Merge pull request #9288 from reyyed/issue#9063fix
[youtube-dl] / youtube_dl / extractor / kaltura.py
1 # coding: utf-8
2 from __future__ import unicode_literals
3
4 import re
5 import base64
6
7 from .common import InfoExtractor
8 from ..compat import (
9     compat_urlparse,
10     compat_parse_qs,
11 )
12 from ..utils import (
13     clean_html,
14     ExtractorError,
15     int_or_none,
16     unsmuggle_url,
17     smuggle_url,
18 )
19
20
21 class KalturaIE(InfoExtractor):
22     _VALID_URL = r'''(?x)
23                 (?:
24                     kaltura:(?P<partner_id>\d+):(?P<id>[0-9a-z_]+)|
25                     https?://
26                         (:?(?:www|cdnapi(?:sec)?)\.)?kaltura\.com/
27                         (?:
28                             (?:
29                                 # flash player
30                                 index\.php/kwidget|
31                                 # html5 player
32                                 html5/html5lib/[^/]+/mwEmbedFrame\.php
33                             )
34                         )(?:/(?P<path>[^?]+))?(?:\?(?P<query>.*))?
35                 )
36                 '''
37     _SERVICE_URL = 'http://cdnapi.kaltura.com'
38     _SERVICE_BASE = '/api_v3/index.php'
39     _TESTS = [
40         {
41             'url': 'kaltura:269692:1_1jc2y3e4',
42             'md5': '3adcbdb3dcc02d647539e53f284ba171',
43             'info_dict': {
44                 'id': '1_1jc2y3e4',
45                 'ext': 'mp4',
46                 'title': 'Straight from the Heart',
47                 'upload_date': '20131219',
48                 'uploader_id': 'mlundberg@wolfgangsvault.com',
49                 'description': 'The Allman Brothers Band, 12/16/1981',
50                 'thumbnail': 're:^https?://.*/thumbnail/.*',
51                 'timestamp': int,
52             },
53         },
54         {
55             'url': 'http://www.kaltura.com/index.php/kwidget/cache_st/1300318621/wid/_269692/uiconf_id/3873291/entry_id/1_1jc2y3e4',
56             'only_matching': True,
57         },
58         {
59             'url': 'https://cdnapisec.kaltura.com/index.php/kwidget/wid/_557781/uiconf_id/22845202/entry_id/1_plr1syf3',
60             'only_matching': True,
61         },
62         {
63             'url': 'https://cdnapisec.kaltura.com/html5/html5lib/v2.30.2/mwEmbedFrame.php/p/1337/uiconf_id/20540612/entry_id/1_sf5ovm7u?wid=_243342',
64             'only_matching': True,
65         }
66     ]
67
68     @staticmethod
69     def _extract_url(webpage):
70         mobj = (
71             re.search(
72                 r"""(?xs)
73                     kWidget\.(?:thumb)?[Ee]mbed\(
74                     \{.*?
75                         (?P<q1>['\"])wid(?P=q1)\s*:\s*
76                         (?P<q2>['\"])_?(?P<partner_id>[^'\"]+)(?P=q2),.*?
77                         (?P<q3>['\"])entry_?[Ii]d(?P=q3)\s*:\s*
78                         (?P<q4>['\"])(?P<id>[^'\"]+)(?P=q4),
79                 """, webpage) or
80             re.search(
81                 r'''(?xs)
82                     (?P<q1>["\'])
83                         (?:https?:)?//cdnapi(?:sec)?\.kaltura\.com/.*?(?:p|partner_id)/(?P<partner_id>\d+).*?
84                     (?P=q1).*?
85                     (?:
86                         entry_?[Ii]d|
87                         (?P<q2>["\'])entry_?[Ii]d(?P=q2)
88                     )\s*:\s*
89                     (?P<q3>["\'])(?P<id>.+?)(?P=q3)
90                 ''', webpage))
91         if mobj:
92             embed_info = mobj.groupdict()
93             url = 'kaltura:%(partner_id)s:%(id)s' % embed_info
94             escaped_pid = re.escape(embed_info['partner_id'])
95             service_url = re.search(
96                 r'<script[^>]+src=["\']((?:https?:)?//.+?)/p/%s/sp/%s00/embedIframeJs' % (escaped_pid, escaped_pid),
97                 webpage)
98             if service_url:
99                 url = smuggle_url(url, {'service_url': service_url.group(1)})
100             return url
101
102     def _kaltura_api_call(self, video_id, actions, service_url=None, *args, **kwargs):
103         params = actions[0]
104         if len(actions) > 1:
105             for i, a in enumerate(actions[1:], start=1):
106                 for k, v in a.items():
107                     params['%d:%s' % (i, k)] = v
108
109         data = self._download_json(
110             (service_url or self._SERVICE_URL) + self._SERVICE_BASE,
111             video_id, query=params, *args, **kwargs)
112
113         status = data if len(actions) == 1 else data[0]
114         if status.get('objectType') == 'KalturaAPIException':
115             raise ExtractorError(
116                 '%s said: %s' % (self.IE_NAME, status['message']))
117
118         return data
119
120     def _get_kaltura_signature(self, video_id, partner_id, service_url=None):
121         actions = [{
122             'apiVersion': '3.1',
123             'expiry': 86400,
124             'format': 1,
125             'service': 'session',
126             'action': 'startWidgetSession',
127             'widgetId': '_%s' % partner_id,
128         }]
129         return self._kaltura_api_call(
130             video_id, actions, service_url, note='Downloading Kaltura signature')['ks']
131
132     def _get_video_info(self, video_id, partner_id, service_url=None):
133         signature = self._get_kaltura_signature(video_id, partner_id, service_url)
134         actions = [
135             {
136                 'action': 'null',
137                 'apiVersion': '3.1.5',
138                 'clientTag': 'kdp:v3.8.5',
139                 'format': 1,  # JSON, 2 = XML, 3 = PHP
140                 'service': 'multirequest',
141                 'ks': signature,
142             },
143             {
144                 'action': 'get',
145                 'entryId': video_id,
146                 'service': 'baseentry',
147                 'version': '-1',
148             },
149             {
150                 'action': 'getbyentryid',
151                 'entryId': video_id,
152                 'service': 'flavorAsset',
153             },
154         ]
155         return self._kaltura_api_call(
156             video_id, actions, service_url, note='Downloading video info JSON')
157
158     def _real_extract(self, url):
159         url, smuggled_data = unsmuggle_url(url, {})
160
161         mobj = re.match(self._VALID_URL, url)
162         partner_id, entry_id = mobj.group('partner_id', 'id')
163         ks = None
164         if partner_id and entry_id:
165             info, flavor_assets = self._get_video_info(entry_id, partner_id, smuggled_data.get('service_url'))
166         else:
167             path, query = mobj.group('path', 'query')
168             if not path and not query:
169                 raise ExtractorError('Invalid URL', expected=True)
170             params = {}
171             if query:
172                 params = compat_parse_qs(query)
173             if path:
174                 splitted_path = path.split('/')
175                 params.update(dict((zip(splitted_path[::2], [[v] for v in splitted_path[1::2]]))))
176             if 'wid' in params:
177                 partner_id = params['wid'][0][1:]
178             elif 'p' in params:
179                 partner_id = params['p'][0]
180             else:
181                 raise ExtractorError('Invalid URL', expected=True)
182             if 'entry_id' in params:
183                 entry_id = params['entry_id'][0]
184                 info, flavor_assets = self._get_video_info(entry_id, partner_id)
185             elif 'uiconf_id' in params and 'flashvars[referenceId]' in params:
186                 reference_id = params['flashvars[referenceId]'][0]
187                 webpage = self._download_webpage(url, reference_id)
188                 entry_data = self._parse_json(self._search_regex(
189                     r'window\.kalturaIframePackageData\s*=\s*({.*});',
190                     webpage, 'kalturaIframePackageData'),
191                     reference_id)['entryResult']
192                 info, flavor_assets = entry_data['meta'], entry_data['contextData']['flavorAssets']
193                 entry_id = info['id']
194             else:
195                 raise ExtractorError('Invalid URL', expected=True)
196             ks = params.get('flashvars[ks]', [None])[0]
197
198         source_url = smuggled_data.get('source_url')
199         if source_url:
200             referrer = base64.b64encode(
201                 '://'.join(compat_urlparse.urlparse(source_url)[:2])
202                 .encode('utf-8')).decode('utf-8')
203         else:
204             referrer = None
205
206         def sign_url(unsigned_url):
207             if ks:
208                 unsigned_url += '/ks/%s' % ks
209             if referrer:
210                 unsigned_url += '?referrer=%s' % referrer
211             return unsigned_url
212
213         data_url = info['dataUrl']
214         if '/flvclipper/' in data_url:
215             data_url = re.sub(r'/flvclipper/.*', '/serveFlavor', data_url)
216
217         formats = []
218         for f in flavor_assets:
219             # Continue if asset is not ready
220             if f['status'] != 2:
221                 continue
222             video_url = sign_url(
223                 '%s/flavorId/%s' % (data_url, f['id']))
224             formats.append({
225                 'format_id': '%(fileExt)s-%(bitrate)s' % f,
226                 'ext': f.get('fileExt'),
227                 'tbr': int_or_none(f['bitrate']),
228                 'fps': int_or_none(f.get('frameRate')),
229                 'filesize_approx': int_or_none(f.get('size'), invscale=1024),
230                 'container': f.get('containerFormat'),
231                 'vcodec': f.get('videoCodecId'),
232                 'height': int_or_none(f.get('height')),
233                 'width': int_or_none(f.get('width')),
234                 'url': video_url,
235             })
236         if '/playManifest/' in data_url:
237             m3u8_url = sign_url(data_url.replace(
238                 'format/url', 'format/applehttp'))
239             formats.extend(self._extract_m3u8_formats(
240                 m3u8_url, entry_id, 'mp4', 'm3u8_native',
241                 m3u8_id='hls', fatal=False))
242
243         self._check_formats(formats, entry_id)
244         self._sort_formats(formats)
245
246         return {
247             'id': entry_id,
248             'title': info['name'],
249             'formats': formats,
250             'description': clean_html(info.get('description')),
251             'thumbnail': info.get('thumbnailUrl'),
252             'duration': info.get('duration'),
253             'timestamp': info.get('createdAt'),
254             'uploader_id': info.get('userId'),
255             'view_count': info.get('plays'),
256         }