]> gitweb @ CieloNegro.org - youtube-dl.git/blob - youtube_dl/extractor/nexx.py
[cbs] extract smpte and vtt subtitles
[youtube-dl.git] / 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                             https?://arc\.nexx\.cloud/api/video/
26                         )
27                         (?P<id>\d+)
28                     '''
29     _TESTS = [{
30         # movie
31         'url': 'https://api.nexx.cloud/v3/748/videos/byid/128907',
32         'md5': '31899fd683de49ad46f4ee67e53e83fe',
33         'info_dict': {
34             'id': '128907',
35             'ext': 'mp4',
36             'title': 'Stiftung Warentest',
37             'alt_title': 'Wie ein Test abläuft',
38             'description': 'md5:d1ddb1ef63de721132abd38639cc2fd2',
39             'creator': 'SPIEGEL TV',
40             'thumbnail': r're:^https?://.*\.jpg$',
41             'duration': 2509,
42             'timestamp': 1384264416,
43             'upload_date': '20131112',
44         },
45     }, {
46         # episode
47         'url': 'https://api.nexx.cloud/v3/741/videos/byid/247858',
48         'info_dict': {
49             'id': '247858',
50             'ext': 'mp4',
51             'title': 'Return of the Golden Child (OV)',
52             'description': 'md5:5d969537509a92b733de21bae249dc63',
53             'release_year': 2017,
54             'thumbnail': r're:^https?://.*\.jpg$',
55             'duration': 1397,
56             'timestamp': 1495033267,
57             'upload_date': '20170517',
58             'episode_number': 2,
59             'season_number': 2,
60         },
61         'params': {
62             'skip_download': True,
63         },
64         'skip': 'HTTP Error 404: Not Found',
65     }, {
66         # does not work via arc
67         'url': 'nexx:741:1269984',
68         'md5': 'c714b5b238b2958dc8d5642addba6886',
69         'info_dict': {
70             'id': '1269984',
71             'ext': 'mp4',
72             'title': '1 TAG ohne KLO... wortwörtlich! 😑',
73             'alt_title': '1 TAG ohne KLO... wortwörtlich! 😑',
74             'thumbnail': r're:^https?://.*\.jpg$',
75             'duration': 607,
76             'timestamp': 1518614955,
77             'upload_date': '20180214',
78         },
79     }, {
80         # free cdn from http://www.spiegel.de/video/eifel-zoo-aufregung-um-ausgebrochene-raubtiere-video-99018031.html
81         'url': 'nexx:747:1533779',
82         'md5': '6bf6883912b82b7069fb86c2297e9893',
83         'info_dict': {
84             'id': '1533779',
85             'ext': 'mp4',
86             'title': 'Aufregung um ausgebrochene Raubtiere',
87             'alt_title': 'Eifel-Zoo',
88             'description': 'md5:f21375c91c74ad741dcb164c427999d2',
89             'thumbnail': r're:^https?://.*\.jpg$',
90             'duration': 111,
91             'timestamp': 1527874460,
92             'upload_date': '20180601',
93         },
94     }, {
95         'url': 'https://api.nexxcdn.com/v3/748/videos/byid/128907',
96         'only_matching': True,
97     }, {
98         'url': 'nexx:748:128907',
99         'only_matching': True,
100     }, {
101         'url': 'nexx:128907',
102         'only_matching': True,
103     }, {
104         'url': 'https://arc.nexx.cloud/api/video/128907.json',
105         'only_matching': True,
106     }]
107
108     @staticmethod
109     def _extract_domain_id(webpage):
110         mobj = re.search(
111             r'<script\b[^>]+\bsrc=["\'](?:https?:)?//require\.nexx(?:\.cloud|cdn\.com)/(?P<id>\d+)',
112             webpage)
113         return mobj.group('id') if mobj else None
114
115     @staticmethod
116     def _extract_urls(webpage):
117         # Reference:
118         # 1. https://nx-s.akamaized.net/files/201510/44.pdf
119
120         entries = []
121
122         # JavaScript Integration
123         domain_id = NexxIE._extract_domain_id(webpage)
124         if domain_id:
125             for video_id in re.findall(
126                     r'(?is)onPLAYReady.+?_play\.init\s*\(.+?\s*,\s*["\']?(\d+)',
127                     webpage):
128                 entries.append(
129                     'https://api.nexx.cloud/v3/%s/videos/byid/%s'
130                     % (domain_id, video_id))
131
132         # TODO: support more embed formats
133
134         return entries
135
136     @staticmethod
137     def _extract_url(webpage):
138         return NexxIE._extract_urls(webpage)[0]
139
140     def _handle_error(self, response):
141         status = int_or_none(try_get(
142             response, lambda x: x['metadata']['status']) or 200)
143         if 200 <= status < 300:
144             return
145         raise ExtractorError(
146             '%s said: %s' % (self.IE_NAME, response['metadata']['errorhint']),
147             expected=True)
148
149     def _call_api(self, domain_id, path, video_id, data=None, headers={}):
150         headers['Content-Type'] = 'application/x-www-form-urlencoded; charset=UTF-8'
151         result = self._download_json(
152             'https://api.nexx.cloud/v3/%s/%s' % (domain_id, path), video_id,
153             'Downloading %s JSON' % path, data=urlencode_postdata(data),
154             headers=headers)
155         self._handle_error(result)
156         return result['result']
157
158     def _extract_free_formats(self, video, video_id):
159         stream_data = video['streamdata']
160         cdn = stream_data['cdnType']
161         assert cdn == 'free'
162
163         hash = video['general']['hash']
164
165         ps = compat_str(stream_data['originalDomain'])
166         if stream_data['applyFolderHierarchy'] == 1:
167             s = ('%04d' % int(video_id))[::-1]
168             ps += '/%s/%s' % (s[0:2], s[2:4])
169         ps += '/%s/%s_' % (video_id, hash)
170
171         t = 'http://%s' + ps
172         fd = stream_data['azureFileDistribution'].split(',')
173         cdn_provider = stream_data['cdnProvider']
174
175         def p0(p):
176             return '_%s' % p if stream_data['applyAzureStructure'] == 1 else ''
177
178         formats = []
179         if cdn_provider == 'ak':
180             t += ','
181             for i in fd:
182                 p = i.split(':')
183                 t += p[1] + p0(int(p[0])) + ','
184             t += '.mp4.csmil/master.%s'
185         elif cdn_provider == 'ce':
186             k = t.split('/')
187             h = k.pop()
188             http_base = t = '/'.join(k)
189             http_base = http_base % stream_data['cdnPathHTTP']
190             t += '/asset.ism/manifest.%s?dcp_ver=aos4&videostream='
191             for i in fd:
192                 p = i.split(':')
193                 tbr = int(p[0])
194                 filename = '%s%s%s.mp4' % (h, p[1], p0(tbr))
195                 f = {
196                     'url': http_base + '/' + filename,
197                     'format_id': '%s-http-%d' % (cdn, tbr),
198                     'tbr': tbr,
199                 }
200                 width_height = p[1].split('x')
201                 if len(width_height) == 2:
202                     f.update({
203                         'width': int_or_none(width_height[0]),
204                         'height': int_or_none(width_height[1]),
205                     })
206                 formats.append(f)
207                 a = filename + ':%s' % (tbr * 1000)
208                 t += a + ','
209             t = t[:-1] + '&audiostream=' + a.split(':')[0]
210         else:
211             assert False
212
213         if cdn_provider == 'ce':
214             formats.extend(self._extract_mpd_formats(
215                 t % (stream_data['cdnPathDASH'], 'mpd'), video_id,
216                 mpd_id='%s-dash' % cdn, fatal=False))
217         formats.extend(self._extract_m3u8_formats(
218             t % (stream_data['cdnPathHLS'], 'm3u8'), video_id, 'mp4',
219             entry_protocol='m3u8_native', m3u8_id='%s-hls' % cdn, fatal=False))
220
221         return formats
222
223     def _extract_azure_formats(self, video, video_id):
224         stream_data = video['streamdata']
225         cdn = stream_data['cdnType']
226         assert cdn == 'azure'
227
228         azure_locator = stream_data['azureLocator']
229
230         def get_cdn_shield_base(shield_type='', static=False):
231             for secure in ('', 's'):
232                 cdn_shield = stream_data.get('cdnShield%sHTTP%s' % (shield_type, secure.upper()))
233                 if cdn_shield:
234                     return 'http%s://%s' % (secure, cdn_shield)
235             else:
236                 if 'fb' in stream_data['azureAccount']:
237                     prefix = 'df' if static else 'f'
238                 else:
239                     prefix = 'd' if static else 'p'
240                 account = int(stream_data['azureAccount'].replace('nexxplayplus', '').replace('nexxplayfb', ''))
241                 return 'http://nx-%s%02d.akamaized.net/' % (prefix, account)
242
243         language = video['general'].get('language_raw') or ''
244
245         azure_stream_base = get_cdn_shield_base()
246         is_ml = ',' in language
247         azure_manifest_url = '%s%s/%s_src%s.ism/Manifest' % (
248             azure_stream_base, azure_locator, video_id, ('_manifest' if is_ml else '')) + '%s'
249
250         protection_token = try_get(
251             video, lambda x: x['protectiondata']['token'], compat_str)
252         if protection_token:
253             azure_manifest_url += '?hdnts=%s' % protection_token
254
255         formats = self._extract_m3u8_formats(
256             azure_manifest_url % '(format=m3u8-aapl)',
257             video_id, 'mp4', 'm3u8_native',
258             m3u8_id='%s-hls' % cdn, fatal=False)
259         formats.extend(self._extract_mpd_formats(
260             azure_manifest_url % '(format=mpd-time-csf)',
261             video_id, mpd_id='%s-dash' % cdn, fatal=False))
262         formats.extend(self._extract_ism_formats(
263             azure_manifest_url % '', video_id, ism_id='%s-mss' % cdn, fatal=False))
264
265         azure_progressive_base = get_cdn_shield_base('Prog', True)
266         azure_file_distribution = stream_data.get('azureFileDistribution')
267         if azure_file_distribution:
268             fds = azure_file_distribution.split(',')
269             if fds:
270                 for fd in fds:
271                     ss = fd.split(':')
272                     if len(ss) == 2:
273                         tbr = int_or_none(ss[0])
274                         if tbr:
275                             f = {
276                                 'url': '%s%s/%s_src_%s_%d.mp4' % (
277                                     azure_progressive_base, azure_locator, video_id, ss[1], tbr),
278                                 'format_id': '%s-http-%d' % (cdn, tbr),
279                                 'tbr': tbr,
280                             }
281                             width_height = ss[1].split('x')
282                             if len(width_height) == 2:
283                                 f.update({
284                                     'width': int_or_none(width_height[0]),
285                                     'height': int_or_none(width_height[1]),
286                                 })
287                             formats.append(f)
288
289         return formats
290
291     def _real_extract(self, url):
292         mobj = re.match(self._VALID_URL, url)
293         domain_id = mobj.group('domain_id') or mobj.group('domain_id_s')
294         video_id = mobj.group('id')
295
296         video = None
297
298         response = self._download_json(
299             'https://arc.nexx.cloud/api/video/%s.json' % video_id,
300             video_id, fatal=False)
301         if response and isinstance(response, dict):
302             result = response.get('result')
303             if result and isinstance(result, dict):
304                 video = result
305
306         # not all videos work via arc, e.g. nexx:741:1269984
307         if not video:
308             # Reverse engineered from JS code (see getDeviceID function)
309             device_id = '%d:%d:%d%d' % (
310                 random.randint(1, 4), int(time.time()),
311                 random.randint(1e4, 99999), random.randint(1, 9))
312
313             result = self._call_api(domain_id, 'session/init', video_id, data={
314                 'nxp_devh': device_id,
315                 'nxp_userh': '',
316                 'precid': '0',
317                 'playlicense': '0',
318                 'screenx': '1920',
319                 'screeny': '1080',
320                 'playerversion': '6.0.00',
321                 'gateway': 'html5',
322                 'adGateway': '',
323                 'explicitlanguage': 'en-US',
324                 'addTextTemplates': '1',
325                 'addDomainData': '1',
326                 'addAdModel': '1',
327             }, headers={
328                 'X-Request-Enable-Auth-Fallback': '1',
329             })
330
331             cid = result['general']['cid']
332
333             # As described in [1] X-Request-Token generation algorithm is
334             # as follows:
335             #   md5( operation + domain_id + domain_secret )
336             # where domain_secret is a static value that will be given by nexx.tv
337             # as per [1]. Here is how this "secret" is generated (reversed
338             # from _play.api.init function, search for clienttoken). So it's
339             # actually not static and not that much of a secret.
340             # 1. https://nexxtvstorage.blob.core.windows.net/files/201610/27.pdf
341             secret = result['device']['clienttoken'][int(device_id[0]):]
342             secret = secret[0:len(secret) - int(device_id[-1])]
343
344             op = 'byid'
345
346             # Reversed from JS code for _play.api.call function (search for
347             # X-Request-Token)
348             request_token = hashlib.md5(
349                 ''.join((op, domain_id, secret)).encode('utf-8')).hexdigest()
350
351             video = self._call_api(
352                 domain_id, 'videos/%s/%s' % (op, video_id), video_id, data={
353                     'additionalfields': 'language,channel,actors,studio,licenseby,slug,subtitle,teaser,description',
354                     'addInteractionOptions': '1',
355                     'addStatusDetails': '1',
356                     'addStreamDetails': '1',
357                     'addCaptions': '1',
358                     'addScenes': '1',
359                     'addHotSpots': '1',
360                     'addBumpers': '1',
361                     'captionFormat': 'data',
362                 }, headers={
363                     'X-Request-CID': cid,
364                     'X-Request-Token': request_token,
365                 })
366
367         general = video['general']
368         title = general['title']
369
370         cdn = video['streamdata']['cdnType']
371
372         if cdn == 'azure':
373             formats = self._extract_azure_formats(video, video_id)
374         elif cdn == 'free':
375             formats = self._extract_free_formats(video, video_id)
376         else:
377             # TODO: reverse more cdns
378             assert False
379
380         self._sort_formats(formats)
381
382         return {
383             'id': video_id,
384             'title': title,
385             'alt_title': general.get('subtitle'),
386             'description': general.get('description'),
387             'release_year': int_or_none(general.get('year')),
388             'creator': general.get('studio') or general.get('studio_adref'),
389             'thumbnail': try_get(
390                 video, lambda x: x['imagedata']['thumb'], compat_str),
391             'duration': parse_duration(general.get('runtime')),
392             'timestamp': int_or_none(general.get('uploaded')),
393             'episode_number': int_or_none(try_get(
394                 video, lambda x: x['episodedata']['episode'])),
395             'season_number': int_or_none(try_get(
396                 video, lambda x: x['episodedata']['season'])),
397             'formats': formats,
398         }
399
400
401 class NexxEmbedIE(InfoExtractor):
402     _VALID_URL = r'https?://embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?P<id>[^/?#&]+)'
403     _TEST = {
404         'url': 'http://embed.nexx.cloud/748/KC1614647Z27Y7T?autoplay=1',
405         'md5': '16746bfc28c42049492385c989b26c4a',
406         'info_dict': {
407             'id': '161464',
408             'ext': 'mp4',
409             'title': 'Nervenkitzel Achterbahn',
410             'alt_title': 'Karussellbauer in Deutschland',
411             'description': 'md5:ffe7b1cc59a01f585e0569949aef73cc',
412             'release_year': 2005,
413             'creator': 'SPIEGEL TV',
414             'thumbnail': r're:^https?://.*\.jpg$',
415             'duration': 2761,
416             'timestamp': 1394021479,
417             'upload_date': '20140305',
418         },
419         'params': {
420             'format': 'bestvideo',
421             'skip_download': True,
422         },
423     }
424
425     @staticmethod
426     def _extract_urls(webpage):
427         # Reference:
428         # 1. https://nx-s.akamaized.net/files/201510/44.pdf
429
430         # iFrame Embed Integration
431         return [mobj.group('url') for mobj in re.finditer(
432             r'<iframe[^>]+\bsrc=(["\'])(?P<url>(?:https?:)?//embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?:(?!\1).)+)\1',
433             webpage)]
434
435     def _real_extract(self, url):
436         embed_id = self._match_id(url)
437
438         webpage = self._download_webpage(url, embed_id)
439
440         return self.url_result(NexxIE._extract_url(webpage), ie=NexxIE.ie_key())