[nexx] Add support for shortcuts and relax domain id extraction
[ytdl] / youtube_dl / extractor / nexx.py
1 # coding: utf-8
2 from __future__ import unicode_literals
3
4 import hashlib
5 import random
6 import re
7 import time
8
9 from .common import InfoExtractor
10 from ..compat import compat_str
11 from ..utils import (
12     ExtractorError,
13     int_or_none,
14     parse_duration,
15     try_get,
16     urlencode_postdata,
17 )
18
19
20 class NexxIE(InfoExtractor):
21     _VALID_URL = r'''(?x)
22                         (?:
23                             https?://api\.nexx(?:\.cloud|cdn\.com)/v3/(?P<domain_id>\d+)/videos/byid/|
24                             nexx:(?P<domain_id_s>\d+):
25                         )
26                         (?P<id>\d+)
27                     '''
28     _TESTS = [{
29         # movie
30         'url': 'https://api.nexx.cloud/v3/748/videos/byid/128907',
31         'md5': '16746bfc28c42049492385c989b26c4a',
32         'info_dict': {
33             'id': '128907',
34             'ext': 'mp4',
35             'title': 'Stiftung Warentest',
36             'alt_title': 'Wie ein Test abläuft',
37             'description': 'md5:d1ddb1ef63de721132abd38639cc2fd2',
38             'release_year': 2013,
39             'creator': 'SPIEGEL TV',
40             'thumbnail': r're:^https?://.*\.jpg$',
41             'duration': 2509,
42             'timestamp': 1384264416,
43             'upload_date': '20131112',
44         },
45         'params': {
46             'format': 'bestvideo',
47         },
48     }, {
49         # episode
50         'url': 'https://api.nexx.cloud/v3/741/videos/byid/247858',
51         'info_dict': {
52             'id': '247858',
53             'ext': 'mp4',
54             'title': 'Return of the Golden Child (OV)',
55             'description': 'md5:5d969537509a92b733de21bae249dc63',
56             'release_year': 2017,
57             'thumbnail': r're:^https?://.*\.jpg$',
58             'duration': 1397,
59             'timestamp': 1495033267,
60             'upload_date': '20170517',
61             'episode_number': 2,
62             'season_number': 2,
63         },
64         'params': {
65             'format': 'bestvideo',
66             'skip_download': True,
67         },
68     }, {
69         'url': 'https://api.nexxcdn.com/v3/748/videos/byid/128907',
70         'only_matching': True,
71     }, {
72         'url': 'nexx:748:128907',
73         'only_matching': True,
74     }]
75
76     @staticmethod
77     def _extract_domain_id(webpage):
78         mobj = re.search(
79             r'<script\b[^>]+\bsrc=["\'](?:https?:)?//require\.nexx(?:\.cloud|cdn\.com)/(?P<id>\d+)',
80             webpage)
81         return mobj.group('id') if mobj else None
82
83     @staticmethod
84     def _extract_urls(webpage):
85         # Reference:
86         # 1. https://nx-s.akamaized.net/files/201510/44.pdf
87
88         entries = []
89
90         # JavaScript Integration
91         domain_id = NexxIE._extract_domain_id(webpage)
92         if domain_id:
93             for video_id in re.findall(
94                     r'(?is)onPLAYReady.+?_play\.init\s*\(.+?\s*,\s*["\']?(\d+)',
95                     webpage):
96                 entries.append(
97                     'https://api.nexx.cloud/v3/%s/videos/byid/%s'
98                     % (domain_id, video_id))
99
100         # TODO: support more embed formats
101
102         return entries
103
104     @staticmethod
105     def _extract_url(webpage):
106         return NexxIE._extract_urls(webpage)[0]
107
108     def _handle_error(self, response):
109         status = int_or_none(try_get(
110             response, lambda x: x['metadata']['status']) or 200)
111         if 200 <= status < 300:
112             return
113         raise ExtractorError(
114             '%s said: %s' % (self.IE_NAME, response['metadata']['errorhint']),
115             expected=True)
116
117     def _call_api(self, domain_id, path, video_id, data=None, headers={}):
118         headers['Content-Type'] = 'application/x-www-form-urlencoded; charset=UTF-8'
119         result = self._download_json(
120             'https://api.nexx.cloud/v3/%s/%s' % (domain_id, path), video_id,
121             'Downloading %s JSON' % path, data=urlencode_postdata(data),
122             headers=headers)
123         self._handle_error(result)
124         return result['result']
125
126     def _real_extract(self, url):
127         mobj = re.match(self._VALID_URL, url)
128         domain_id = mobj.group('domain_id') or mobj.group('domain_id_s')
129         video_id = mobj.group('id')
130
131         # Reverse engineered from JS code (see getDeviceID function)
132         device_id = '%d:%d:%d%d' % (
133             random.randint(1, 4), int(time.time()),
134             random.randint(1e4, 99999), random.randint(1, 9))
135
136         result = self._call_api(domain_id, 'session/init', video_id, data={
137             'nxp_devh': device_id,
138             'nxp_userh': '',
139             'precid': '0',
140             'playlicense': '0',
141             'screenx': '1920',
142             'screeny': '1080',
143             'playerversion': '6.0.00',
144             'gateway': 'html5',
145             'adGateway': '',
146             'explicitlanguage': 'en-US',
147             'addTextTemplates': '1',
148             'addDomainData': '1',
149             'addAdModel': '1',
150         }, headers={
151             'X-Request-Enable-Auth-Fallback': '1',
152         })
153
154         cid = result['general']['cid']
155
156         # As described in [1] X-Request-Token generation algorithm is
157         # as follows:
158         #   md5( operation + domain_id + domain_secret )
159         # where domain_secret is a static value that will be given by nexx.tv
160         # as per [1]. Here is how this "secret" is generated (reversed
161         # from _play.api.init function, search for clienttoken). So it's
162         # actually not static and not that much of a secret.
163         # 1. https://nexxtvstorage.blob.core.windows.net/files/201610/27.pdf
164         secret = result['device']['clienttoken'][int(device_id[0]):]
165         secret = secret[0:len(secret) - int(device_id[-1])]
166
167         op = 'byid'
168
169         # Reversed from JS code for _play.api.call function (search for
170         # X-Request-Token)
171         request_token = hashlib.md5(
172             ''.join((op, domain_id, secret)).encode('utf-8')).hexdigest()
173
174         video = self._call_api(
175             domain_id, 'videos/%s/%s' % (op, video_id), video_id, data={
176                 'additionalfields': 'language,channel,actors,studio,licenseby,slug,subtitle,teaser,description',
177                 'addInteractionOptions': '1',
178                 'addStatusDetails': '1',
179                 'addStreamDetails': '1',
180                 'addCaptions': '1',
181                 'addScenes': '1',
182                 'addHotSpots': '1',
183                 'addBumpers': '1',
184                 'captionFormat': 'data',
185             }, headers={
186                 'X-Request-CID': cid,
187                 'X-Request-Token': request_token,
188             })
189
190         general = video['general']
191         title = general['title']
192
193         stream_data = video['streamdata']
194         language = general.get('language_raw') or ''
195
196         # TODO: reverse more cdns and formats
197
198         cdn = stream_data['cdnType']
199         assert cdn == 'azure'
200
201         azure_locator = stream_data['azureLocator']
202
203         AZURE_URL = 'http://nx-p%02d.akamaized.net/'
204
205         for secure in ('s', ''):
206             cdn_shield = stream_data.get('cdnShieldHTTP%s' % secure.upper())
207             if cdn_shield:
208                 azure_base = 'http%s://%s' % (secure, cdn_shield)
209                 break
210         else:
211             azure_base = AZURE_URL % int(stream_data['azureAccount'].replace('nexxplayplus', ''))
212
213         is_ml = ',' in language
214         azure_m3u8_url = '%s%s/%s_src%s.ism/Manifest(format=m3u8-aapl)' % (
215             azure_base, azure_locator, video_id, ('_manifest' if is_ml else ''))
216
217         protection_token = try_get(
218             video, lambda x: x['protectiondata']['token'], compat_str)
219         if protection_token:
220             azure_m3u8_url += '?hdnts=%s' % protection_token
221
222         formats = self._extract_m3u8_formats(
223             azure_m3u8_url, video_id, 'mp4', entry_protocol='m3u8_native',
224             m3u8_id='%s-hls' % cdn)
225         self._sort_formats(formats)
226
227         return {
228             'id': video_id,
229             'title': title,
230             'alt_title': general.get('subtitle'),
231             'description': general.get('description'),
232             'release_year': int_or_none(general.get('year')),
233             'creator': general.get('studio') or general.get('studio_adref'),
234             'thumbnail': try_get(
235                 video, lambda x: x['imagedata']['thumb'], compat_str),
236             'duration': parse_duration(general.get('runtime')),
237             'timestamp': int_or_none(general.get('uploaded')),
238             'episode_number': int_or_none(try_get(
239                 video, lambda x: x['episodedata']['episode'])),
240             'season_number': int_or_none(try_get(
241                 video, lambda x: x['episodedata']['season'])),
242             'formats': formats,
243         }
244
245
246 class NexxEmbedIE(InfoExtractor):
247     _VALID_URL = r'https?://embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?P<id>[^/?#&]+)'
248     _TEST = {
249         'url': 'http://embed.nexx.cloud/748/KC1614647Z27Y7T?autoplay=1',
250         'md5': '16746bfc28c42049492385c989b26c4a',
251         'info_dict': {
252             'id': '161464',
253             'ext': 'mp4',
254             'title': 'Nervenkitzel Achterbahn',
255             'alt_title': 'Karussellbauer in Deutschland',
256             'description': 'md5:ffe7b1cc59a01f585e0569949aef73cc',
257             'release_year': 2005,
258             'creator': 'SPIEGEL TV',
259             'thumbnail': r're:^https?://.*\.jpg$',
260             'duration': 2761,
261             'timestamp': 1394021479,
262             'upload_date': '20140305',
263         },
264         'params': {
265             'format': 'bestvideo',
266             'skip_download': True,
267         },
268     }
269
270     @staticmethod
271     def _extract_urls(webpage):
272         # Reference:
273         # 1. https://nx-s.akamaized.net/files/201510/44.pdf
274
275         # iFrame Embed Integration
276         return [mobj.group('url') for mobj in re.finditer(
277             r'<iframe[^>]+\bsrc=(["\'])(?P<url>(?:https?:)?//embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?:(?!\1).)+)\1',
278             webpage)]
279
280     def _real_extract(self, url):
281         embed_id = self._match_id(url)
282
283         webpage = self._download_webpage(url, embed_id)
284
285         return self.url_result(NexxIE._extract_url(webpage), ie=NexxIE.ie_key())