yt_dlp/extractor/theplatform.py

   1 import binascii
   2 import hashlib
   3 import hmac
   4 import re
   5 import time
   6
   7 from .adobepass import AdobePassIE
   8 from .once import OnceIE
   9 from ..networking import HEADRequest, Request
  10 from ..utils import (
  11     ExtractorError,
  12     determine_ext,
  13     find_xpath_attr,
  14     float_or_none,
  15     int_or_none,
  16     mimetype2ext,
  17     parse_qs,
  18     traverse_obj,
  19     unsmuggle_url,
  20     update_url,
  21     update_url_query,
  22     urlhandle_detect_ext,
  23     xpath_with_ns,
  24 )
  25
  26 default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  27 _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  28
  29
  30 class ThePlatformBaseIE(OnceIE):
  31     _TP_TLD = 'com'
  32
  33     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  34         meta = self._download_xml(
  35             smil_url, video_id, note=note, query={'format': 'SMIL'},
  36             headers=self.geo_verification_headers())
  37         error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src')
  38         if error_element is not None:
  39             exception = find_xpath_attr(
  40                 error_element, _x('.//smil:param'), 'name', 'exception')
  41             if exception is not None:
  42                 if exception.get('value') == 'GeoLocationBlocked':
  43                     self.raise_geo_restricted(error_element.attrib['abstract'])
  44                 elif error_element.attrib['src'].startswith(
  45                         'http://link.theplatform.%s/s/errorFiles/Unavailable.'
  46                         % self._TP_TLD):
  47                     raise ExtractorError(
  48                         error_element.attrib['abstract'], expected=True)
  49
  50         smil_formats, subtitles = self._parse_smil_formats_and_subtitles(
  51             meta, smil_url, video_id, namespace=default_ns,
  52             # the parameters are from syfy.com, other sites may use others,
  53             # they also work for nbc.com
  54             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  55             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  56
  57         formats = []
  58         for _format in smil_formats:
  59             if OnceIE.suitable(_format['url']):
  60                 formats.extend(self._extract_once_formats(_format['url']))
  61             else:
  62                 media_url = _format['url']
  63                 if determine_ext(media_url) == 'm3u8':
  64                     hdnea2 = self._get_cookies(media_url).get('hdnea2')
  65                     if hdnea2:
  66                         _format['url'] = update_url_query(media_url, {'hdnea3': hdnea2.value})
  67
  68                 formats.append(_format)
  69
  70         return formats, subtitles
  71
  72     def _download_theplatform_metadata(self, path, video_id):
  73         info_url = 'http://link.theplatform.%s/s/%s?format=preview' % (self._TP_TLD, path)
  74         return self._download_json(info_url, video_id)
  75
  76     def _parse_theplatform_metadata(self, info):
  77         subtitles = {}
  78         captions = info.get('captions')
  79         if isinstance(captions, list):
  80             for caption in captions:
  81                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  82                 subtitles.setdefault(lang, []).append({
  83                     'ext': mimetype2ext(mime),
  84                     'url': src,
  85                 })
  86
  87         duration = info.get('duration')
  88         tp_chapters = info.get('chapters', [])
  89         chapters = []
  90         if tp_chapters:
  91             def _add_chapter(start_time, end_time):
  92                 start_time = float_or_none(start_time, 1000)
  93                 end_time = float_or_none(end_time, 1000)
  94                 if start_time is None or end_time is None:
  95                     return
  96                 chapters.append({
  97                     'start_time': start_time,
  98                     'end_time': end_time,
  99                 })
 100
 101             for chapter in tp_chapters[:-1]:
 102                 _add_chapter(chapter.get('startTime'), chapter.get('endTime'))
 103             _add_chapter(tp_chapters[-1].get('startTime'), tp_chapters[-1].get('endTime') or duration)
 104
 105         def extract_site_specific_field(field):
 106             # A number of sites have custom-prefixed keys, e.g. 'cbc$seasonNumber'
 107             return traverse_obj(info, lambda k, v: v and k.endswith(f'${field}'), get_all=False)
 108
 109         return {
 110             'title': info['title'],
 111             'subtitles': subtitles,
 112             'description': info['description'],
 113             'thumbnail': info['defaultThumbnailUrl'],
 114             'duration': float_or_none(duration, 1000),
 115             'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
 116             'uploader': info.get('billingCode'),
 117             'chapters': chapters,
 118             'creator': traverse_obj(info, ('author', {str})) or None,
 119             'categories': traverse_obj(info, (
 120                 'categories', lambda _, v: v.get('label') in ('category', None), 'name', {str})) or None,
 121             'tags': traverse_obj(info, ('keywords', {lambda x: re.split(r'[;,]\s?', x) if x else None})),
 122             'location': extract_site_specific_field('region'),
 123             'series': extract_site_specific_field('show'),
 124             'season_number': int_or_none(extract_site_specific_field('seasonNumber')),
 125             'media_type': extract_site_specific_field('programmingType') or extract_site_specific_field('type'),
 126         }
 127
 128     def _extract_theplatform_metadata(self, path, video_id):
 129         info = self._download_theplatform_metadata(path, video_id)
 130         return self._parse_theplatform_metadata(info)
 131
 132
 133 class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
 134     _VALID_URL = r'''(?x)
 135         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
 136            (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)?|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
 137          |theplatform:)(?P<id>[^/\?&]+)'''
 138     _EMBED_REGEX = [
 139         r'''(?x)
 140             <meta\s+
 141                 property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
 142                 content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2''',
 143         r'(?s)<(?:iframe|script)[^>]+src=(["\'])(?P<url>(?:https?:)?//player\.theplatform\.com/p/.+?)\1'
 144     ]
 145
 146     _TESTS = [{
 147         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
 148         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
 149         'info_dict': {
 150             'id': 'e9I_cZgTgIPd',
 151             'ext': 'flv',
 152             'title': 'Blackberry\'s big, bold Z30',
 153             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
 154             'duration': 247,
 155             'timestamp': 1383239700,
 156             'upload_date': '20131031',
 157             'uploader': 'CBSI-NEW',
 158         },
 159         'params': {
 160             # rtmp download
 161             'skip_download': True,
 162         },
 163         'skip': '404 Not Found',
 164     }, {
 165         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 166         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 167         'info_dict': {
 168             'id': '22d_qsQ6MIRT',
 169             'ext': 'flv',
 170             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 171             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 172             'timestamp': 1426176191,
 173             'upload_date': '20150312',
 174             'uploader': 'CBSI-NEW',
 175         },
 176         'params': {
 177             # rtmp download
 178             'skip_download': True,
 179         },
 180         'skip': 'CNet no longer uses ThePlatform',
 181     }, {
 182         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 183         'info_dict': {
 184             'id': 'yMBg9E8KFxZD',
 185             'ext': 'mp4',
 186             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 187             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 188             'uploader': 'EGSM',
 189         },
 190         'skip': 'Dead link',
 191     }, {
 192         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 193         'only_matching': True,
 194     }, {
 195         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 196         'md5': 'fb96bb3d85118930a5b055783a3bd992',
 197         'info_dict': {
 198             'id': 'tdy_or_siri_150701',
 199             'ext': 'mp4',
 200             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 201             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 202             'duration': 83.0,
 203             'thumbnail': r're:^https?://.*\.jpg$',
 204             'timestamp': 1435752600,
 205             'upload_date': '20150701',
 206             'uploader': 'NBCU-NEWS',
 207         },
 208         'skip': 'Error: Player PID "nbcNewsOffsite" is disabled',
 209     }, {
 210         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 211         # geo-restricted (US), HLS encrypted with AES-128
 212         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 213         'only_matching': True,
 214     }]
 215
 216     @classmethod
 217     def _extract_embed_urls(cls, url, webpage):
 218         # Are whitespaces ignored in URLs?
 219         # https://github.com/ytdl-org/youtube-dl/issues/12044
 220         for embed_url in super()._extract_embed_urls(url, webpage):
 221             yield re.sub(r'\s', '', embed_url)
 222
 223     @staticmethod
 224     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 225         flags = '10' if include_qs else '00'
 226         expiration_date = '%x' % (int(time.time()) + life)
 227
 228         def str_to_hex(str):
 229             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 230
 231         def hex_to_bytes(hex):
 232             return binascii.a2b_hex(hex.encode('ascii'))
 233
 234         relative_path = re.match(r'https?://link\.theplatform\.com/s/([^?]+)', url).group(1)
 235         clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path))
 236         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 237         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 238         return '%s&sig=%s' % (url, sig)
 239
 240     def _real_extract(self, url):
 241         url, smuggled_data = unsmuggle_url(url, {})
 242         self._initialize_geo_bypass({
 243             'countries': smuggled_data.get('geo_countries'),
 244         })
 245
 246         mobj = self._match_valid_url(url)
 247         provider_id = mobj.group('provider_id')
 248         video_id = mobj.group('id')
 249
 250         if not provider_id:
 251             provider_id = 'dJ5BDC'
 252
 253         path = provider_id + '/'
 254         if mobj.group('media'):
 255             path += mobj.group('media')
 256         path += video_id
 257
 258         qs_dict = parse_qs(url)
 259         if 'guid' in qs_dict:
 260             webpage = self._download_webpage(url, video_id)
 261             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 262             feed_id = None
 263             # feed id usually locates in the last script.
 264             # Seems there's no pattern for the interested script filename, so
 265             # I try one by one
 266             for script in reversed(scripts):
 267                 feed_script = self._download_webpage(
 268                     self._proto_relative_url(script, 'http:'),
 269                     video_id, 'Downloading feed script')
 270                 feed_id = self._search_regex(
 271                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 272                     'default feed id', default=None)
 273                 if feed_id is not None:
 274                     break
 275             if feed_id is None:
 276                 raise ExtractorError('Unable to find feed id')
 277             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 278                 provider_id, feed_id, qs_dict['guid'][0]))
 279
 280         if smuggled_data.get('force_smil_url', False):
 281             smil_url = url
 282         # Explicitly specified SMIL (see https://github.com/ytdl-org/youtube-dl/issues/7385)
 283         elif '/guid/' in url:
 284             headers = {}
 285             source_url = smuggled_data.get('source_url')
 286             if source_url:
 287                 headers['Referer'] = source_url
 288             request = Request(url, headers=headers)
 289             webpage = self._download_webpage(request, video_id)
 290             smil_url = self._search_regex(
 291                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 292                 webpage, 'smil url', group='url')
 293             path = self._search_regex(
 294                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 295             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4'
 296         elif mobj.group('config'):
 297             config_url = url + '&form=json'
 298             config_url = config_url.replace('swf/', 'config/')
 299             config_url = config_url.replace('onsite/', 'onsite/config/')
 300             config = self._download_json(config_url, video_id, 'Downloading config')
 301             if 'releaseUrl' in config:
 302                 release_url = config['releaseUrl']
 303             else:
 304                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 305             smil_url = release_url + '&formats=MPEG4&manifest=f4m'
 306         else:
 307             smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 308
 309         sig = smuggled_data.get('sig')
 310         if sig:
 311             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 312
 313         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 314
 315         # With some sites, manifest URL must be forced to extract HLS formats
 316         if not traverse_obj(formats, lambda _, v: v['format_id'].startswith('hls')):
 317             m3u8_url = update_url(url, query='mbr=true&manifest=m3u', fragment=None)
 318             urlh = self._request_webpage(
 319                 HEADRequest(m3u8_url), video_id, 'Checking for HLS formats', 'No HLS formats found', fatal=False)
 320             if urlh and urlhandle_detect_ext(urlh) == 'm3u8':
 321                 m3u8_fmts, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
 322                     m3u8_url, video_id, m3u8_id='hls', fatal=False)
 323                 formats.extend(m3u8_fmts)
 324                 self._merge_subtitles(m3u8_subs, target=subtitles)
 325
 326         ret = self._extract_theplatform_metadata(path, video_id)
 327         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 328         ret.update({
 329             'id': video_id,
 330             'formats': formats,
 331             'subtitles': combined_subtitles,
 332         })
 333
 334         return ret
 335
 336
 337 class ThePlatformFeedIE(ThePlatformBaseIE):
 338     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
 339     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[^&]+))'
 340     _TESTS = [{
 341         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 342         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 343         'md5': '6e32495b5073ab414471b615c5ded394',
 344         'info_dict': {
 345             'id': 'n_hardball_5biden_140207',
 346             'ext': 'mp4',
 347             'title': 'The Biden factor: will Joe run in 2016?',
 348             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 349             'thumbnail': r're:^https?://.*\.jpg$',
 350             'upload_date': '20140208',
 351             'timestamp': 1391824260,
 352             'duration': 467.0,
 353             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 354             'uploader': 'NBCU-NEWS',
 355         },
 356     }, {
 357         'url': 'http://feed.theplatform.com/f/2E2eJC/nnd_NBCNews?byGuid=nn_netcast_180306.Copy.01',
 358         'only_matching': True,
 359     }]
 360
 361     def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}, account_id=None):
 362         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
 363         entry = self._download_json(real_url, video_id)['entries'][0]
 364         main_smil_url = 'http://link.theplatform.com/s/%s/media/guid/%d/%s' % (provider_id, account_id, entry['guid']) if account_id else entry.get('plmedia$publicUrl')
 365
 366         formats = []
 367         subtitles = {}
 368         first_video_id = None
 369         duration = None
 370         asset_types = []
 371         for item in entry['media$content']:
 372             smil_url = item['plfile$url']
 373             cur_video_id = ThePlatformIE._match_id(smil_url)
 374             if first_video_id is None:
 375                 first_video_id = cur_video_id
 376                 duration = float_or_none(item.get('plfile$duration'))
 377             file_asset_types = item.get('plfile$assetTypes') or parse_qs(smil_url)['assetTypes']
 378             for asset_type in file_asset_types:
 379                 if asset_type in asset_types:
 380                     continue
 381                 asset_types.append(asset_type)
 382                 query = {
 383                     'mbr': 'true',
 384                     'formats': item['plfile$format'],
 385                     'assetTypes': asset_type,
 386                 }
 387                 if asset_type in asset_types_query:
 388                     query.update(asset_types_query[asset_type])
 389                 cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
 390                     main_smil_url or smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
 391                 formats.extend(cur_formats)
 392                 subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 393
 394         thumbnails = [{
 395             'url': thumbnail['plfile$url'],
 396             'width': int_or_none(thumbnail.get('plfile$width')),
 397             'height': int_or_none(thumbnail.get('plfile$height')),
 398         } for thumbnail in entry.get('media$thumbnails', [])]
 399
 400         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 401         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 402
 403         ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 404         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 405         ret.update({
 406             'id': video_id,
 407             'formats': formats,
 408             'subtitles': subtitles,
 409             'thumbnails': thumbnails,
 410             'duration': duration,
 411             'timestamp': timestamp,
 412             'categories': categories,
 413         })
 414         if custom_fields:
 415             ret.update(custom_fields(entry))
 416
 417         return ret
 418
 419     def _real_extract(self, url):
 420         mobj = self._match_valid_url(url)
 421
 422         video_id = mobj.group('id')
 423         provider_id = mobj.group('provider_id')
 424         feed_id = mobj.group('feed_id')
 425         filter_query = mobj.group('filter')
 426
 427         return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)