video_id = self._match_id(url)
webpage = self._download_webpage(
- 'http://www.xvideos.com/video%s/' % video_id, video_id)
+ 'https://www.xvideos.com/video%s/' % video_id, video_id)
mobj = re.search(r'<h1 class="inlineError">(.+?)</h1>', webpage)
if mobj:
webpage, 'title', default=None,
group='title') or self._og_search_title(webpage)
- thumbnail = self._search_regex(
- (r'setThumbUrl\(\s*(["\'])(?P<thumbnail>(?:(?!\1).)+)\1',
- r'url_bigthumb=(?P<thumbnail>.+?)&'),
- webpage, 'thumbnail', fatal=False, group='thumbnail')
+ thumbnails = []
+ for preference, thumbnail in enumerate(('', '169')):
+ thumbnail_url = self._search_regex(
+ r'setThumbUrl%s\(\s*(["\'])(?P<thumbnail>(?:(?!\1).)+)\1' % thumbnail,
+ webpage, 'thumbnail', default=None, group='thumbnail')
+ if thumbnail_url:
+ thumbnails.append({
+ 'url': thumbnail_url,
+ 'preference': preference,
+ })
+
duration = int_or_none(self._og_search_property(
'duration', webpage, default=None)) or parse_duration(
self._search_regex(
'formats': formats,
'title': title,
'duration': duration,
- 'thumbnail': thumbnail,
+ 'thumbnails': thumbnails,
'age_limit': 18,
}