITES.keys())
  34     IE_DESC = 'Condé Nast media group: %s' % ', '.join(sorted(_SITES.values()))
  35
  36     _TEST = {
  37         'url': 'http://video.wired.com/watch/3d-printed-speakers-lit-with-led',
  38         'file': '5171b343c2b4c00dd0c1ccb3.mp4',
  39         'md5': '1921f713ed48aabd715691f774c451f7',
  40         'info_dict': {
  41             'title': '3D Printed Speakers Lit With LED',
  42             'description': 'Check out these beautiful 3D printed LED speakers.  You can\'t actually buy them, but LumiGeek is working on a board that will let you make you\'re own.',
  43         }
  44     }
  45
  46     def _extract_series(self, url, webpage):
  47         title = self._html_search_regex(r'<div class="cne-series-info">.*?<h1>(.+?)</h1>',
  48                                         webpage, 'series title', flags=re.DOTALL)
  49         url_object = compat_urllib_parse_urlparse(url)
  50         base_url = '%s://%s' % (url_object.scheme, url_object.netloc)
  51         m_paths = re.finditer(r'<p class="cne-thumb-title">.*?<a href="(/watch/.+?)["\?]',
  52                               webpage, flags=re.DOTALL)
  53         paths = orderedSet(m.group(1) for m in m_paths)
  54         build_url = lambda path: compat_urlparse.urljoin(base_url, path)
  55         entries = [self.url_result(build_url(path), 'CondeNast') for path in paths]
  56         return self.playlist_result(entries, playlist_title=title)
  57
  58     def _extract_video(self, webpage):
  59         description = self._html_search_regex([r'<div class="cne-video-description">(.+?)</div>',
  60                                                r'<div class="video-post-content">(.+?)</div>',
  61                                                ],
  62                                               webpage, 'description',
  63                                               fatal=False, flags=re.DOTALL)
  64         params = self._search_regex(r'var params = {(.+?)}[;,]', webpage,
  65                                     'player params', flags=re.DOTALL)
  66         video_id = self._search_regex(r'videoId: [\'"](.+?)[\'"]', params, 'video id')
  67         player_id = self._search_regex(r'playerId: [\'"](.+?)[\'"]', params, 'player id')
  68         target = self._search_regex(r'target: [\'"](.+?)[\'"]', params, 'target')
  69         data = compat_urllib_parse.urlencode({'videoId': video_id,
  70                                               'playerId': player_id,
  71                                               'target': target,
  72                                               })
  73         base_info_url = self._search_regex(r'url = [\'"](.+?)[\'"][,;]',
  74                                            webpage, 'base info url',
  75                                            default='http://player.cnevids.com/player/loader.js?')
  76         info_url = base_info_url + data
  77         info_page = self._download_webpage(info_url, video_id,
  78                                            'Downloading video info')
  79         video_info = self._search_regex(r'var video = ({.+?});', info_page, 'video info')
  80         video_info = json.loads(video_info)
  81
  82         def _formats_sort_key(f):
  83             type_ord = 1 if f['type'] == 'video/mp4' else 0
  84             quality_ord = 1 if f['quality'] == 'high' else 0
  85             return (quality_ord, type_ord)
  86         best_format = sorted(video_info['sources'][0], key=_formats_sort_key)[-1]
  87
  88         return {'id': video_id,
  89                 'url': best_format['src'],
  90                 'ext': best_format['type'].split('/')[-1],
  91                 'title': video_info['title'],
  92                 'thumbnail': video_info['poster_frame'],
  93                 'description': description,
  94                 }
  95
  96     def _real_extract(self, url):
  97         mobj = re.match(self._VALID_URL, url)
  98         site = mobj.group('site')
  99         url_type = mobj.group('type')
 100         id = mobj.group('id')
 101
 102         self.to_screen(u'Extracting from %s with the Condé Nast extractor' % self._SITES[site])
 103         webpage = self._download_webpage(url, id)
 104
 105         if url_type == 'series':
 106             return self._extract_series(url, webpage)
 107         else:
 108             return self._extract_video(webpage)