yt_dlp/extractor/nexx.py

   1 import hashlib
   2 import random
   3 import re
   4 import time
   5
   6 from .common import InfoExtractor
   7 from ..utils import (
   8     ExtractorError,
   9     int_or_none,
  10     parse_duration,
  11     srt_subtitles_timecode,
  12     traverse_obj,
  13     try_get,
  14     urlencode_postdata,
  15 )
  16
  17
  18 class NexxIE(InfoExtractor):
  19     _VALID_URL = r'''(?x)
  20                         (?:
  21                             https?://api\.nexx(?:\.cloud|cdn\.com)/v3(?:\.\d)?/(?P<domain_id>\d+)/videos/byid/|
  22                             nexx:(?:(?P<domain_id_s>\d+):)?|
  23                             https?://arc\.nexx\.cloud/api/video/
  24                         )
  25                         (?P<id>\d+)
  26                     '''
  27     _TESTS = [{
  28         # movie
  29         'url': 'https://api.nexx.cloud/v3/748/videos/byid/128907',
  30         'md5': '31899fd683de49ad46f4ee67e53e83fe',
  31         'info_dict': {
  32             'id': '128907',
  33             'ext': 'mp4',
  34             'title': 'Stiftung Warentest',
  35             'alt_title': 'Wie ein Test abläuft',
  36             'description': 'md5:d1ddb1ef63de721132abd38639cc2fd2',
  37             'creator': 'SPIEGEL TV',
  38             'thumbnail': r're:^https?://.*\.jpg$',
  39             'duration': 2509,
  40             'timestamp': 1384264416,
  41             'upload_date': '20131112',
  42         },
  43         'skip': 'Spiegel nexx CDNs are now disabled',
  44     }, {
  45         # episode with captions
  46         'url': 'https://api.nexx.cloud/v3.1/741/videos/byid/1701834',
  47         'info_dict': {
  48             'id': '1701834',
  49             'ext': 'mp4',
  50             'title': 'Mein Leben mit \'nem TikTok E-Boy 😤',
  51             'alt_title': 'Mein Leben mit \'nem TikTok E-Boy 😤',
  52             'description': 'md5:f84f395a881fd143f952c892deab528d',
  53             'thumbnail': r're:^https?://.*\.jpg$',
  54             'duration': 770,
  55             'timestamp': 1595600027,
  56             'upload_date': '20200724',
  57             'episode_number': 2,
  58             'season_number': 2,
  59             'episode': 'Episode 2',
  60             'season': 'Season 2',
  61         },
  62         'params': {
  63             'skip_download': True,
  64         },
  65     }, {
  66         'url': 'nexx:741:1269984',
  67         'md5': 'd5f14e14b592501e51addd5abef95a7f',
  68         'info_dict': {
  69             'id': '1269984',
  70             'ext': 'mp4',
  71             'title': '1 TAG ohne KLO... wortwörtlich! ?',
  72             'alt_title': '1 TAG ohne KLO... wortwörtlich! ?',
  73             'description': 'md5:2016393a31991a900946432ccdd09a6f',
  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         'skip': 'Spiegel nexx CDNs are now disabled',
  95     }, {
  96         'url': 'https://api.nexxcdn.com/v3/748/videos/byid/128907',
  97         'only_matching': True,
  98     }, {
  99         'url': 'nexx:748:128907',
 100         'only_matching': True,
 101     }, {
 102         'url': 'nexx:128907',
 103         'only_matching': True,
 104     }, {
 105         'url': 'https://arc.nexx.cloud/api/video/128907.json',
 106         'only_matching': True,
 107     }]
 108
 109     @staticmethod
 110     def _extract_domain_id(webpage):
 111         mobj = re.search(
 112             r'<script\b[^>]+\bsrc=["\'](?:https?:)?//(?:require|arc)\.nexx(?:\.cloud|cdn\.com)/(?:sdk/)?(?P<id>\d+)',
 113             webpage)
 114         return mobj.group('id') if mobj else None
 115
 116     @classmethod
 117     def _extract_embed_urls(cls, url, webpage):
 118         # Reference:
 119         # 1. https://nx-s.akamaized.net/files/201510/44.pdf
 120
 121         entries = []
 122
 123         # JavaScript Integration
 124         domain_id = NexxIE._extract_domain_id(webpage)
 125         if domain_id:
 126             for video_id in re.findall(
 127                     r'(?is)onPLAYReady.+?_play\.(?:init|(?:control\.)?addPlayer)\s*\(.+?\s*,\s*["\']?(\d+)',
 128                     webpage):
 129                 entries.append(
 130                     f'https://api.nexx.cloud/v3/{domain_id}/videos/byid/{video_id}')
 131
 132         # TODO: support more embed formats
 133
 134         return entries
 135
 136     def _handle_error(self, response):
 137         if traverse_obj(response, ('metadata', 'notice'), expected_type=str):
 138             self.report_warning('{} said: {}'.format(self.IE_NAME, response['metadata']['notice']))
 139         status = int_or_none(try_get(
 140             response, lambda x: x['metadata']['status']) or 200)
 141         if 200 <= status < 300:
 142             return
 143         raise ExtractorError(
 144             '{} said: {}'.format(self.IE_NAME, response['metadata']['errorhint']),
 145             expected=True)
 146
 147     def _call_api(self, domain_id, path, video_id, data=None, headers={}):
 148         headers['Content-Type'] = 'application/x-www-form-urlencoded; charset=UTF-8'
 149         result = self._download_json(
 150             f'https://api.nexx.cloud/v3/{domain_id}/{path}', video_id,
 151             f'Downloading {path} JSON', data=urlencode_postdata(data),
 152             headers=headers)
 153         self._handle_error(result)
 154         return result['result']
 155
 156     def _extract_free_formats(self, video, video_id):
 157         stream_data = video['streamdata']
 158         cdn = stream_data['cdnType']
 159         assert cdn == 'free'
 160
 161         video_hash = video['general']['hash']
 162
 163         ps = str(stream_data['originalDomain'])
 164         if stream_data['applyFolderHierarchy'] == 1:
 165             s = ('%04d' % int(video_id))[::-1]
 166             ps += f'/{s[0:2]}/{s[2:4]}'
 167         ps += f'/{video_id}/{video_hash}_'
 168
 169         t = 'http://%s' + ps
 170         fd = stream_data['azureFileDistribution'].split(',')
 171         cdn_provider = stream_data['cdnProvider']
 172
 173         def p0(p):
 174             return f'_{p}' if stream_data['applyAzureStructure'] == 1 else ''
 175
 176         formats = []
 177         if cdn_provider == 'ak':
 178             t += ','
 179             for i in fd:
 180                 p = i.split(':')
 181                 t += p[1] + p0(int(p[0])) + ','
 182             t += '.mp4.csmil/master.%s'
 183         elif cdn_provider == 'ce':
 184             k = t.split('/')
 185             h = k.pop()
 186             http_base = t = '/'.join(k)
 187             http_base = http_base % stream_data['cdnPathHTTP']
 188             t += '/asset.ism/manifest.%s?dcp_ver=aos4&videostream='
 189             for i in fd:
 190                 p = i.split(':')
 191                 tbr = int(p[0])
 192                 filename = f'{h}{p[1]}{p0(tbr)}.mp4'
 193                 f = {
 194                     'url': http_base + '/' + filename,
 195                     'format_id': f'{cdn}-http-{tbr}',
 196                     'tbr': tbr,
 197                 }
 198                 width_height = p[1].split('x')
 199                 if len(width_height) == 2:
 200                     f.update({
 201                         'width': int_or_none(width_height[0]),
 202                         'height': int_or_none(width_height[1]),
 203                     })
 204                 formats.append(f)
 205                 a = filename + f':{tbr * 1000}'
 206                 t += a + ','
 207             t = t[:-1] + '&audiostream=' + a.split(':')[0]
 208         else:
 209             assert False
 210
 211         if cdn_provider == 'ce':
 212             formats.extend(self._extract_mpd_formats(
 213                 t % (stream_data['cdnPathDASH'], 'mpd'), video_id,
 214                 mpd_id=f'{cdn}-dash', fatal=False))
 215         formats.extend(self._extract_m3u8_formats(
 216             t % (stream_data['cdnPathHLS'], 'm3u8'), video_id, 'mp4',
 217             entry_protocol='m3u8_native', m3u8_id=f'{cdn}-hls', fatal=False))
 218
 219         return formats
 220
 221     def _extract_3q_formats(self, video, video_id):
 222         stream_data = video['streamdata']
 223         cdn = stream_data['cdnType']
 224         assert cdn == '3q'
 225
 226         q_acc, q_prefix, q_locator, q_hash = stream_data['qAccount'], stream_data['qPrefix'], stream_data['qLocator'], stream_data['qHash']
 227         protection_key = traverse_obj(
 228             video, ('protectiondata', 'key'), expected_type=str)
 229
 230         def get_cdn_shield_base(shield_type=''):
 231             for secure in ('', 's'):
 232                 cdn_shield = stream_data.get(f'cdnShield{shield_type}HTTP{secure.upper()}')
 233                 if cdn_shield:
 234                     return f'http{secure}://{cdn_shield}'
 235             return f'http://sdn-global-{"prog" if shield_type.lower() == "prog" else "streaming"}-cache.3qsdn.com/' + (f's/{protection_key}/' if protection_key else '')
 236
 237         stream_base = get_cdn_shield_base()
 238
 239         formats = []
 240         formats.extend(self._extract_m3u8_formats(
 241             f'{stream_base}{q_acc}/files/{q_prefix}/{q_locator}/{q_acc}-{stream_data.get("qHEVCHash") or q_hash}.ism/manifest.m3u8',
 242             video_id, 'mp4', m3u8_id=f'{cdn}-hls', fatal=False))
 243         formats.extend(self._extract_mpd_formats(
 244             f'{stream_base}{q_acc}/files/{q_prefix}/{q_locator}/{q_acc}-{q_hash}.ism/manifest.mpd',
 245             video_id, mpd_id=f'{cdn}-dash', fatal=False))
 246
 247         progressive_base = get_cdn_shield_base('Prog')
 248         q_references = stream_data.get('qReferences') or ''
 249         fds = q_references.split(',')
 250         for fd in fds:
 251             ss = fd.split(':')
 252             if len(ss) != 3:
 253                 continue
 254             tbr = int_or_none(ss[1], scale=1000)
 255             formats.append({
 256                 'url': f'{progressive_base}{q_acc}/uploads/{q_acc}-{ss[2]}.webm',
 257                 'format_id': f'{cdn}-{ss[0]}{f"-{tbr}" if tbr else ""}',
 258                 'tbr': tbr,
 259             })
 260
 261         azure_file_distribution = stream_data.get('azureFileDistribution') or ''
 262         fds = azure_file_distribution.split(',')
 263         for fd in fds:
 264             ss = fd.split(':')
 265             if len(ss) != 3:
 266                 continue
 267             tbr = int_or_none(ss[0])
 268             width, height = ss[1].split('x') if len(ss[1].split('x')) == 2 else (None, None)
 269             f = {
 270                 'url': f'{progressive_base}{q_acc}/files/{q_prefix}/{q_locator}/{ss[2]}.mp4',
 271                 'format_id': f'{cdn}-http-{f"-{tbr}" if tbr else ""}',
 272                 'tbr': tbr,
 273                 'width': int_or_none(width),
 274                 'height': int_or_none(height),
 275             }
 276             formats.append(f)
 277
 278         return formats
 279
 280     def _extract_azure_formats(self, video, video_id):
 281         stream_data = video['streamdata']
 282         cdn = stream_data['cdnType']
 283         assert cdn == 'azure'
 284
 285         azure_locator = stream_data['azureLocator']
 286
 287         def get_cdn_shield_base(shield_type='', static=False):
 288             for secure in ('', 's'):
 289                 cdn_shield = stream_data.get(f'cdnShield{shield_type}HTTP{secure.upper()}')
 290                 if cdn_shield:
 291                     return f'http{secure}://{cdn_shield}'
 292             if 'fb' in stream_data['azureAccount']:
 293                 prefix = 'df' if static else 'f'
 294             else:
 295                 prefix = 'd' if static else 'p'
 296             account = int(stream_data['azureAccount'].replace('nexxplayplus', '').replace('nexxplayfb', ''))
 297             return 'http://nx-%s%02d.akamaized.net/' % (prefix, account)
 298
 299         language = video['general'].get('language_raw') or ''
 300
 301         azure_stream_base = get_cdn_shield_base()
 302         is_ml = ',' in language
 303         azure_manifest_url = '{}{}/{}_src{}.ism/Manifest'.format(
 304             azure_stream_base, azure_locator, video_id, ('_manifest' if is_ml else '')) + '%s'
 305
 306         protection_token = try_get(
 307             video, lambda x: x['protectiondata']['token'], str)
 308         if protection_token:
 309             azure_manifest_url += f'?hdnts={protection_token}'
 310
 311         formats = self._extract_m3u8_formats(
 312             azure_manifest_url % '(format=m3u8-aapl)',
 313             video_id, 'mp4', 'm3u8_native',
 314             m3u8_id=f'{cdn}-hls', fatal=False)
 315         formats.extend(self._extract_mpd_formats(
 316             azure_manifest_url % '(format=mpd-time-csf)',
 317             video_id, mpd_id=f'{cdn}-dash', fatal=False))
 318         formats.extend(self._extract_ism_formats(
 319             azure_manifest_url % '', video_id, ism_id=f'{cdn}-mss', fatal=False))
 320
 321         azure_progressive_base = get_cdn_shield_base('Prog', True)
 322         azure_file_distribution = stream_data.get('azureFileDistribution')
 323         if azure_file_distribution:
 324             fds = azure_file_distribution.split(',')
 325             if fds:
 326                 for fd in fds:
 327                     ss = fd.split(':')
 328                     if len(ss) == 2:
 329                         tbr = int_or_none(ss[0])
 330                         if tbr:
 331                             f = {
 332                                 'url': f'{azure_progressive_base}{azure_locator}/{video_id}_src_{ss[1]}_{tbr}.mp4',
 333                                 'format_id': f'{cdn}-http-{tbr}',
 334                                 'tbr': tbr,
 335                             }
 336                             width_height = ss[1].split('x')
 337                             if len(width_height) == 2:
 338                                 f.update({
 339                                     'width': int_or_none(width_height[0]),
 340                                     'height': int_or_none(width_height[1]),
 341                                 })
 342                             formats.append(f)
 343
 344         return formats
 345
 346     def _real_extract(self, url):
 347         mobj = self._match_valid_url(url)
 348         domain_id = mobj.group('domain_id') or mobj.group('domain_id_s')
 349         video_id = mobj.group('id')
 350
 351         video = None
 352
 353         def find_video(result):
 354             if isinstance(result, dict):
 355                 return result
 356             elif isinstance(result, list):
 357                 vid = int(video_id)
 358                 for v in result:
 359                     if try_get(v, lambda x: x['general']['ID'], int) == vid:
 360                         return v
 361             return None
 362
 363         response = self._download_json(
 364             f'https://arc.nexx.cloud/api/video/{video_id}.json',
 365             video_id, fatal=False)
 366         if response and isinstance(response, dict):
 367             result = response.get('result')
 368             if result:
 369                 video = find_video(result)
 370
 371         # not all videos work via arc, e.g. nexx:741:1269984
 372         if not video:
 373             # Reverse engineered from JS code (see getDeviceID function)
 374             device_id = f'{random.randint(1, 4)}:{int(time.time())}:{random.randint(10000, 99999)}{random.randint(1, 9)}'
 375
 376             result = self._call_api(domain_id, 'session/init', video_id, data={
 377                 'nxp_devh': device_id,
 378                 'nxp_userh': '',
 379                 'precid': '0',
 380                 'playlicense': '0',
 381                 'screenx': '1920',
 382                 'screeny': '1080',
 383                 'playerversion': '6.0.00',
 384                 'gateway': 'html5',
 385                 'adGateway': '',
 386                 'explicitlanguage': 'en-US',
 387                 'addTextTemplates': '1',
 388                 'addDomainData': '1',
 389                 'addAdModel': '1',
 390             }, headers={
 391                 'X-Request-Enable-Auth-Fallback': '1',
 392             })
 393
 394             cid = result['general']['cid']
 395
 396             # As described in [1] X-Request-Token generation algorithm is
 397             # as follows:
 398             #   md5( operation + domain_id + domain_secret )
 399             # where domain_secret is a static value that will be given by nexx.tv
 400             # as per [1]. Here is how this "secret" is generated (reversed
 401             # from _play._factory.data.getDomainData function, search for
 402             # domaintoken or enableAPIAccess). So it's actually not static
 403             # and not that much of a secret.
 404             # 1. https://nexxtvstorage.blob.core.windows.net/files/201610/27.pdf
 405             secret = result['device']['domaintoken'][int(device_id[0]):]
 406             secret = secret[0:len(secret) - int(device_id[-1])]
 407
 408             op = 'byid'
 409
 410             # Reversed from JS code for _play.api.call function (search for
 411             # X-Request-Token)
 412             request_token = hashlib.md5(
 413                 ''.join((op, domain_id, secret)).encode()).hexdigest()
 414
 415             result = self._call_api(
 416                 domain_id, f'videos/{op}/{video_id}', video_id, data={
 417                     'additionalfields': 'language,channel,format,licenseby,slug,fileversion,episode,season',
 418                     'addInteractionOptions': '1',
 419                     'addStatusDetails': '1',
 420                     'addStreamDetails': '1',
 421                     'addFeatures': '1',
 422                     # Caption format selection doesn't seem to be enforced?
 423                     'addCaptions': 'vtt',
 424                     'addScenes': '1',
 425                     'addChapters': '1',
 426                     'addHotSpots': '1',
 427                     'addConnectedMedia': 'persons',
 428                     'addBumpers': '1',
 429                 }, headers={
 430                     'X-Request-CID': cid,
 431                     'X-Request-Token': request_token,
 432                 })
 433             video = find_video(result)
 434
 435         general = video['general']
 436         title = general['title']
 437
 438         cdn = video['streamdata']['cdnType']
 439
 440         if cdn == 'azure':
 441             formats = self._extract_azure_formats(video, video_id)
 442         elif cdn == 'free':
 443             formats = self._extract_free_formats(video, video_id)
 444         elif cdn == '3q':
 445             formats = self._extract_3q_formats(video, video_id)
 446         else:
 447             self.raise_no_formats(f'{cdn} formats are currently not supported', video_id)
 448
 449         subtitles = {}
 450         for sub in video.get('captiondata') or []:
 451             if sub.get('data'):
 452                 subtitles.setdefault(sub.get('language', 'en'), []).append({
 453                     'ext': 'srt',
 454                     'data': '\n\n'.join(
 455                         f'{i + 1}\n{srt_subtitles_timecode(line["fromms"] / 1000)} --> {srt_subtitles_timecode(line["toms"] / 1000)}\n{line["caption"]}'
 456                         for i, line in enumerate(sub['data'])),
 457                     'name': sub.get('language_long') or sub.get('title'),
 458                 })
 459             elif sub.get('url'):
 460                 subtitles.setdefault(sub.get('language', 'en'), []).append({
 461                     'url': sub['url'],
 462                     'ext': sub.get('format'),
 463                     'name': sub.get('language_long') or sub.get('title'),
 464                 })
 465
 466         return {
 467             'id': video_id,
 468             'title': title,
 469             'alt_title': general.get('subtitle'),
 470             'description': general.get('description'),
 471             'release_year': int_or_none(general.get('year')),
 472             'creator': general.get('studio') or general.get('studio_adref') or None,
 473             'thumbnail': try_get(
 474                 video, lambda x: x['imagedata']['thumb'], str),
 475             'duration': parse_duration(general.get('runtime')),
 476             'timestamp': int_or_none(general.get('uploaded')),
 477             'episode_number': traverse_obj(
 478                 video, (('episodedata', 'general'), 'episode'), expected_type=int, get_all=False),
 479             'season_number': traverse_obj(
 480                 video, (('episodedata', 'general'), 'season'), expected_type=int, get_all=False),
 481             'cast': traverse_obj(video, ('connectedmedia', ..., 'title'), expected_type=str),
 482             'formats': formats,
 483             'subtitles': subtitles,
 484         }
 485
 486
 487 class NexxEmbedIE(InfoExtractor):
 488     _VALID_URL = r'https?://embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?:video/)?(?P<id>[^/?#&]+)'
 489     # Reference. https://nx-s.akamaized.net/files/201510/44.pdf
 490     _EMBED_REGEX = [r'<iframe[^>]+\bsrc=(["\'])(?P<url>(?:https?:)?//embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?:(?!\1).)+)\1']
 491     _TESTS = [{
 492         'url': 'http://embed.nexx.cloud/748/KC1614647Z27Y7T?autoplay=1',
 493         'md5': '16746bfc28c42049492385c989b26c4a',
 494         'info_dict': {
 495             'id': '161464',
 496             'ext': 'mp4',
 497             'title': 'Nervenkitzel Achterbahn',
 498             'alt_title': 'Karussellbauer in Deutschland',
 499             'description': 'md5:ffe7b1cc59a01f585e0569949aef73cc',
 500             'creator': 'SPIEGEL TV',
 501             'thumbnail': r're:^https?://.*\.jpg$',
 502             'duration': 2761,
 503             'timestamp': 1394021479,
 504             'upload_date': '20140305',
 505         },
 506         'params': {
 507             'skip_download': True,
 508         },
 509     }, {
 510         'url': 'https://embed.nexx.cloud/11888/video/DSRTO7UVOX06S7',
 511         'only_matching': True,
 512     }]
 513
 514     def _real_extract(self, url):
 515         embed_id = self._match_id(url)
 516
 517         webpage = self._download_webpage(url, embed_id)
 518
 519         return self.url_result(NexxIE._extract_url(webpage), ie=NexxIE.ie_key())