|
@ -16,7 +16,7 @@ from ..aes import aes_decrypt_text |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
class SpankwireIE(InfoExtractor): |
|
|
class SpankwireIE(InfoExtractor): |
|
|
_VALID_URL = r'https?://(?:www\.)?(?P<url>spankwire\.com/[^/]*/video(?P<videoid>[0-9]+)/?)' |
|
|
|
|
|
|
|
|
_VALID_URL = r'https?://(?:www\.)?(?P<url>spankwire\.com/[^/]*/video(?P<id>[0-9]+)/?)' |
|
|
_TESTS = [{ |
|
|
_TESTS = [{ |
|
|
# download URL pattern: */<height>P_<tbr>K_<video_id>.mp4 |
|
|
# download URL pattern: */<height>P_<tbr>K_<video_id>.mp4 |
|
|
'url': 'http://www.spankwire.com/Buckcherry-s-X-Rated-Music-Video-Crazy-Bitch/video103545/', |
|
|
'url': 'http://www.spankwire.com/Buckcherry-s-X-Rated-Music-Video-Crazy-Bitch/video103545/', |
|
@ -49,10 +49,9 @@ class SpankwireIE(InfoExtractor): |
|
|
|
|
|
|
|
|
def _real_extract(self, url): |
|
|
def _real_extract(self, url): |
|
|
mobj = re.match(self._VALID_URL, url) |
|
|
mobj = re.match(self._VALID_URL, url) |
|
|
video_id = mobj.group('videoid') |
|
|
|
|
|
url = 'http://www.' + mobj.group('url') |
|
|
|
|
|
|
|
|
video_id = mobj.group('id') |
|
|
|
|
|
|
|
|
req = compat_urllib_request.Request(url) |
|
|
|
|
|
|
|
|
req = compat_urllib_request.Request('http://www.' + mobj.group('url')) |
|
|
req.add_header('Cookie', 'age_verified=1') |
|
|
req.add_header('Cookie', 'age_verified=1') |
|
|
webpage = self._download_webpage(req, video_id) |
|
|
webpage = self._download_webpage(req, video_id) |
|
|
|
|
|
|
|
|