[tv2:article] Fix extraction (Closes #10188)
[youtube-dl] / youtube_dl / extractor / tv2.py
1 # encoding: utf-8
2 from __future__ import unicode_literals
3
4 import re
5
6 from .common import InfoExtractor
7 from ..utils import (
8     determine_ext,
9     int_or_none,
10     float_or_none,
11     js_to_json,
12     parse_iso8601,
13     remove_end,
14 )
15
16
17 class TV2IE(InfoExtractor):
18     _VALID_URL = r'https?://(?:www\.)?tv2\.no/v/(?P<id>\d+)'
19     _TEST = {
20         'url': 'http://www.tv2.no/v/916509/',
21         'info_dict': {
22             'id': '916509',
23             'ext': 'mp4',
24             'title': 'Se Frode Gryttens hyllest av Steven Gerrard',
25             'description': 'TV 2 Sportens huspoet tar avskjed med Liverpools kaptein Steven Gerrard.',
26             'timestamp': 1431715610,
27             'upload_date': '20150515',
28             'duration': 156.967,
29             'view_count': int,
30             'categories': list,
31         },
32         'params': {
33             # m3u8 download
34             'skip_download': True,
35         },
36     }
37
38     def _real_extract(self, url):
39         video_id = self._match_id(url)
40
41         formats = []
42         format_urls = []
43         for protocol in ('HDS', 'HLS'):
44             data = self._download_json(
45                 'http://sumo.tv2.no/api/web/asset/%s/play.json?protocol=%s&videoFormat=SMIL+ISMUSP' % (video_id, protocol),
46                 video_id, 'Downloading play JSON')['playback']
47             for item in data['items']['item']:
48                 video_url = item.get('url')
49                 if not video_url or video_url in format_urls:
50                     continue
51                 format_id = '%s-%s' % (protocol.lower(), item.get('mediaFormat'))
52                 if not self._is_valid_url(video_url, video_id, format_id):
53                     continue
54                 format_urls.append(video_url)
55                 ext = determine_ext(video_url)
56                 if ext == 'f4m':
57                     formats.extend(self._extract_f4m_formats(
58                         video_url, video_id, f4m_id=format_id))
59                 elif ext == 'm3u8':
60                     formats.extend(self._extract_m3u8_formats(
61                         video_url, video_id, 'mp4', m3u8_id=format_id))
62                 elif ext == 'ism' or video_url.endswith('.ism/Manifest'):
63                     pass
64                 else:
65                     formats.append({
66                         'url': video_url,
67                         'format_id': format_id,
68                         'tbr': int_or_none(item.get('bitrate')),
69                         'filesize': int_or_none(item.get('fileSize')),
70                     })
71         self._sort_formats(formats)
72
73         asset = self._download_json(
74             'http://sumo.tv2.no/api/web/asset/%s.json' % video_id,
75             video_id, 'Downloading metadata JSON')['asset']
76
77         title = asset['title']
78         description = asset.get('description')
79         timestamp = parse_iso8601(asset.get('createTime'))
80         duration = float_or_none(asset.get('accurateDuration') or asset.get('duration'))
81         view_count = int_or_none(asset.get('views'))
82         categories = asset.get('keywords', '').split(',')
83
84         thumbnails = [{
85             'id': thumbnail.get('@type'),
86             'url': thumbnail.get('url'),
87         } for _, thumbnail in asset.get('imageVersions', {}).items()]
88
89         return {
90             'id': video_id,
91             'url': video_url,
92             'title': title,
93             'description': description,
94             'thumbnails': thumbnails,
95             'timestamp': timestamp,
96             'duration': duration,
97             'view_count': view_count,
98             'categories': categories,
99             'formats': formats,
100         }
101
102
103 class TV2ArticleIE(InfoExtractor):
104     _VALID_URL = r'https?://(?:www\.)?tv2\.no/(?:a|\d{4}/\d{2}/\d{2}(/[^/]+)+)/(?P<id>\d+)'
105     _TESTS = [{
106         'url': 'http://www.tv2.no/2015/05/16/nyheter/alesund/krim/pingvin/6930542',
107         'info_dict': {
108             'id': '6930542',
109             'title': 'Russen hetses etter pingvintyveri - innrømmer å ha åpnet luken på buret',
110             'description': 'md5:339573779d3eea3542ffe12006190954',
111         },
112         'playlist_count': 2,
113     }, {
114         'url': 'http://www.tv2.no/a/6930542',
115         'only_matching': True,
116     }]
117
118     def _real_extract(self, url):
119         playlist_id = self._match_id(url)
120
121         webpage = self._download_webpage(url, playlist_id)
122
123         # Old embed pattern (looks unused nowadays)
124         assets = re.findall(r'data-assetid=["\'](\d+)', webpage)
125
126         if not assets:
127             # New embed pattern
128             for v in re.findall('TV2ContentboxVideo\(({.+?})\)', webpage):
129                 video = self._parse_json(
130                     v, playlist_id, transform_source=js_to_json, fatal=False)
131                 if not video:
132                     continue
133                 asset = video.get('assetId')
134                 if asset:
135                     assets.append(asset)
136
137         entries = [
138             self.url_result('http://www.tv2.no/v/%s' % asset_id, 'TV2')
139             for asset_id in assets]
140
141         title = remove_end(self._og_search_title(webpage), ' - TV2.no')
142         description = remove_end(self._og_search_description(webpage), ' - TV2.no')
143
144         return self.playlist_result(entries, playlist_id, title, description)