_ Git - youtube-dl/blob - youtube_dl/extractor/tele5.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 from .common import InfoExtractor
   5 from .nexx import NexxIE
   6 from ..compat import compat_urlparse
   7
   8
   9 class Tele5IE(InfoExtractor):
  10     _VALID_URL = r'https?://(?:www\.)?tele5\.de/(?:[^/]+/)*(?P<id>[^/?#&]+)'
  11     _TESTS = [{
  12         'url': 'https://www.tele5.de/mediathek/filme-online/videos?vid=1549416',
  13         'info_dict': {
  14             'id': '1549416',
  15             'ext': 'mp4',
  16             'upload_date': '20180814',
  17             'timestamp': 1534290623,
  18             'title': 'Pandorum',
  19         },
  20         'params': {
  21             'skip_download': True,
  22         },
  23     }, {
  24         'url': 'https://www.tele5.de/kalkofes-mattscheibe/video-clips/politik-und-gesellschaft?ve_id=1551191',
  25         'only_matching': True,
  26     }, {
  27         'url': 'https://www.tele5.de/video-clip/?ve_id=1609440',
  28         'only_matching': True,
  29     }, {
  30         'url': 'https://www.tele5.de/filme/schlefaz-dragon-crusaders/',
  31         'only_matching': True,
  32     }, {
  33         'url': 'https://www.tele5.de/filme/making-of/avengers-endgame/',
  34         'only_matching': True,
  35     }, {
  36         'url': 'https://www.tele5.de/star-trek/raumschiff-voyager/ganze-folge/das-vinculum/',
  37         'only_matching': True,
  38     }, {
  39         'url': 'https://www.tele5.de/anders-ist-sevda/',
  40         'only_matching': True,
  41     }]
  42
  43     def _real_extract(self, url):
  44         qs = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
  45         video_id = (qs.get('vid') or qs.get('ve_id') or [None])[0]
  46
  47         if not video_id:
  48             display_id = self._match_id(url)
  49             webpage = self._download_webpage(url, display_id)
  50             video_id = self._html_search_regex(
  51                 (r'id\s*=\s*["\']video-player["\'][^>]+data-id\s*=\s*["\'](\d+)',
  52                  r'\s+id\s*=\s*["\']player_(\d{6,})',
  53                  r'\bdata-id\s*=\s*["\'](\d{6,})'), webpage, 'video id')
  54
  55         return self.url_result(
  56             'https://api.nexx.cloud/v3/759/videos/byid/%s' % video_id,
  57             ie=NexxIE.ie_key(), video_id=video_id)