]> jfr.im git - yt-dlp.git/blob - yt_dlp/extractor/mojvideo.py
[ie/orf:on] Improve extraction (#9677)
[yt-dlp.git] / yt_dlp / extractor / mojvideo.py
1 from .common import InfoExtractor
2 from ..utils import (
3 ExtractorError,
4 parse_duration,
5 )
6
7
8 class MojvideoIE(InfoExtractor):
9 _VALID_URL = r'https?://(?:www\.)?mojvideo\.com/video-(?P<display_id>[^/]+)/(?P<id>[a-f0-9]+)'
10 _TEST = {
11 'url': 'http://www.mojvideo.com/video-v-avtu-pred-mano-rdecelaska-alfi-nipic/3d1ed4497707730b2906',
12 'md5': 'f7fd662cc8ce2be107b0d4f2c0483ae7',
13 'info_dict': {
14 'id': '3d1ed4497707730b2906',
15 'display_id': 'v-avtu-pred-mano-rdecelaska-alfi-nipic',
16 'ext': 'mp4',
17 'title': 'V avtu pred mano rdečelaska - Alfi Nipič',
18 'thumbnail': r're:^http://.*\.jpg$',
19 'duration': 242,
20 }
21 }
22
23 def _real_extract(self, url):
24 mobj = self._match_valid_url(url)
25 video_id = mobj.group('id')
26 display_id = mobj.group('display_id')
27
28 # XML is malformed
29 playerapi = self._download_webpage(
30 'http://www.mojvideo.com/playerapi.php?v=%s&t=1' % video_id, display_id)
31
32 if '<error>true</error>' in playerapi:
33 error_desc = self._html_search_regex(
34 r'<errordesc>([^<]*)</errordesc>', playerapi, 'error description', fatal=False)
35 raise ExtractorError('%s said: %s' % (self.IE_NAME, error_desc), expected=True)
36
37 title = self._html_extract_title(playerapi)
38 video_url = self._html_search_regex(
39 r'<file>([^<]+)</file>', playerapi, 'video URL')
40 thumbnail = self._html_search_regex(
41 r'<preview>([^<]+)</preview>', playerapi, 'thumbnail', fatal=False)
42 duration = parse_duration(self._html_search_regex(
43 r'<duration>([^<]+)</duration>', playerapi, 'duration', fatal=False))
44
45 return {
46 'id': video_id,
47 'display_id': display_id,
48 'url': video_url,
49 'title': title,
50 'thumbnail': thumbnail,
51 'duration': duration,
52 }