mirror of
https://github.com/ytdl-org/youtube-dl
synced 2024-11-14 22:37:59 +09:00
41 lines
1.3 KiB
Python
41 lines
1.3 KiB
Python
from __future__ import unicode_literals
|
|
|
|
import re
|
|
|
|
from .common import InfoExtractor
|
|
|
|
|
|
class KeekIE(InfoExtractor):
|
|
_VALID_URL = r'https?://(?:www\.)?keek\.com/(?:!|\w+/keeks/)(?P<videoID>\w+)'
|
|
IE_NAME = 'keek'
|
|
_TEST = {
|
|
'url': 'https://www.keek.com/ytdl/keeks/NODfbab',
|
|
'file': 'NODfbab.mp4',
|
|
'md5': '9b0636f8c0f7614afa4ea5e4c6e57e83',
|
|
'info_dict': {
|
|
'uploader': 'ytdl',
|
|
'title': 'test chars: "\'/\\\u00e4<>This is a test video for youtube-dl.For more information, contact phihag@phihag.de .',
|
|
},
|
|
}
|
|
|
|
def _real_extract(self, url):
|
|
m = re.match(self._VALID_URL, url)
|
|
video_id = m.group('videoID')
|
|
|
|
video_url = 'http://cdn.keek.com/keek/video/%s' % video_id
|
|
thumbnail = 'http://cdn.keek.com/keek/thumbnail/%s/w100/h75' % video_id
|
|
webpage = self._download_webpage(url, video_id)
|
|
|
|
uploader = self._html_search_regex(
|
|
r'<div class="user-name-and-bio">[\S\s]+?<h2>(?P<uploader>.+?)</h2>',
|
|
webpage, 'uploader', fatal=False)
|
|
|
|
return {
|
|
'id': video_id,
|
|
'url': video_url,
|
|
'ext': 'mp4',
|
|
'title': self._og_search_title(webpage),
|
|
'thumbnail': thumbnail,
|
|
'uploader': uploader
|
|
}
|