1 # -*- coding: utf-8 -*-
2 from __future__ import unicode_literals
6 from .common import InfoExtractor
10 compat_urllib_parse_urlparse,
18 class CeskaTelevizeIE(InfoExtractor):
19 _VALID_URL = r'https?://www\.ceskatelevize\.cz/(porady|ivysilani)/(.+/)?(?P<id>[^?#]+)'
23 'url': 'http://www.ceskatelevize.cz/ivysilani/ivysilani/10441294653-hyde-park-civilizace/214411058091220',
25 'id': '214411058091220',
27 'title': 'Hyde Park Civilizace',
28 'description': 'Věda a současná civilizace. Interaktivní pořad - prostor pro vaše otázky a komentáře',
29 'thumbnail': 're:^https?://.*\.jpg',
34 'skip_download': True,
38 'url': 'http://www.ceskatelevize.cz/ivysilani/10532695142-prvni-republika/bonus/14716-zpevacka-z-duparny-bobina',
42 'title': 'První republika: Zpěvačka z Dupárny Bobina',
43 'description': 'Sága mapující atmosféru první republiky od r. 1918 do r. 1945.',
44 'thumbnail': 're:^https?://.*\.jpg',
49 'skip_download': True,
54 def _real_extract(self, url):
55 url = url.replace('/porady/', '/ivysilani/').replace('/video/', '')
57 mobj = re.match(self._VALID_URL, url)
58 video_id = mobj.group('id')
60 webpage = self._download_webpage(url, video_id)
62 NOT_AVAILABLE_STRING = 'This content is not available at your territory due to limited copyright.'
63 if '%s</p>' % NOT_AVAILABLE_STRING in webpage:
64 raise ExtractorError(NOT_AVAILABLE_STRING, expected=True)
66 typ = self._html_search_regex(r'getPlaylistUrl\(\[\{"type":"(.+?)","id":".+?"\}\],', webpage, 'type')
67 episode_id = self._html_search_regex(r'getPlaylistUrl\(\[\{"type":".+?","id":"(.+?)"\}\],', webpage, 'episode_id')
70 'playlist[0][type]': typ,
71 'playlist[0][id]': episode_id,
72 'requestUrl': compat_urllib_parse_urlparse(url).path,
73 'requestSource': 'iVysilani',
76 req = compat_urllib_request.Request(
77 'http://www.ceskatelevize.cz/ivysilani/ajax/get-client-playlist',
78 data=compat_urllib_parse.urlencode(data))
80 req.add_header('Content-type', 'application/x-www-form-urlencoded')
81 req.add_header('x-addr', '127.0.0.1')
82 req.add_header('X-Requested-With', 'XMLHttpRequest')
83 req.add_header('Referer', url)
85 playlistpage = self._download_json(req, video_id)
87 playlist_url = playlistpage['url']
88 if playlist_url == 'error_region':
89 raise ExtractorError(NOT_AVAILABLE_STRING, expected=True)
91 req = compat_urllib_request.Request(compat_urllib_parse.unquote(playlist_url))
92 req.add_header('Referer', url)
94 playlist = self._download_json(req, video_id)
96 item = playlist['playlist'][0]
98 for format_id, stream_url in item['streamUrls'].items():
99 formats.extend(self._extract_m3u8_formats(stream_url, video_id, 'mp4'))
100 self._sort_formats(formats)
102 title = self._og_search_title(webpage)
103 description = self._og_search_description(webpage)
104 duration = float_or_none(item.get('duration'))
105 thumbnail = item.get('previewImageUrl')
110 'description': description,
111 'thumbnail': thumbnail,
112 'duration': duration,