-# coding: utf-8
from __future__ import unicode_literals
-import re
-
from .common import InfoExtractor
class TweakersIE(InfoExtractor):
- _VALID_URL = r'https?://tweakers\.net/video/(?P<id>[0-9]+).*'
+ _VALID_URL = r'https?://tweakers\.net/video/(?P<id>\d+)'
_TEST = {
'url': 'https://tweakers.net/video/9926/new-nintendo-3ds-xl-op-alle-fronten-beter.html',
- 'md5': 'f7f7f3027166a7f32f024b4ae6571ced',
+ 'md5': '3147e4ddad366f97476a93863e4557c8',
'info_dict': {
'id': '9926',
'ext': 'mp4',
- 'title': 'New-Nintendo-3Ds-Xl-Op-Alle-Fronten-Beter',
- # TODO more properties, either as:
- # * A value
- # * MD5 checksum; start the string with md5:
- # * A regular expression; start the string with re:
- # * Any Python type (for example int or float)
+ 'title': 'New Nintendo 3DS XL - Op alle fronten beter',
+ 'description': 'md5:f97324cc71e86e11c853f0763820e3ba',
+ 'thumbnail': 're:^https?://.*\.jpe?g$',
+ 'duration': 386,
}
}
def _real_extract(self, url):
- splitted_url = re.split('.html|/', url)
- del splitted_url[-1] # To remove extra '/' at the end
- video_id = splitted_url[4]
- title = splitted_url[5].title() # Retrieve title for URL and capitalize
- splitted_url[3] = splitted_url[3] + '/player' # Add /player to get the player page
- player_url = '/'.join(splitted_url) + '.html'
- player_page = self._download_webpage(player_url, video_id)
-
- return {
- 'id': video_id,
- 'ext': 'mp4',
- 'title': title,
- 'url': re.findall('http.*mp4', player_page)[0],
- 'player_url': player_url
- }
+ playlist_id = self._match_id(url)
+ entries = self._extract_xspf_playlist(
+ 'https://tweakers.net/video/s1playlist/%s/playlist.xspf' % playlist_id, playlist_id)
+ return self.playlist_result(entries, playlist_id)