]> jfr.im git - yt-dlp.git/blob - yt_dlp/extractor/mojvideo.py
[npr] Make SMIL extraction non-fatal (#2099)
[yt-dlp.git] / yt_dlp / extractor / mojvideo.py
1 # coding: utf-8
2 from __future__ import unicode_literals
3
4
5 from .common import InfoExtractor
6 from ..utils import (
7 ExtractorError,
8 parse_duration,
9 )
10
11
12 class MojvideoIE(InfoExtractor):
13 _VALID_URL = r'https?://(?:www\.)?mojvideo\.com/video-(?P<display_id>[^/]+)/(?P<id>[a-f0-9]+)'
14 _TEST = {
15 'url': 'http://www.mojvideo.com/video-v-avtu-pred-mano-rdecelaska-alfi-nipic/3d1ed4497707730b2906',
16 'md5': 'f7fd662cc8ce2be107b0d4f2c0483ae7',
17 'info_dict': {
18 'id': '3d1ed4497707730b2906',
19 'display_id': 'v-avtu-pred-mano-rdecelaska-alfi-nipic',
20 'ext': 'mp4',
21 'title': 'V avtu pred mano rdečelaska - Alfi Nipič',
22 'thumbnail': r're:^http://.*\.jpg$',
23 'duration': 242,
24 }
25 }
26
27 def _real_extract(self, url):
28 mobj = self._match_valid_url(url)
29 video_id = mobj.group('id')
30 display_id = mobj.group('display_id')
31
32 # XML is malformed
33 playerapi = self._download_webpage(
34 'http://www.mojvideo.com/playerapi.php?v=%s&t=1' % video_id, display_id)
35
36 if '<error>true</error>' in playerapi:
37 error_desc = self._html_search_regex(
38 r'<errordesc>([^<]*)</errordesc>', playerapi, 'error description', fatal=False)
39 raise ExtractorError('%s said: %s' % (self.IE_NAME, error_desc), expected=True)
40
41 title = self._html_search_regex(
42 r'<title>([^<]+)</title>', playerapi, 'title')
43 video_url = self._html_search_regex(
44 r'<file>([^<]+)</file>', playerapi, 'video URL')
45 thumbnail = self._html_search_regex(
46 r'<preview>([^<]+)</preview>', playerapi, 'thumbnail', fatal=False)
47 duration = parse_duration(self._html_search_regex(
48 r'<duration>([^<]+)</duration>', playerapi, 'duration', fatal=False))
49
50 return {
51 'id': video_id,
52 'display_id': display_id,
53 'url': video_url,
54 'title': title,
55 'thumbnail': thumbnail,
56 'duration': duration,
57 }