Raphaël G. Git Repositories - youtubedl/blob - youtube_dl/extractor/vbox7.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5
   6 from .common import InfoExtractor
   7 from ..utils import ExtractorError
   8
   9
  10 class Vbox7IE(InfoExtractor):
  11     _VALID_URL = r'''(?x)
  12                     https?://
  13                         (?:[^/]+\.)?vbox7\.com/
  14                         (?:
  15                             play:|
  16                             (?:
  17                                 emb/external\.php|
  18                                 player/ext\.swf
  19                             )\?.*?\bvid=
  20                         )
  21                         (?P<id>[\da-fA-F]+)
  22                     '''
  23     _TESTS = [{
  24         'url': 'http://vbox7.com/play:0946fff23c',
  25         'md5': 'a60f9ab3a3a2f013ef9a967d5f7be5bf',
  26         'info_dict': {
  27             'id': '0946fff23c',
  28             'ext': 'mp4',
  29             'title': 'Борисов: Притеснен съм за бъдещето на България',
  30             'description': 'По думите му е опасно страната ни да бъде обявена за "сигурна"',
  31             'thumbnail': r're:^https?://.*\.jpg$',
  32             'timestamp': 1470982814,
  33             'upload_date': '20160812',
  34             'uploader': 'zdraveibulgaria',
  35         },
  36         'params': {
  37             'proxy': '127.0.0.1:8118',
  38         },
  39     }, {
  40         'url': 'http://vbox7.com/play:249bb972c2',
  41         'md5': '99f65c0c9ef9b682b97313e052734c3f',
  42         'info_dict': {
  43             'id': '249bb972c2',
  44             'ext': 'mp4',
  45             'title': 'Смях! Чудо - чист за секунди - Скрита камера',
  46         },
  47         'skip': 'georestricted',
  48     }, {
  49         'url': 'http://vbox7.com/emb/external.php?vid=a240d20f9c&autoplay=1',
  50         'only_matching': True,
  51     }, {
  52         'url': 'http://i49.vbox7.com/player/ext.swf?vid=0946fff23c&autoplay=1',
  53         'only_matching': True,
  54     }]
  55
  56     @staticmethod
  57     def _extract_url(webpage):
  58         mobj = re.search(
  59             r'<iframe[^>]+src=(?P<q>["\'])(?P<url>(?:https?:)?//vbox7\.com/emb/external\.php.+?)(?P=q)',
  60             webpage)
  61         if mobj:
  62             return mobj.group('url')
  63
  64     def _real_extract(self, url):
  65         video_id = self._match_id(url)
  66
  67         response = self._download_json(
  68             'https://www.vbox7.com/ajax/video/nextvideo.php?vid=%s' % video_id,
  69             video_id)
  70
  71         if 'error' in response:
  72             raise ExtractorError(
  73                 '%s said: %s' % (self.IE_NAME, response['error']), expected=True)
  74
  75         video = response['options']
  76
  77         title = video['title']
  78         video_url = video['src']
  79
  80         if '/na.mp4' in video_url:
  81             self.raise_geo_restricted()
  82
  83         uploader = video.get('uploader')
  84
  85         webpage = self._download_webpage(
  86             'http://vbox7.com/play:%s' % video_id, video_id, fatal=None)
  87
  88         info = {}
  89
  90         if webpage:
  91             info = self._search_json_ld(
  92                 webpage.replace('"/*@context"', '"@context"'), video_id,
  93                 fatal=False)
  94
  95         info.update({
  96             'id': video_id,
  97             'title': title,
  98             'url': video_url,
  99             'uploader': uploader,
 100             'thumbnail': self._proto_relative_url(
 101                 info.get('thumbnail') or self._og_search_thumbnail(webpage),
 102                 'http:'),
 103         })
 104         return info