[ustream] Support HLS streams (closes #10698)
[youtube-dl] / youtube_dl / extractor / ustream.py
1 from __future__ import unicode_literals
2
3 import random
4 import re
5
6 from .common import InfoExtractor
7 from ..compat import (
8     compat_str,
9     compat_urlparse,
10 )
11 from ..utils import (
12     encode_data_uri,
13     ExtractorError,
14     int_or_none,
15     float_or_none,
16     mimetype2ext,
17     str_or_none,
18 )
19
20
21 class UstreamIE(InfoExtractor):
22     _VALID_URL = r'https?://(?:www\.)?ustream\.tv/(?P<type>recorded|embed|embed/recorded)/(?P<id>\d+)'
23     IE_NAME = 'ustream'
24     _TESTS = [{
25         'url': 'http://www.ustream.tv/recorded/20274954',
26         'md5': '088f151799e8f572f84eb62f17d73e5c',
27         'info_dict': {
28             'id': '20274954',
29             'ext': 'flv',
30             'title': 'Young Americans for Liberty February 7, 2012 2:28 AM',
31             'description': 'Young Americans for Liberty February 7, 2012 2:28 AM',
32             'timestamp': 1328577035,
33             'upload_date': '20120207',
34             'uploader': 'yaliberty',
35             'uploader_id': '6780869',
36         },
37     }, {
38         # From http://sportscanada.tv/canadagames/index.php/week2/figure-skating/444
39         # Title and uploader available only from params JSON
40         'url': 'http://www.ustream.tv/embed/recorded/59307601?ub=ff0000&lc=ff0000&oc=ffffff&uc=ffffff&v=3&wmode=direct',
41         'md5': '5a2abf40babeac9812ed20ae12d34e10',
42         'info_dict': {
43             'id': '59307601',
44             'ext': 'flv',
45             'title': '-CG11- Canada Games Figure Skating',
46             'uploader': 'sportscanadatv',
47         },
48         'skip': 'This Pro Broadcaster has chosen to remove this video from the ustream.tv site.',
49     }, {
50         'url': 'http://www.ustream.tv/embed/10299409',
51         'info_dict': {
52             'id': '10299409',
53         },
54         'playlist_count': 3,
55     }, {
56         'url': 'http://www.ustream.tv/recorded/91343263',
57         'info_dict': {
58             'id': '91343263',
59             'ext': 'mp4',
60             'title': 'GitHub Universe - General Session - Day 1',
61             'upload_date': '20160914',
62             'description': 'GitHub Universe - General Session - Day 1',
63             'timestamp': 1473872730,
64             'uploader': 'wa0dnskeqkr',
65             'uploader_id': '38977840',
66         },
67         'params': {
68             'skip_download': True,  # m3u8 download
69         },
70     }]
71
72     def _get_stream_info(self, url, video_id, app_id_ver, extra_note=None):
73         def num_to_hex(n):
74             return hex(n)[2:]
75
76         rnd = random.randrange
77
78         if not extra_note:
79             extra_note = ''
80
81         conn_info = self._download_json(
82             'http://r%d-1-%s-recorded-lp-live.ums.ustream.tv/1/ustream' % (rnd(1e8), video_id),
83             video_id, note='Downloading connection info' + extra_note,
84             query={
85                 'type': 'viewer',
86                 'appId': app_id_ver[0],
87                 'appVersion': app_id_ver[1],
88                 'rsid': '%s:%s' % (num_to_hex(rnd(1e8)), num_to_hex(rnd(1e8))),
89                 'rpin': '_rpin.%d' % rnd(1e15),
90                 'referrer': url,
91                 'media': video_id,
92                 'application': 'recorded',
93             })
94         host = conn_info[0]['args'][0]['host']
95         connection_id = conn_info[0]['args'][0]['connectionId']
96
97         return self._download_json(
98             'http://%s/1/ustream?connectionId=%s' % (host, connection_id),
99             video_id, note='Downloading stream info' + extra_note)
100
101     def _get_streams(self, url, video_id, app_id_ver):
102         # Sometimes the return dict does not have 'stream'
103         for trial_count in range(3):
104             stream_info = self._get_stream_info(
105                 url, video_id, app_id_ver,
106                 extra_note=' (try %d)' % (trial_count + 1) if trial_count > 0 else '')
107             if 'stream' in stream_info[0]['args'][0]:
108                 return stream_info[0]['args'][0]['stream']
109         return []
110
111     def _parse_segmented_mp4(self, dash_stream_info):
112         def resolve_dash_template(template, idx, chunk_hash):
113             return template.replace('%', compat_str(idx), 1).replace('%', chunk_hash)
114
115         formats = []
116         for stream in dash_stream_info['streams']:
117             # Use only one provider to avoid too many formats
118             provider = dash_stream_info['providers'][0]
119             fragments = [{
120                 'url': resolve_dash_template(
121                     provider['url'] + stream['initUrl'], 0, dash_stream_info['hashes']['0'])
122             }]
123             for idx in range(dash_stream_info['videoLength'] // dash_stream_info['chunkTime']):
124                 fragments.append({
125                     'url': resolve_dash_template(
126                         provider['url'] + stream['segmentUrl'], idx,
127                         dash_stream_info['hashes'][compat_str(idx // 10 * 10)])
128                 })
129             content_type = stream['contentType']
130             kind = content_type.split('/')[0]
131             f = {
132                 'format_id': '-'.join(filter(None, [
133                     'dash', kind, str_or_none(stream.get('bitrate'))])),
134                 'protocol': 'http_dash_segments',
135                 # TODO: generate a MPD doc for external players?
136                 'url': encode_data_uri(b'<MPD/>', 'text/xml'),
137                 'ext': mimetype2ext(content_type),
138                 'height': stream.get('height'),
139                 'width': stream.get('width'),
140                 'fragments': fragments,
141             }
142             if kind == 'video':
143                 f.update({
144                     'vcodec': stream.get('codec'),
145                     'acodec': 'none',
146                     'vbr': stream.get('bitrate'),
147                 })
148             else:
149                 f.update({
150                     'vcodec': 'none',
151                     'acodec': stream.get('codec'),
152                     'abr': stream.get('bitrate'),
153                 })
154             formats.append(f)
155         return formats
156
157     def _real_extract(self, url):
158         m = re.match(self._VALID_URL, url)
159         video_id = m.group('id')
160
161         # some sites use this embed format (see: https://github.com/rg3/youtube-dl/issues/2990)
162         if m.group('type') == 'embed/recorded':
163             video_id = m.group('id')
164             desktop_url = 'http://www.ustream.tv/recorded/' + video_id
165             return self.url_result(desktop_url, 'Ustream')
166         if m.group('type') == 'embed':
167             video_id = m.group('id')
168             webpage = self._download_webpage(url, video_id)
169             content_video_ids = self._parse_json(self._search_regex(
170                 r'ustream\.vars\.offAirContentVideoIds=([^;]+);', webpage,
171                 'content video IDs'), video_id)
172             return self.playlist_result(
173                 map(lambda u: self.url_result('http://www.ustream.tv/recorded/' + u, 'Ustream'), content_video_ids),
174                 video_id)
175
176         params = self._download_json(
177             'https://api.ustream.tv/videos/%s.json' % video_id, video_id)
178
179         error = params.get('error')
180         if error:
181             raise ExtractorError(
182                 '%s returned error: %s' % (self.IE_NAME, error), expected=True)
183
184         video = params['video']
185
186         title = video['title']
187         filesize = float_or_none(video.get('file_size'))
188
189         formats = [{
190             'id': video_id,
191             'url': video_url,
192             'ext': format_id,
193             'filesize': filesize,
194         } for format_id, video_url in video['media_urls'].items() if video_url]
195
196         if not formats:
197             hls_streams = self._get_streams(url, video_id, app_id_ver=(11, 2))
198             if hls_streams:
199                 # m3u8_native leads to intermittent ContentTooShortError
200                 formats.extend(self._extract_m3u8_formats(
201                     hls_streams[0]['url'], video_id, ext='mp4', m3u8_id='hls'))
202
203             '''
204             # DASH streams handling is incomplete as 'url' is missing
205             dash_streams = self._get_streams(url, video_id, app_id_ver=(3, 1))
206             if dash_streams:
207                 formats.extend(self._parse_segmented_mp4(dash_streams))
208             '''
209
210         self._sort_formats(formats)
211
212         description = video.get('description')
213         timestamp = int_or_none(video.get('created_at'))
214         duration = float_or_none(video.get('length'))
215         view_count = int_or_none(video.get('views'))
216
217         uploader = video.get('owner', {}).get('username')
218         uploader_id = video.get('owner', {}).get('id')
219
220         thumbnails = [{
221             'id': thumbnail_id,
222             'url': thumbnail_url,
223         } for thumbnail_id, thumbnail_url in video.get('thumbnail', {}).items()]
224
225         return {
226             'id': video_id,
227             'title': title,
228             'description': description,
229             'thumbnails': thumbnails,
230             'timestamp': timestamp,
231             'duration': duration,
232             'view_count': view_count,
233             'uploader': uploader,
234             'uploader_id': uploader_id,
235             'formats': formats,
236         }
237
238
239 class UstreamChannelIE(InfoExtractor):
240     _VALID_URL = r'https?://(?:www\.)?ustream\.tv/channel/(?P<slug>.+)'
241     IE_NAME = 'ustream:channel'
242     _TEST = {
243         'url': 'http://www.ustream.tv/channel/channeljapan',
244         'info_dict': {
245             'id': '10874166',
246         },
247         'playlist_mincount': 17,
248     }
249
250     def _real_extract(self, url):
251         m = re.match(self._VALID_URL, url)
252         display_id = m.group('slug')
253         webpage = self._download_webpage(url, display_id)
254         channel_id = self._html_search_meta('ustream:channel_id', webpage)
255
256         BASE = 'http://www.ustream.tv'
257         next_url = '/ajax/socialstream/videos/%s/1.json' % channel_id
258         video_ids = []
259         while next_url:
260             reply = self._download_json(
261                 compat_urlparse.urljoin(BASE, next_url), display_id,
262                 note='Downloading video information (next: %d)' % (len(video_ids) + 1))
263             video_ids.extend(re.findall(r'data-content-id="(\d.*)"', reply['data']))
264             next_url = reply['nextUrl']
265
266         entries = [
267             self.url_result('http://www.ustream.tv/recorded/' + vid, 'Ustream')
268             for vid in video_ids]
269         return {
270             '_type': 'playlist',
271             'id': channel_id,
272             'display_id': display_id,
273             'entries': entries,
274         }