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 ..compat import compat_str
   8 from ..utils import (
   9     ExtractorError,
  10     int_or_none,
  11     parse_duration,
  12     srt_subtitles_timecode,
  13     traverse_obj,
  14     try_get,
  15     urlencode_postdata,
  16 )
  17
  18
  19 class NexxIE(InfoExtractor):
  20     _VALID_URL = r'''(?x)
  21                         (?:
  22                             https?://api\.nexx(?:\.cloud|cdn\.com)/v3(?:\.\d)?/(?P<domain_id>\d+)/videos/byid/|
  23                             nexx:(?:(?P<domain_id_s>\d+):)?|
  24                             https?://arc\.nexx\.cloud/api/video/
  25                         )
  26                         (?P<id>\d+)
  27                     '''
  28     _TESTS = [{
  29         # movie
  30         'url': 'https://api.nexx.cloud/v3/748/videos/byid/128907',
  31         'md5': '31899fd683de49ad46f4ee67e53e83fe',
  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             'creator': 'SPIEGEL TV',
  39             'thumbnail': r're:^https?://.*\.jpg$',
  40             'duration': 2509,
  41             'timestamp': 1384264416,
  42             'upload_date': '20131112',
  43         },
  44         'skip': 'Spiegel nexx CDNs are now disabled'
  45     }, {
  46         # episode with captions
  47         'url': 'https://api.nexx.cloud/v3.1/741/videos/byid/1701834',
  48         'info_dict': {
  49             'id': '1701834',
  50             'ext': 'mp4',
  51             'title': 'Mein Leben mit \'nem TikTok E-Boy 😤',
  52             'alt_title': 'Mein Leben mit \'nem TikTok E-Boy 😤',
  53             'description': 'md5:f84f395a881fd143f952c892deab528d',
  54             'thumbnail': r're:^https?://.*\.jpg$',
  55             'duration': 770,
  56             'timestamp': 1595600027,
  57             'upload_date': '20200724',
  58             'episode_number': 2,
  59             'season_number': 2,
  60             'episode': 'Episode 2',
  61             'season': 'Season 2',
  62         },
  63         'params': {
  64             'skip_download': True,
  65         },
  66     }, {
  67         'url': 'nexx:741:1269984',
  68         'md5': 'd5f14e14b592501e51addd5abef95a7f',
  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             'description': 'md5:2016393a31991a900946432ccdd09a6f',
  75             'thumbnail': r're:^https?://.*\.jpg$',
  76             'duration': 607,
  77             'timestamp': 1518614955,
  78             'upload_date': '20180214',
  79         },
  80     }, {
  81         # free cdn from http://www.spiegel.de/video/eifel-zoo-aufregung-um-ausgebrochene-raubtiere-video-99018031.html
  82         'url': 'nexx:747:1533779',
  83         'md5': '6bf6883912b82b7069fb86c2297e9893',
  84         'info_dict': {
  85             'id': '1533779',
  86             'ext': 'mp4',
  87             'title': 'Aufregung um ausgebrochene Raubtiere',
  88             'alt_title': 'Eifel-Zoo',
  89             'description': 'md5:f21375c91c74ad741dcb164c427999d2',
  90             'thumbnail': r're:^https?://.*\.jpg$',
  91             'duration': 111,
  92             'timestamp': 1527874460,
  93             'upload_date': '20180601',
  94         },
  95         'skip': 'Spiegel nexx CDNs are now disabled'
  96     }, {
  97         'url': 'https://api.nexxcdn.com/v3/748/videos/byid/128907',
  98         'only_matching': True,
  99     }, {
 100         'url': 'nexx:748:128907',
 101         'only_matching': True,
 102     }, {
 103         'url': 'nexx:128907',
 104         'only_matching': True,
 105     }, {
 106         'url': 'https://arc.nexx.cloud/api/video/128907.json',
 107         'only_matching': True,
 108     }]
 109
 110     @staticmethod
 111     def _extract_domain_id(webpage):
 112         mobj = re.search(
 113             r'<script\b[^>]+\bsrc=["\'](?:https?:)?//(?:require|arc)\.nexx(?:\.cloud|cdn\.com)/(?:sdk/)?(?P<id>\d+)',
 114             webpage)
 115         return mobj.group('id') if mobj else None
 116
 117     @classmethod
 118     def _extract_embed_urls(cls, url, webpage):
 119         # Reference:
 120         # 1. https://nx-s.akamaized.net/files/201510/44.pdf
 121
 122         entries = []
 123
 124         # JavaScript Integration
 125         domain_id = NexxIE._extract_domain_id(webpage)
 126         if domain_id:
 127             for video_id in re.findall(
 128                     r'(?is)onPLAYReady.+?_play\.(?:init|(?:control\.)?addPlayer)\s*\(.+?\s*,\s*["\']?(\d+)',
 129                     webpage):
 130                 entries.append(
 131                     'https://api.nexx.cloud/v3/%s/videos/byid/%s'
 132                     % (domain_id, video_id))
 133
 134         # TODO: support more embed formats
 135
 136         return entries
 137
 138     def _handle_error(self, response):
 139         if traverse_obj(response, ('metadata', 'notice'), expected_type=str):
 140             self.report_warning('%s said: %s' % (self.IE_NAME, response['metadata']['notice']))
 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_3q_formats(self, video, video_id):
 224         stream_data = video['streamdata']
 225         cdn = stream_data['cdnType']
 226         assert cdn == '3q'
 227
 228         q_acc, q_prefix, q_locator, q_hash = stream_data['qAccount'], stream_data['qPrefix'], stream_data['qLocator'], stream_data['qHash']
 229         protection_key = traverse_obj(
 230             video, ('protectiondata', 'key'), expected_type=str)
 231
 232         def get_cdn_shield_base(shield_type=''):
 233             for secure in ('', 's'):
 234                 cdn_shield = stream_data.get('cdnShield%sHTTP%s' % (shield_type, secure.upper()))
 235                 if cdn_shield:
 236                     return 'http%s://%s' % (secure, cdn_shield)
 237             return f'http://sdn-global-{"prog" if shield_type.lower() == "prog" else "streaming"}-cache.3qsdn.com/' + (f's/{protection_key}/' if protection_key else '')
 238
 239         stream_base = get_cdn_shield_base()
 240
 241         formats = []
 242         formats.extend(self._extract_m3u8_formats(
 243             f'{stream_base}{q_acc}/files/{q_prefix}/{q_locator}/{q_acc}-{stream_data.get("qHEVCHash") or q_hash}.ism/manifest.m3u8',
 244             video_id, 'mp4', m3u8_id=f'{cdn}-hls', fatal=False))
 245         formats.extend(self._extract_mpd_formats(
 246             f'{stream_base}{q_acc}/files/{q_prefix}/{q_locator}/{q_acc}-{q_hash}.ism/manifest.mpd',
 247             video_id, mpd_id=f'{cdn}-dash', fatal=False))
 248
 249         progressive_base = get_cdn_shield_base('Prog')
 250         q_references = stream_data.get('qReferences') or ''
 251         fds = q_references.split(',')
 252         for fd in fds:
 253             ss = fd.split(':')
 254             if len(ss) != 3:
 255                 continue
 256             tbr = int_or_none(ss[1], scale=1000)
 257             formats.append({
 258                 'url': f'{progressive_base}{q_acc}/uploads/{q_acc}-{ss[2]}.webm',
 259                 'format_id': f'{cdn}-{ss[0]}{"-%s" % tbr if tbr else ""}',
 260                 'tbr': tbr,
 261             })
 262
 263         azure_file_distribution = stream_data.get('azureFileDistribution') or ''
 264         fds = azure_file_distribution.split(',')
 265         for fd in fds:
 266             ss = fd.split(':')
 267             if len(ss) != 3:
 268                 continue
 269             tbr = int_or_none(ss[0])
 270             width, height = ss[1].split('x') if len(ss[1].split('x')) == 2 else (None, None)
 271             f = {
 272                 'url': f'{progressive_base}{q_acc}/files/{q_prefix}/{q_locator}/{ss[2]}.mp4',
 273                 'format_id': f'{cdn}-http-{"-%s" % tbr if tbr else ""}',
 274                 'tbr': tbr,
 275                 'width': int_or_none(width),
 276                 'height': int_or_none(height),
 277             }
 278             formats.append(f)
 279
 280         return formats
 281
 282     def _extract_azure_formats(self, video, video_id):
 283         stream_data = video['streamdata']
 284         cdn = stream_data['cdnType']
 285         assert cdn == 'azure'
 286
 287         azure_locator = stream_data['azureLocator']
 288
 289         def get_cdn_shield_base(shield_type='', static=False):
 290             for secure in ('', 's'):
 291                 cdn_shield = stream_data.get('cdnShield%sHTTP%s' % (shield_type, secure.upper()))
 292                 if cdn_shield:
 293                     return 'http%s://%s' % (secure, cdn_shield)
 294             else:
 295                 if 'fb' in stream_data['azureAccount']:
 296                     prefix = 'df' if static else 'f'
 297                 else:
 298                     prefix = 'd' if static else 'p'
 299                 account = int(stream_data['azureAccount'].replace('nexxplayplus', '').replace('nexxplayfb', ''))
 300                 return 'http://nx-%s%02d.akamaized.net/' % (prefix, account)
 301
 302         language = video['general'].get('language_raw') or ''
 303
 304         azure_stream_base = get_cdn_shield_base()
 305         is_ml = ',' in language
 306         azure_manifest_url = '%s%s/%s_src%s.ism/Manifest' % (
 307             azure_stream_base, azure_locator, video_id, ('_manifest' if is_ml else '')) + '%s'
 308
 309         protection_token = try_get(
 310             video, lambda x: x['protectiondata']['token'], compat_str)
 311         if protection_token:
 312             azure_manifest_url += '?hdnts=%s' % protection_token
 313
 314         formats = self._extract_m3u8_formats(
 315             azure_manifest_url % '(format=m3u8-aapl)',
 316             video_id, 'mp4', 'm3u8_native',
 317             m3u8_id='%s-hls' % cdn, fatal=False)
 318         formats.extend(self._extract_mpd_formats(
 319             azure_manifest_url % '(format=mpd-time-csf)',
 320             video_id, mpd_id='%s-dash' % cdn, fatal=False))
 321         formats.extend(self._extract_ism_formats(
 322             azure_manifest_url % '', video_id, ism_id='%s-mss' % cdn, fatal=False))
 323
 324         azure_progressive_base = get_cdn_shield_base('Prog', True)
 325         azure_file_distribution = stream_data.get('azureFileDistribution')
 326         if azure_file_distribution:
 327             fds = azure_file_distribution.split(',')
 328             if fds:
 329                 for fd in fds:
 330                     ss = fd.split(':')
 331                     if len(ss) == 2:
 332                         tbr = int_or_none(ss[0])
 333                         if tbr:
 334                             f = {
 335                                 'url': '%s%s/%s_src_%s_%d.mp4' % (
 336                                     azure_progressive_base, azure_locator, video_id, ss[1], tbr),
 337                                 'format_id': '%s-http-%d' % (cdn, tbr),
 338                                 'tbr': tbr,
 339                             }
 340                             width_height = ss[1].split('x')
 341                             if len(width_height) == 2:
 342                                 f.update({
 343                                     'width': int_or_none(width_height[0]),
 344                                     'height': int_or_none(width_height[1]),
 345                                 })
 346                             formats.append(f)
 347
 348         return formats
 349
 350     def _real_extract(self, url):
 351         mobj = self._match_valid_url(url)
 352         domain_id = mobj.group('domain_id') or mobj.group('domain_id_s')
 353         video_id = mobj.group('id')
 354
 355         video = None
 356
 357         def find_video(result):
 358             if isinstance(result, dict):
 359                 return result
 360             elif isinstance(result, list):
 361                 vid = int(video_id)
 362                 for v in result:
 363                     if try_get(v, lambda x: x['general']['ID'], int) == vid:
 364                         return v
 365             return None
 366
 367         response = self._download_json(
 368             'https://arc.nexx.cloud/api/video/%s.json' % video_id,
 369             video_id, fatal=False)
 370         if response and isinstance(response, dict):
 371             result = response.get('result')
 372             if result:
 373                 video = find_video(result)
 374
 375         # not all videos work via arc, e.g. nexx:741:1269984
 376         if not video:
 377             # Reverse engineered from JS code (see getDeviceID function)
 378             device_id = '%d:%d:%d%d' % (
 379                 random.randint(1, 4), int(time.time()),
 380                 random.randint(1e4, 99999), random.randint(1, 9))
 381
 382             result = self._call_api(domain_id, 'session/init', video_id, data={
 383                 'nxp_devh': device_id,
 384                 'nxp_userh': '',
 385                 'precid': '0',
 386                 'playlicense': '0',
 387                 'screenx': '1920',
 388                 'screeny': '1080',
 389                 'playerversion': '6.0.00',
 390                 'gateway': 'html5',
 391                 'adGateway': '',
 392                 'explicitlanguage': 'en-US',
 393                 'addTextTemplates': '1',
 394                 'addDomainData': '1',
 395                 'addAdModel': '1',
 396             }, headers={
 397                 'X-Request-Enable-Auth-Fallback': '1',
 398             })
 399
 400             cid = result['general']['cid']
 401
 402             # As described in [1] X-Request-Token generation algorithm is
 403             # as follows:
 404             #   md5( operation + domain_id + domain_secret )
 405             # where domain_secret is a static value that will be given by nexx.tv
 406             # as per [1]. Here is how this "secret" is generated (reversed
 407             # from _play._factory.data.getDomainData function, search for
 408             # domaintoken or enableAPIAccess). So it's actually not static
 409             # and not that much of a secret.
 410             # 1. https://nexxtvstorage.blob.core.windows.net/files/201610/27.pdf
 411             secret = result['device']['domaintoken'][int(device_id[0]):]
 412             secret = secret[0:len(secret) - int(device_id[-1])]
 413
 414             op = 'byid'
 415
 416             # Reversed from JS code for _play.api.call function (search for
 417             # X-Request-Token)
 418             request_token = hashlib.md5(
 419                 ''.join((op, domain_id, secret)).encode('utf-8')).hexdigest()
 420
 421             result = self._call_api(
 422                 domain_id, 'videos/%s/%s' % (op, video_id), video_id, data={
 423                     'additionalfields': 'language,channel,format,licenseby,slug,fileversion,episode,season',
 424                     'addInteractionOptions': '1',
 425                     'addStatusDetails': '1',
 426                     'addStreamDetails': '1',
 427                     'addFeatures': '1',
 428                     # Caption format selection doesn't seem to be enforced?
 429                     'addCaptions': 'vtt',
 430                     'addScenes': '1',
 431                     'addChapters': '1',
 432                     'addHotSpots': '1',
 433                     'addConnectedMedia': 'persons',
 434                     'addBumpers': '1',
 435                 }, headers={
 436                     'X-Request-CID': cid,
 437                     'X-Request-Token': request_token,
 438                 })
 439             video = find_video(result)
 440
 441         general = video['general']
 442         title = general['title']
 443
 444         cdn = video['streamdata']['cdnType']
 445
 446         if cdn == 'azure':
 447             formats = self._extract_azure_formats(video, video_id)
 448         elif cdn == 'free':
 449             formats = self._extract_free_formats(video, video_id)
 450         elif cdn == '3q':
 451             formats = self._extract_3q_formats(video, video_id)
 452         else:
 453             self.raise_no_formats(f'{cdn} formats are currently not supported', video_id)
 454
 455         subtitles = {}
 456         for sub in video.get('captiondata') or []:
 457             if sub.get('data'):
 458                 subtitles.setdefault(sub.get('language', 'en'), []).append({
 459                     'ext': 'srt',
 460                     'data': '\n\n'.join(
 461                         f'{i + 1}\n{srt_subtitles_timecode(line["fromms"] / 1000)} --> {srt_subtitles_timecode(line["toms"] / 1000)}\n{line["caption"]}'
 462                         for i, line in enumerate(sub['data'])),
 463                     'name': sub.get('language_long') or sub.get('title')
 464                 })
 465             elif sub.get('url'):
 466                 subtitles.setdefault(sub.get('language', 'en'), []).append({
 467                     'url': sub['url'],
 468                     'ext': sub.get('format'),
 469                     'name': sub.get('language_long') or sub.get('title')
 470                 })
 471
 472         return {
 473             'id': video_id,
 474             'title': title,
 475             'alt_title': general.get('subtitle'),
 476             'description': general.get('description'),
 477             'release_year': int_or_none(general.get('year')),
 478             'creator': general.get('studio') or general.get('studio_adref') or None,
 479             'thumbnail': try_get(
 480                 video, lambda x: x['imagedata']['thumb'], compat_str),
 481             'duration': parse_duration(general.get('runtime')),
 482             'timestamp': int_or_none(general.get('uploaded')),
 483             'episode_number': traverse_obj(
 484                 video, (('episodedata', 'general'), 'episode'), expected_type=int, get_all=False),
 485             'season_number': traverse_obj(
 486                 video, (('episodedata', 'general'), 'season'), expected_type=int, get_all=False),
 487             'cast': traverse_obj(video, ('connectedmedia', ..., 'title'), expected_type=str),
 488             'formats': formats,
 489             'subtitles': subtitles,
 490         }
 491
 492
 493 class NexxEmbedIE(InfoExtractor):
 494     _VALID_URL = r'https?://embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?:video/)?(?P<id>[^/?#&]+)'
 495     # Reference. https://nx-s.akamaized.net/files/201510/44.pdf
 496     _EMBED_REGEX = [r'<iframe[^>]+\bsrc=(["\'])(?P<url>(?:https?:)?//embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?:(?!\1).)+)\1']
 497     _TESTS = [{
 498         'url': 'http://embed.nexx.cloud/748/KC1614647Z27Y7T?autoplay=1',
 499         'md5': '16746bfc28c42049492385c989b26c4a',
 500         'info_dict': {
 501             'id': '161464',
 502             'ext': 'mp4',
 503             'title': 'Nervenkitzel Achterbahn',
 504             'alt_title': 'Karussellbauer in Deutschland',
 505             'description': 'md5:ffe7b1cc59a01f585e0569949aef73cc',
 506             'creator': 'SPIEGEL TV',
 507             'thumbnail': r're:^https?://.*\.jpg$',
 508             'duration': 2761,
 509             'timestamp': 1394021479,
 510             'upload_date': '20140305',
 511         },
 512         'params': {
 513             'skip_download': True,
 514         },
 515     }, {
 516         'url': 'https://embed.nexx.cloud/11888/video/DSRTO7UVOX06S7',
 517         'only_matching': True,
 518     }]
 519
 520     def _real_extract(self, url):
 521         embed_id = self._match_id(url)
 522
 523         webpage = self._download_webpage(url, embed_id)
 524
 525         return self.url_result(NexxIE._extract_url(webpage), ie=NexxIE.ie_key())