1
0
mirror of https://github.com/l1ving/youtube-dl synced 2024-11-24 14:12:53 +08:00
youtube-dl/youtube_dl/extractor/malemotion.py
2015-07-17 23:41:14 +06:00

47 lines
1.4 KiB
Python

# coding: utf-8
from __future__ import unicode_literals
from .common import InfoExtractor
from ..compat import compat_urllib_parse_unquote
class MalemotionIE(InfoExtractor):
_VALID_URL = r'https?://malemotion\.com/video/(.+?)\.(?P<id>.+?)(#|$)'
_TEST = {
'url': 'http://malemotion.com/video/bete-de-concours.ltc',
'md5': '3013e53a0afbde2878bc39998c33e8a5',
'info_dict': {
'id': 'ltc',
'ext': 'mp4',
'title': 'Bête de Concours',
'age_limit': 18,
},
}
def _real_extract(self, url):
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
video_url = compat_urllib_parse_unquote(self._search_regex(
r'<source type="video/mp4" src="(.+?)"', webpage, 'video URL'))
video_title = self._html_search_regex(
r'<title>(.*?)</title', webpage, 'title')
video_thumbnail = self._search_regex(
r'<video .+?poster="(.+?)"', webpage, 'thumbnail', fatal=False)
formats = [{
'url': video_url,
'ext': 'mp4',
'format_id': 'mp4',
'preference': 1,
}]
self._sort_formats(formats)
return {
'id': video_id,
'formats': formats,
'title': video_title,
'thumbnail': video_thumbnail,
'age_limit': 18,
}