yt_dlp/extractor/theplatform.py

   1 import re
   2 import time
   3 import hmac
   4 import binascii
   5 import hashlib
   6
   7
   8 from .once import OnceIE
   9 from .adobepass import AdobePassIE
  10 from ..networking import Request
  11 from ..utils import (
  12     determine_ext,
  13     ExtractorError,
  14     float_or_none,
  15     int_or_none,
  16     parse_qs,
  17     unsmuggle_url,
  18     update_url_query,
  19     xpath_with_ns,
  20     mimetype2ext,
  21     find_xpath_attr,
  22     traverse_obj,
  23     update_url,
  24     urlhandle_detect_ext,
  25 )
  26 from ..networking import HEADRequest
  27
  28 default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  29 _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  30
  31
  32 class ThePlatformBaseIE(OnceIE):
  33     _TP_TLD = 'com'
  34
  35     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  36         meta = self._download_xml(
  37             smil_url, video_id, note=note, query={'format': 'SMIL'},
  38             headers=self.geo_verification_headers())
  39         error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src')
  40         if error_element is not None:
  41             exception = find_xpath_attr(
  42                 error_element, _x('.//smil:param'), 'name', 'exception')
  43             if exception is not None:
  44                 if exception.get('value') == 'GeoLocationBlocked':
  45                     self.raise_geo_restricted(error_element.attrib['abstract'])
  46                 elif error_element.attrib['src'].startswith(
  47                         'http://link.theplatform.%s/s/errorFiles/Unavailable.'
  48                         % self._TP_TLD):
  49                     raise ExtractorError(
  50                         error_element.attrib['abstract'], expected=True)
  51
  52         smil_formats, subtitles = self._parse_smil_formats_and_subtitles(
  53             meta, smil_url, video_id, namespace=default_ns,
  54             # the parameters are from syfy.com, other sites may use others,
  55             # they also work for nbc.com
  56             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  57             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  58
  59         formats = []
  60         for _format in smil_formats:
  61             if OnceIE.suitable(_format['url']):
  62                 formats.extend(self._extract_once_formats(_format['url']))
  63             else:
  64                 media_url = _format['url']
  65                 if determine_ext(media_url) == 'm3u8':
  66                     hdnea2 = self._get_cookies(media_url).get('hdnea2')
  67                     if hdnea2:
  68                         _format['url'] = update_url_query(media_url, {'hdnea3': hdnea2.value})
  69
  70                 formats.append(_format)
  71
  72         return formats, subtitles
  73
  74     def _download_theplatform_metadata(self, path, video_id):
  75         info_url = 'http://link.theplatform.%s/s/%s?format=preview' % (self._TP_TLD, path)
  76         return self._download_json(info_url, video_id)
  77
  78     def _parse_theplatform_metadata(self, info):
  79         subtitles = {}
  80         captions = info.get('captions')
  81         if isinstance(captions, list):
  82             for caption in captions:
  83                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  84                 subtitles.setdefault(lang, []).append({
  85                     'ext': mimetype2ext(mime),
  86                     'url': src,
  87                 })
  88
  89         duration = info.get('duration')
  90         tp_chapters = info.get('chapters', [])
  91         chapters = []
  92         if tp_chapters:
  93             def _add_chapter(start_time, end_time):
  94                 start_time = float_or_none(start_time, 1000)
  95                 end_time = float_or_none(end_time, 1000)
  96                 if start_time is None or end_time is None:
  97                     return
  98                 chapters.append({
  99                     'start_time': start_time,
 100                     'end_time': end_time,
 101                 })
 102
 103             for chapter in tp_chapters[:-1]:
 104                 _add_chapter(chapter.get('startTime'), chapter.get('endTime'))
 105             _add_chapter(tp_chapters[-1].get('startTime'), tp_chapters[-1].get('endTime') or duration)
 106
 107         def extract_site_specific_field(field):
 108             # A number of sites have custom-prefixed keys, e.g. 'cbc$seasonNumber'
 109             return traverse_obj(info, lambda k, v: v and k.endswith(f'${field}'), get_all=False)
 110
 111         return {
 112             'title': info['title'],
 113             'subtitles': subtitles,
 114             'description': info['description'],
 115             'thumbnail': info['defaultThumbnailUrl'],
 116             'duration': float_or_none(duration, 1000),
 117             'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
 118             'uploader': info.get('billingCode'),
 119             'chapters': chapters,
 120             'creator': traverse_obj(info, ('author', {str})) or None,
 121             'categories': traverse_obj(info, (
 122                 'categories', lambda _, v: v.get('label') in ('category', None), 'name', {str})) or None,
 123             'tags': traverse_obj(info, ('keywords', {lambda x: re.split(r'[;,]\s?', x) if x else None})),
 124             'location': extract_site_specific_field('region'),
 125             'series': extract_site_specific_field('show'),
 126             'season_number': int_or_none(extract_site_specific_field('seasonNumber')),
 127             'media_type': extract_site_specific_field('programmingType') or extract_site_specific_field('type'),
 128         }
 129
 130     def _extract_theplatform_metadata(self, path, video_id):
 131         info = self._download_theplatform_metadata(path, video_id)
 132         return self._parse_theplatform_metadata(info)
 133
 134
 135 class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
 136     _VALID_URL = r'''(?x)
 137         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
 138            (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)?|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
 139          |theplatform:)(?P<id>[^/\?&]+)'''
 140     _EMBED_REGEX = [
 141         r'''(?x)
 142             <meta\s+
 143                 property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
 144                 content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2''',
 145         r'(?s)<(?:iframe|script)[^>]+src=(["\'])(?P<url>(?:https?:)?//player\.theplatform\.com/p/.+?)\1'
 146     ]
 147
 148     _TESTS = [{
 149         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
 150         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
 151         'info_dict': {
 152             'id': 'e9I_cZgTgIPd',
 153             'ext': 'flv',
 154             'title': 'Blackberry\'s big, bold Z30',
 155             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
 156             'duration': 247,
 157             'timestamp': 1383239700,
 158             'upload_date': '20131031',
 159             'uploader': 'CBSI-NEW',
 160         },
 161         'params': {
 162             # rtmp download
 163             'skip_download': True,
 164         },
 165         'skip': '404 Not Found',
 166     }, {
 167         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 168         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 169         'info_dict': {
 170             'id': '22d_qsQ6MIRT',
 171             'ext': 'flv',
 172             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 173             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 174             'timestamp': 1426176191,
 175             'upload_date': '20150312',
 176             'uploader': 'CBSI-NEW',
 177         },
 178         'params': {
 179             # rtmp download
 180             'skip_download': True,
 181         },
 182         'skip': 'CNet no longer uses ThePlatform',
 183     }, {
 184         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 185         'info_dict': {
 186             'id': 'yMBg9E8KFxZD',
 187             'ext': 'mp4',
 188             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 189             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 190             'uploader': 'EGSM',
 191         },
 192         'skip': 'Dead link',
 193     }, {
 194         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 195         'only_matching': True,
 196     }, {
 197         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 198         'md5': 'fb96bb3d85118930a5b055783a3bd992',
 199         'info_dict': {
 200             'id': 'tdy_or_siri_150701',
 201             'ext': 'mp4',
 202             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 203             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 204             'duration': 83.0,
 205             'thumbnail': r're:^https?://.*\.jpg$',
 206             'timestamp': 1435752600,
 207             'upload_date': '20150701',
 208             'uploader': 'NBCU-NEWS',
 209         },
 210         'skip': 'Error: Player PID "nbcNewsOffsite" is disabled',
 211     }, {
 212         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 213         # geo-restricted (US), HLS encrypted with AES-128
 214         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 215         'only_matching': True,
 216     }]
 217
 218     @classmethod
 219     def _extract_embed_urls(cls, url, webpage):
 220         # Are whitespaces ignored in URLs?
 221         # https://github.com/ytdl-org/youtube-dl/issues/12044
 222         for embed_url in super()._extract_embed_urls(url, webpage):
 223             yield re.sub(r'\s', '', embed_url)
 224
 225     @staticmethod
 226     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 227         flags = '10' if include_qs else '00'
 228         expiration_date = '%x' % (int(time.time()) + life)
 229
 230         def str_to_hex(str):
 231             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 232
 233         def hex_to_bytes(hex):
 234             return binascii.a2b_hex(hex.encode('ascii'))
 235
 236         relative_path = re.match(r'https?://link\.theplatform\.com/s/([^?]+)', url).group(1)
 237         clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path))
 238         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 239         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 240         return '%s&sig=%s' % (url, sig)
 241
 242     def _real_extract(self, url):
 243         url, smuggled_data = unsmuggle_url(url, {})
 244         self._initialize_geo_bypass({
 245             'countries': smuggled_data.get('geo_countries'),
 246         })
 247
 248         mobj = self._match_valid_url(url)
 249         provider_id = mobj.group('provider_id')
 250         video_id = mobj.group('id')
 251
 252         if not provider_id:
 253             provider_id = 'dJ5BDC'
 254
 255         path = provider_id + '/'
 256         if mobj.group('media'):
 257             path += mobj.group('media')
 258         path += video_id
 259
 260         qs_dict = parse_qs(url)
 261         if 'guid' in qs_dict:
 262             webpage = self._download_webpage(url, video_id)
 263             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 264             feed_id = None
 265             # feed id usually locates in the last script.
 266             # Seems there's no pattern for the interested script filename, so
 267             # I try one by one
 268             for script in reversed(scripts):
 269                 feed_script = self._download_webpage(
 270                     self._proto_relative_url(script, 'http:'),
 271                     video_id, 'Downloading feed script')
 272                 feed_id = self._search_regex(
 273                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 274                     'default feed id', default=None)
 275                 if feed_id is not None:
 276                     break
 277             if feed_id is None:
 278                 raise ExtractorError('Unable to find feed id')
 279             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 280                 provider_id, feed_id, qs_dict['guid'][0]))
 281
 282         if smuggled_data.get('force_smil_url', False):
 283             smil_url = url
 284         # Explicitly specified SMIL (see https://github.com/ytdl-org/youtube-dl/issues/7385)
 285         elif '/guid/' in url:
 286             headers = {}
 287             source_url = smuggled_data.get('source_url')
 288             if source_url:
 289                 headers['Referer'] = source_url
 290             request = Request(url, headers=headers)
 291             webpage = self._download_webpage(request, video_id)
 292             smil_url = self._search_regex(
 293                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 294                 webpage, 'smil url', group='url')
 295             path = self._search_regex(
 296                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 297             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4'
 298         elif mobj.group('config'):
 299             config_url = url + '&form=json'
 300             config_url = config_url.replace('swf/', 'config/')
 301             config_url = config_url.replace('onsite/', 'onsite/config/')
 302             config = self._download_json(config_url, video_id, 'Downloading config')
 303             if 'releaseUrl' in config:
 304                 release_url = config['releaseUrl']
 305             else:
 306                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 307             smil_url = release_url + '&formats=MPEG4&manifest=f4m'
 308         else:
 309             smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 310
 311         sig = smuggled_data.get('sig')
 312         if sig:
 313             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 314
 315         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 316
 317         # With some sites, manifest URL must be forced to extract HLS formats
 318         if not traverse_obj(formats, lambda _, v: v['format_id'].startswith('hls')):
 319             m3u8_url = update_url(url, query='mbr=true&manifest=m3u', fragment=None)
 320             urlh = self._request_webpage(
 321                 HEADRequest(m3u8_url), video_id, 'Checking for HLS formats', 'No HLS formats found', fatal=False)
 322             if urlh and urlhandle_detect_ext(urlh) == 'm3u8':
 323                 m3u8_fmts, m3u8_subs = self._extract_m3u8_formats_and_subtitles(
 324                     m3u8_url, video_id, m3u8_id='hls', fatal=False)
 325                 formats.extend(m3u8_fmts)
 326                 self._merge_subtitles(m3u8_subs, target=subtitles)
 327
 328         ret = self._extract_theplatform_metadata(path, video_id)
 329         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 330         ret.update({
 331             'id': video_id,
 332             'formats': formats,
 333             'subtitles': combined_subtitles,
 334         })
 335
 336         return ret
 337
 338
 339 class ThePlatformFeedIE(ThePlatformBaseIE):
 340     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
 341     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[^&]+))'
 342     _TESTS = [{
 343         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 344         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 345         'md5': '6e32495b5073ab414471b615c5ded394',
 346         'info_dict': {
 347             'id': 'n_hardball_5biden_140207',
 348             'ext': 'mp4',
 349             'title': 'The Biden factor: will Joe run in 2016?',
 350             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 351             'thumbnail': r're:^https?://.*\.jpg$',
 352             'upload_date': '20140208',
 353             'timestamp': 1391824260,
 354             'duration': 467.0,
 355             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 356             'uploader': 'NBCU-NEWS',
 357         },
 358     }, {
 359         'url': 'http://feed.theplatform.com/f/2E2eJC/nnd_NBCNews?byGuid=nn_netcast_180306.Copy.01',
 360         'only_matching': True,
 361     }]
 362
 363     def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}, account_id=None):
 364         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
 365         entry = self._download_json(real_url, video_id)['entries'][0]
 366         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$publicUrl')
 367
 368         formats = []
 369         subtitles = {}
 370         first_video_id = None
 371         duration = None
 372         asset_types = []
 373         for item in entry['media$content']:
 374             smil_url = item['plfile$url']
 375             cur_video_id = ThePlatformIE._match_id(smil_url)
 376             if first_video_id is None:
 377                 first_video_id = cur_video_id
 378                 duration = float_or_none(item.get('plfile$duration'))
 379             file_asset_types = item.get('plfile$assetTypes') or parse_qs(smil_url)['assetTypes']
 380             for asset_type in file_asset_types:
 381                 if asset_type in asset_types:
 382                     continue
 383                 asset_types.append(asset_type)
 384                 query = {
 385                     'mbr': 'true',
 386                     'formats': item['plfile$format'],
 387                     'assetTypes': asset_type,
 388                 }
 389                 if asset_type in asset_types_query:
 390                     query.update(asset_types_query[asset_type])
 391                 cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
 392                     main_smil_url or smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
 393                 formats.extend(cur_formats)
 394                 subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 395
 396         thumbnails = [{
 397             'url': thumbnail['plfile$url'],
 398             'width': int_or_none(thumbnail.get('plfile$width')),
 399             'height': int_or_none(thumbnail.get('plfile$height')),
 400         } for thumbnail in entry.get('media$thumbnails', [])]
 401
 402         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 403         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 404
 405         ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 406         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 407         ret.update({
 408             'id': video_id,
 409             'formats': formats,
 410             'subtitles': subtitles,
 411             'thumbnails': thumbnails,
 412             'duration': duration,
 413             'timestamp': timestamp,
 414             'categories': categories,
 415         })
 416         if custom_fields:
 417             ret.update(custom_fields(entry))
 418
 419         return ret
 420
 421     def _real_extract(self, url):
 422         mobj = self._match_valid_url(url)
 423
 424         video_id = mobj.group('id')
 425         provider_id = mobj.group('provider_id')
 426         feed_id = mobj.group('feed_id')
 427         filter_query = mobj.group('filter')
 428
 429         return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)