yt_dlp/extractor/aljazeera.py

   1 from __future__ import unicode_literals
   2
   3 import json
   4 import re
   5
   6 from .common import InfoExtractor
   7
   8
   9 class AlJazeeraIE(InfoExtractor):
  10     _VALID_URL = r'https?://(?:www\.)?aljazeera\.com/(?P<type>program/[^/]+|(?:feature|video)s)/\d{4}/\d{1,2}/\d{1,2}/(?P<id>[^/?&#]+)'
  11
  12     _TESTS = [{
  13         'url': 'https://www.aljazeera.com/program/episode/2014/9/19/deliverance',
  14         'info_dict': {
  15             'id': '3792260579001',
  16             'ext': 'mp4',
  17             'title': 'The Slum - Episode 1: Deliverance',
  18             'description': 'As a birth attendant advocating for family planning, Remy is on the frontline of Tondo\'s battle with overcrowding.',
  19             'uploader_id': '665003303001',
  20             'timestamp': 1411116829,
  21             'upload_date': '20140919',
  22         },
  23         'add_ie': ['BrightcoveNew'],
  24         'skip': 'Not accessible from Travis CI server',
  25     }, {
  26         'url': 'https://www.aljazeera.com/videos/2017/5/11/sierra-leone-709-carat-diamond-to-be-auctioned-off',
  27         'only_matching': True,
  28     }, {
  29         'url': 'https://www.aljazeera.com/features/2017/8/21/transforming-pakistans-buses-into-art',
  30         'only_matching': True,
  31     }]
  32     BRIGHTCOVE_URL_TEMPLATE = 'http://players.brightcove.net/%s/%s_default/index.html?videoId=%s'
  33
  34     def _real_extract(self, url):
  35         post_type, name = re.match(self._VALID_URL, url).groups()
  36         post_type = {
  37             'features': 'post',
  38             'program': 'episode',
  39             'videos': 'video',
  40         }[post_type.split('/')[0]]
  41         video = self._download_json(
  42             'https://www.aljazeera.com/graphql', name, query={
  43                 'operationName': 'SingleArticleQuery',
  44                 'variables': json.dumps({
  45                     'name': name,
  46                     'postType': post_type,
  47                 }),
  48             }, headers={
  49                 'wp-site': 'aje',
  50             })['data']['article']['video']
  51         video_id = video['id']
  52         account_id = video.get('accountId') or '665003303001'
  53         player_id = video.get('playerId') or 'BkeSH5BDb'
  54         return self.url_result(
  55             self.BRIGHTCOVE_URL_TEMPLATE % (account_id, player_id, video_id),
  56             'BrightcoveNew', video_id)