# coding: utf-8
from __future__ import unicode_literals
+import re
+
from .common import InfoExtractor
+from ..utils import determine_ext
class Lecture2GoIE(InfoExtractor):
- _VALID_URL = r'https?://lecture2go.uni-hamburg.de/veranstaltungen/-/v/(?P<id>[0-9]+)'
+ _VALID_URL = r'https?://lecture2go\.uni-hamburg\.de/veranstaltungen/-/v/(?P<id>\d+)'
_TEST = {
'url': 'https://lecture2go.uni-hamburg.de/veranstaltungen/-/v/17473',
'md5': 'a9e76f83b3ef58019c4b7dbc35f406c1',
video_id = self._match_id(url)
webpage = self._download_webpage(url, video_id)
- title = self._html_search_regex(r'<em class="title">(.*?)</em>', webpage, 'title')
- video_url = self._search_regex(r'b.isFirefox..a.useHTML5\).b.setOption.a,"src","(.*.mp4)"\).else', webpage, 'video_url')
- creator = self._html_search_regex(r'<div id="description">(.*)</div>', webpage, 'creator')
+ title = self._html_search_regex(r'<em[^>]+class="title">(.+)</em>', webpage, 'title')
+
+ formats = []
+ for url in set(re.findall(r'"src","([^"]+)"', webpage)):
+ ext = determine_ext(url)
+ if ext == 'f4m':
+ formats.extend(self._extract_f4m_formats(url, video_id))
+ elif ext == 'm3u8':
+ formats.extend(self._extract_m3u8_formats(url, video_id))
+ else:
+ formats.append({
+ 'url': url,
+ })
+
+ self._sort_formats(formats)
+
+ creator = self._html_search_regex(r'<div[^>]+id="description">([^<]+)</div>', webpage, 'creator')
return {
'id': video_id,
'title': title,
- 'url': video_url,
+ 'formats': formats,
'creator': creator
}