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     @staticmethod
 118     def _extract_urls(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     @staticmethod
 139     def _extract_url(webpage):
 140         return NexxIE._extract_urls(webpage)[0]
 141
 142     def _handle_error(self, response):
 143         if traverse_obj(response, ('metadata', 'notice'), expected_type=str):
 144             self.report_warning('%s said: %s' % (self.IE_NAME, response['metadata']['notice']))
 145         status = int_or_none(try_get(
 146             response, lambda x: x['metadata']['status']) or 200)
 147         if 200 <= status < 300:
 148             return
 149         raise ExtractorError(
 150             '%s said: %s' % (self.IE_NAME, response['metadata']['errorhint']),
 151             expected=True)
 152
 153     def _call_api(self, domain_id, path, video_id, data=None, headers={}):
 154         headers['Content-Type'] = 'application/x-www-form-urlencoded; charset=UTF-8'
 155         result = self._download_json(
 156             'https://api.nexx.cloud/v3/%s/%s' % (domain_id, path), video_id,
 157             'Downloading %s JSON' % path, data=urlencode_postdata(data),
 158             headers=headers)
 159         self._handle_error(result)
 160         return result['result']
 161
 162     def _extract_free_formats(self, video, video_id):
 163         stream_data = video['streamdata']
 164         cdn = stream_data['cdnType']
 165         assert cdn == 'free'
 166
 167         hash = video['general']['hash']
 168
 169         ps = compat_str(stream_data['originalDomain'])
 170         if stream_data['applyFolderHierarchy'] == 1:
 171             s = ('%04d' % int(video_id))[::-1]
 172             ps += '/%s/%s' % (s[0:2], s[2:4])
 173         ps += '/%s/%s_' % (video_id, hash)
 174
 175         t = 'http://%s' + ps
 176         fd = stream_data['azureFileDistribution'].split(',')
 177         cdn_provider = stream_data['cdnProvider']
 178
 179         def p0(p):
 180             return '_%s' % p if stream_data['applyAzureStructure'] == 1 else ''
 181
 182         formats = []
 183         if cdn_provider == 'ak':
 184             t += ','
 185             for i in fd:
 186                 p = i.split(':')
 187                 t += p[1] + p0(int(p[0])) + ','
 188             t += '.mp4.csmil/master.%s'
 189         elif cdn_provider == 'ce':
 190             k = t.split('/')
 191             h = k.pop()
 192             http_base = t = '/'.join(k)
 193             http_base = http_base % stream_data['cdnPathHTTP']
 194             t += '/asset.ism/manifest.%s?dcp_ver=aos4&videostream='
 195             for i in fd:
 196                 p = i.split(':')
 197                 tbr = int(p[0])
 198                 filename = '%s%s%s.mp4' % (h, p[1], p0(tbr))
 199                 f = {
 200                     'url': http_base + '/' + filename,
 201                     'format_id': '%s-http-%d' % (cdn, tbr),
 202                     'tbr': tbr,
 203                 }
 204                 width_height = p[1].split('x')
 205                 if len(width_height) == 2:
 206                     f.update({
 207                         'width': int_or_none(width_height[0]),
 208                         'height': int_or_none(width_height[1]),
 209                     })
 210                 formats.append(f)
 211                 a = filename + ':%s' % (tbr * 1000)
 212                 t += a + ','
 213             t = t[:-1] + '&audiostream=' + a.split(':')[0]
 214         else:
 215             assert False
 216
 217         if cdn_provider == 'ce':
 218             formats.extend(self._extract_mpd_formats(
 219                 t % (stream_data['cdnPathDASH'], 'mpd'), video_id,
 220                 mpd_id='%s-dash' % cdn, fatal=False))
 221         formats.extend(self._extract_m3u8_formats(
 222             t % (stream_data['cdnPathHLS'], 'm3u8'), video_id, 'mp4',
 223             entry_protocol='m3u8_native', m3u8_id='%s-hls' % cdn, fatal=False))
 224
 225         return formats
 226
 227     def _extract_3q_formats(self, video, video_id):
 228         stream_data = video['streamdata']
 229         cdn = stream_data['cdnType']
 230         assert cdn == '3q'
 231
 232         q_acc, q_prefix, q_locator, q_hash = stream_data['qAccount'], stream_data['qPrefix'], stream_data['qLocator'], stream_data['qHash']
 233         protection_key = traverse_obj(
 234             video, ('protectiondata', 'key'), expected_type=str)
 235
 236         def get_cdn_shield_base(shield_type=''):
 237             for secure in ('', 's'):
 238                 cdn_shield = stream_data.get('cdnShield%sHTTP%s' % (shield_type, secure.upper()))
 239                 if cdn_shield:
 240                     return 'http%s://%s' % (secure, cdn_shield)
 241             return f'http://sdn-global-{"prog" if shield_type.lower() == "prog" else "streaming"}-cache.3qsdn.com/' + (f's/{protection_key}/' if protection_key else '')
 242
 243         stream_base = get_cdn_shield_base()
 244
 245         formats = []
 246         formats.extend(self._extract_m3u8_formats(
 247             f'{stream_base}{q_acc}/files/{q_prefix}/{q_locator}/{q_acc}-{stream_data.get("qHEVCHash") or q_hash}.ism/manifest.m3u8',
 248             video_id, 'mp4', m3u8_id=f'{cdn}-hls', fatal=False))
 249         formats.extend(self._extract_mpd_formats(
 250             f'{stream_base}{q_acc}/files/{q_prefix}/{q_locator}/{q_acc}-{q_hash}.ism/manifest.mpd',
 251             video_id, mpd_id=f'{cdn}-dash', fatal=False))
 252
 253         progressive_base = get_cdn_shield_base('Prog')
 254         q_references = stream_data.get('qReferences') or ''
 255         fds = q_references.split(',')
 256         for fd in fds:
 257             ss = fd.split(':')
 258             if len(ss) != 3:
 259                 continue
 260             tbr = int_or_none(ss[1], scale=1000)
 261             formats.append({
 262                 'url': f'{progressive_base}{q_acc}/uploads/{q_acc}-{ss[2]}.webm',
 263                 'format_id': f'{cdn}-{ss[0]}{"-%s" % tbr if tbr else ""}',
 264                 'tbr': tbr,
 265             })
 266
 267         azure_file_distribution = stream_data.get('azureFileDistribution') or ''
 268         fds = azure_file_distribution.split(',')
 269         for fd in fds:
 270             ss = fd.split(':')
 271             if len(ss) != 3:
 272                 continue
 273             tbr = int_or_none(ss[0])
 274             width, height = ss[1].split('x') if len(ss[1].split('x')) == 2 else (None, None)
 275             f = {
 276                 'url': f'{progressive_base}{q_acc}/files/{q_prefix}/{q_locator}/{ss[2]}.mp4',
 277                 'format_id': f'{cdn}-http-{"-%s" % tbr if tbr else ""}',
 278                 'tbr': tbr,
 279                 'width': int_or_none(width),
 280                 'height': int_or_none(height),
 281             }
 282             formats.append(f)
 283
 284         return formats
 285
 286     def _extract_azure_formats(self, video, video_id):
 287         stream_data = video['streamdata']
 288         cdn = stream_data['cdnType']
 289         assert cdn == 'azure'
 290
 291         azure_locator = stream_data['azureLocator']
 292
 293         def get_cdn_shield_base(shield_type='', static=False):
 294             for secure in ('', 's'):
 295                 cdn_shield = stream_data.get('cdnShield%sHTTP%s' % (shield_type, secure.upper()))
 296                 if cdn_shield:
 297                     return 'http%s://%s' % (secure, cdn_shield)
 298             else:
 299                 if 'fb' in stream_data['azureAccount']:
 300                     prefix = 'df' if static else 'f'
 301                 else:
 302                     prefix = 'd' if static else 'p'
 303                 account = int(stream_data['azureAccount'].replace('nexxplayplus', '').replace('nexxplayfb', ''))
 304                 return 'http://nx-%s%02d.akamaized.net/' % (prefix, account)
 305
 306         language = video['general'].get('language_raw') or ''
 307
 308         azure_stream_base = get_cdn_shield_base()
 309         is_ml = ',' in language
 310         azure_manifest_url = '%s%s/%s_src%s.ism/Manifest' % (
 311             azure_stream_base, azure_locator, video_id, ('_manifest' if is_ml else '')) + '%s'
 312
 313         protection_token = try_get(
 314             video, lambda x: x['protectiondata']['token'], compat_str)
 315         if protection_token:
 316             azure_manifest_url += '?hdnts=%s' % protection_token
 317
 318         formats = self._extract_m3u8_formats(
 319             azure_manifest_url % '(format=m3u8-aapl)',
 320             video_id, 'mp4', 'm3u8_native',
 321             m3u8_id='%s-hls' % cdn, fatal=False)
 322         formats.extend(self._extract_mpd_formats(
 323             azure_manifest_url % '(format=mpd-time-csf)',
 324             video_id, mpd_id='%s-dash' % cdn, fatal=False))
 325         formats.extend(self._extract_ism_formats(
 326             azure_manifest_url % '', video_id, ism_id='%s-mss' % cdn, fatal=False))
 327
 328         azure_progressive_base = get_cdn_shield_base('Prog', True)
 329         azure_file_distribution = stream_data.get('azureFileDistribution')
 330         if azure_file_distribution:
 331             fds = azure_file_distribution.split(',')
 332             if fds:
 333                 for fd in fds:
 334                     ss = fd.split(':')
 335                     if len(ss) == 2:
 336                         tbr = int_or_none(ss[0])
 337                         if tbr:
 338                             f = {
 339                                 'url': '%s%s/%s_src_%s_%d.mp4' % (
 340                                     azure_progressive_base, azure_locator, video_id, ss[1], tbr),
 341                                 'format_id': '%s-http-%d' % (cdn, tbr),
 342                                 'tbr': tbr,
 343                             }
 344                             width_height = ss[1].split('x')
 345                             if len(width_height) == 2:
 346                                 f.update({
 347                                     'width': int_or_none(width_height[0]),
 348                                     'height': int_or_none(width_height[1]),
 349                                 })
 350                             formats.append(f)
 351
 352         return formats
 353
 354     def _real_extract(self, url):
 355         mobj = self._match_valid_url(url)
 356         domain_id = mobj.group('domain_id') or mobj.group('domain_id_s')
 357         video_id = mobj.group('id')
 358
 359         video = None
 360
 361         def find_video(result):
 362             if isinstance(result, dict):
 363                 return result
 364             elif isinstance(result, list):
 365                 vid = int(video_id)
 366                 for v in result:
 367                     if try_get(v, lambda x: x['general']['ID'], int) == vid:
 368                         return v
 369             return None
 370
 371         response = self._download_json(
 372             'https://arc.nexx.cloud/api/video/%s.json' % video_id,
 373             video_id, fatal=False)
 374         if response and isinstance(response, dict):
 375             result = response.get('result')
 376             if result:
 377                 video = find_video(result)
 378
 379         # not all videos work via arc, e.g. nexx:741:1269984
 380         if not video:
 381             # Reverse engineered from JS code (see getDeviceID function)
 382             device_id = '%d:%d:%d%d' % (
 383                 random.randint(1, 4), int(time.time()),
 384                 random.randint(1e4, 99999), random.randint(1, 9))
 385
 386             result = self._call_api(domain_id, 'session/init', video_id, data={
 387                 'nxp_devh': device_id,
 388                 'nxp_userh': '',
 389                 'precid': '0',
 390                 'playlicense': '0',
 391                 'screenx': '1920',
 392                 'screeny': '1080',
 393                 'playerversion': '6.0.00',
 394                 'gateway': 'html5',
 395                 'adGateway': '',
 396                 'explicitlanguage': 'en-US',
 397                 'addTextTemplates': '1',
 398                 'addDomainData': '1',
 399                 'addAdModel': '1',
 400             }, headers={
 401                 'X-Request-Enable-Auth-Fallback': '1',
 402             })
 403
 404             cid = result['general']['cid']
 405
 406             # As described in [1] X-Request-Token generation algorithm is
 407             # as follows:
 408             #   md5( operation + domain_id + domain_secret )
 409             # where domain_secret is a static value that will be given by nexx.tv
 410             # as per [1]. Here is how this "secret" is generated (reversed
 411             # from _play._factory.data.getDomainData function, search for
 412             # domaintoken or enableAPIAccess). So it's actually not static
 413             # and not that much of a secret.
 414             # 1. https://nexxtvstorage.blob.core.windows.net/files/201610/27.pdf
 415             secret = result['device']['domaintoken'][int(device_id[0]):]
 416             secret = secret[0:len(secret) - int(device_id[-1])]
 417
 418             op = 'byid'
 419
 420             # Reversed from JS code for _play.api.call function (search for
 421             # X-Request-Token)
 422             request_token = hashlib.md5(
 423                 ''.join((op, domain_id, secret)).encode('utf-8')).hexdigest()
 424
 425             result = self._call_api(
 426                 domain_id, 'videos/%s/%s' % (op, video_id), video_id, data={
 427                     'additionalfields': 'language,channel,format,licenseby,slug,fileversion,episode,season',
 428                     'addInteractionOptions': '1',
 429                     'addStatusDetails': '1',
 430                     'addStreamDetails': '1',
 431                     'addFeatures': '1',
 432                     # Caption format selection doesn't seem to be enforced?
 433                     'addCaptions': 'vtt',
 434                     'addScenes': '1',
 435                     'addChapters': '1',
 436                     'addHotSpots': '1',
 437                     'addConnectedMedia': 'persons',
 438                     'addBumpers': '1',
 439                 }, headers={
 440                     'X-Request-CID': cid,
 441                     'X-Request-Token': request_token,
 442                 })
 443             video = find_video(result)
 444
 445         general = video['general']
 446         title = general['title']
 447
 448         cdn = video['streamdata']['cdnType']
 449
 450         if cdn == 'azure':
 451             formats = self._extract_azure_formats(video, video_id)
 452         elif cdn == 'free':
 453             formats = self._extract_free_formats(video, video_id)
 454         elif cdn == '3q':
 455             formats = self._extract_3q_formats(video, video_id)
 456         else:
 457             self.raise_no_formats(f'{cdn} formats are currently not supported', video_id)
 458
 459         self._sort_formats(formats)
 460
 461         subtitles = {}
 462         for sub in video.get('captiondata') or []:
 463             if sub.get('data'):
 464                 subtitles.setdefault(sub.get('language', 'en'), []).append({
 465                     'ext': 'srt',
 466                     'data': '\n\n'.join(
 467                         f'{i + 1}\n{srt_subtitles_timecode(line["fromms"] / 1000)} --> {srt_subtitles_timecode(line["toms"] / 1000)}\n{line["caption"]}'
 468                         for i, line in enumerate(sub['data'])),
 469                     'name': sub.get('language_long') or sub.get('title')
 470                 })
 471             elif sub.get('url'):
 472                 subtitles.setdefault(sub.get('language', 'en'), []).append({
 473                     'url': sub['url'],
 474                     'ext': sub.get('format'),
 475                     'name': sub.get('language_long') or sub.get('title')
 476                 })
 477
 478         return {
 479             'id': video_id,
 480             'title': title,
 481             'alt_title': general.get('subtitle'),
 482             'description': general.get('description'),
 483             'release_year': int_or_none(general.get('year')),
 484             'creator': general.get('studio') or general.get('studio_adref') or None,
 485             'thumbnail': try_get(
 486                 video, lambda x: x['imagedata']['thumb'], compat_str),
 487             'duration': parse_duration(general.get('runtime')),
 488             'timestamp': int_or_none(general.get('uploaded')),
 489             'episode_number': traverse_obj(
 490                 video, (('episodedata', 'general'), 'episode'), expected_type=int, get_all=False),
 491             'season_number': traverse_obj(
 492                 video, (('episodedata', 'general'), 'season'), expected_type=int, get_all=False),
 493             'cast': traverse_obj(video, ('connectedmedia', ..., 'title'), expected_type=str),
 494             'formats': formats,
 495             'subtitles': subtitles,
 496         }
 497
 498
 499 class NexxEmbedIE(InfoExtractor):
 500     _VALID_URL = r'https?://embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?:video/)?(?P<id>[^/?#&]+)'
 501     _TESTS = [{
 502         'url': 'http://embed.nexx.cloud/748/KC1614647Z27Y7T?autoplay=1',
 503         'md5': '16746bfc28c42049492385c989b26c4a',
 504         'info_dict': {
 505             'id': '161464',
 506             'ext': 'mp4',
 507             'title': 'Nervenkitzel Achterbahn',
 508             'alt_title': 'Karussellbauer in Deutschland',
 509             'description': 'md5:ffe7b1cc59a01f585e0569949aef73cc',
 510             'creator': 'SPIEGEL TV',
 511             'thumbnail': r're:^https?://.*\.jpg$',
 512             'duration': 2761,
 513             'timestamp': 1394021479,
 514             'upload_date': '20140305',
 515         },
 516         'params': {
 517             'skip_download': True,
 518         },
 519     }, {
 520         'url': 'https://embed.nexx.cloud/11888/video/DSRTO7UVOX06S7',
 521         'only_matching': True,
 522     }]
 523
 524     @staticmethod
 525     def _extract_urls(webpage):
 526         # Reference:
 527         # 1. https://nx-s.akamaized.net/files/201510/44.pdf
 528
 529         # iFrame Embed Integration
 530         return [mobj.group('url') for mobj in re.finditer(
 531             r'<iframe[^>]+\bsrc=(["\'])(?P<url>(?:https?:)?//embed\.nexx(?:\.cloud|cdn\.com)/\d+/(?:(?!\1).)+)\1',
 532             webpage)]
 533
 534     def _real_extract(self, url):
 535         embed_id = self._match_id(url)
 536
 537         webpage = self._download_webpage(url, embed_id)
 538
 539         return self.url_result(NexxIE._extract_url(webpage), ie=NexxIE.ie_key())