yt_dlp/extractor/vbox7.py

   1 from .common import InfoExtractor
   2 from ..utils import ExtractorError
   3
   4
   5 class Vbox7IE(InfoExtractor):
   6     _VALID_URL = r'''(?x)
   7                     https?://
   8                         (?:[^/]+\.)?vbox7\.com/
   9                         (?:
  10                             play:|
  11                             (?:
  12                                 emb/external\.php|
  13                                 player/ext\.swf
  14                             )\?.*?\bvid=
  15                         )
  16                         (?P<id>[\da-fA-F]+)
  17                     '''
  18     _EMBED_REGEX = [r'<iframe[^>]+src=(?P<q>["\'])(?P<url>(?:https?:)?//vbox7\.com/emb/external\.php.+?)(?P=q)']
  19     _GEO_COUNTRIES = ['BG']
  20     _TESTS = [{
  21         'url': 'http://vbox7.com/play:0946fff23c',
  22         'md5': 'a60f9ab3a3a2f013ef9a967d5f7be5bf',
  23         'info_dict': {
  24             'id': '0946fff23c',
  25             'ext': 'mp4',
  26             'title': 'Борисов: Притеснен съм за бъдещето на България',
  27             'description': 'По думите му е опасно страната ни да бъде обявена за "сигурна"',
  28             'thumbnail': r're:^https?://.*\.jpg$',
  29             'timestamp': 1470982814,
  30             'upload_date': '20160812',
  31             'uploader': 'zdraveibulgaria',
  32         },
  33         'params': {
  34             'proxy': '127.0.0.1:8118',
  35         },
  36     }, {
  37         'url': 'http://vbox7.com/play:249bb972c2',
  38         'md5': '99f65c0c9ef9b682b97313e052734c3f',
  39         'info_dict': {
  40             'id': '249bb972c2',
  41             'ext': 'mp4',
  42             'title': 'Смях! Чудо - чист за секунди - Скрита камера',
  43         },
  44         'skip': 'georestricted',
  45     }, {
  46         'url': 'http://vbox7.com/emb/external.php?vid=a240d20f9c&autoplay=1',
  47         'only_matching': True,
  48     }, {
  49         'url': 'http://i49.vbox7.com/player/ext.swf?vid=0946fff23c&autoplay=1',
  50         'only_matching': True,
  51     }]
  52
  53     def _real_extract(self, url):
  54         video_id = self._match_id(url)
  55
  56         response = self._download_json(
  57             'https://www.vbox7.com/ajax/video/nextvideo.php?vid=%s' % video_id,
  58             video_id)
  59
  60         if 'error' in response:
  61             raise ExtractorError(
  62                 '%s said: %s' % (self.IE_NAME, response['error']), expected=True)
  63
  64         video = response['options']
  65
  66         title = video['title']
  67         video_url = video['src']
  68
  69         if '/na.mp4' in video_url:
  70             self.raise_geo_restricted(countries=self._GEO_COUNTRIES)
  71
  72         uploader = video.get('uploader')
  73
  74         webpage = self._download_webpage(
  75             'http://vbox7.com/play:%s' % video_id, video_id, fatal=None)
  76
  77         info = {}
  78
  79         if webpage:
  80             info = self._search_json_ld(
  81                 webpage.replace('"/*@context"', '"@context"'), video_id,
  82                 fatal=False)
  83
  84         info.update({
  85             'id': video_id,
  86             'title': title,
  87             'url': video_url,
  88             'uploader': uploader,
  89             'thumbnail': self._proto_relative_url(
  90                 info.get('thumbnail') or self._og_search_thumbnail(webpage),
  91                 'http:'),
  92         })
  93         return info