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