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
 127     _TESTS = [{
 128         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
 129         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
 130         'info_dict': {
 131             'id': 'e9I_cZgTgIPd',
 132             'ext': 'flv',
 133             'title': 'Blackberry\'s big, bold Z30',
 134             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
 135             'duration': 247,
 136             'timestamp': 1383239700,
 137             'upload_date': '20131031',
 138             'uploader': 'CBSI-NEW',
 139         },
 140         'params': {
 141             # rtmp download
 142             'skip_download': True,
 143         },
 144         'skip': '404 Not Found',
 145     }, {
 146         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 147         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 148         'info_dict': {
 149             'id': '22d_qsQ6MIRT',
 150             'ext': 'flv',
 151             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 152             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 153             'timestamp': 1426176191,
 154             'upload_date': '20150312',
 155             'uploader': 'CBSI-NEW',
 156         },
 157         'params': {
 158             # rtmp download
 159             'skip_download': True,
 160         }
 161     }, {
 162         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 163         'info_dict': {
 164             'id': 'yMBg9E8KFxZD',
 165             'ext': 'mp4',
 166             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 167             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 168             'uploader': 'EGSM',
 169         }
 170     }, {
 171         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 172         'only_matching': True,
 173     }, {
 174         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 175         'md5': 'fb96bb3d85118930a5b055783a3bd992',
 176         'info_dict': {
 177             'id': 'tdy_or_siri_150701',
 178             'ext': 'mp4',
 179             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 180             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 181             'duration': 83.0,
 182             'thumbnail': r're:^https?://.*\.jpg$',
 183             'timestamp': 1435752600,
 184             'upload_date': '20150701',
 185             'uploader': 'NBCU-NEWS',
 186         },
 187     }, {
 188         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 189         # geo-restricted (US), HLS encrypted with AES-128
 190         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 191         'only_matching': True,
 192     }]
 193
 194     @classmethod
 195     def _extract_urls(cls, webpage):
 196         m = re.search(
 197             r'''(?x)
 198                     <meta\s+
 199                         property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
 200                         content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2
 201             ''', webpage)
 202         if m:
 203             return [m.group('url')]
 204
 205         # Are whitespaces ignored in URLs?
 206         # https://github.com/ytdl-org/youtube-dl/issues/12044
 207         matches = re.findall(
 208             r'(?s)<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage)
 209         if matches:
 210             return [re.sub(r'\s', '', list(zip(*matches))[1][0])]
 211
 212     @staticmethod
 213     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 214         flags = '10' if include_qs else '00'
 215         expiration_date = '%x' % (int(time.time()) + life)
 216
 217         def str_to_hex(str):
 218             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 219
 220         def hex_to_bytes(hex):
 221             return binascii.a2b_hex(hex.encode('ascii'))
 222
 223         relative_path = re.match(r'https?://link\.theplatform\.com/s/([^?]+)', url).group(1)
 224         clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path))
 225         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 226         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 227         return '%s&sig=%s' % (url, sig)
 228
 229     def _real_extract(self, url):
 230         url, smuggled_data = unsmuggle_url(url, {})
 231         self._initialize_geo_bypass({
 232             'countries': smuggled_data.get('geo_countries'),
 233         })
 234
 235         mobj = self._match_valid_url(url)
 236         provider_id = mobj.group('provider_id')
 237         video_id = mobj.group('id')
 238
 239         if not provider_id:
 240             provider_id = 'dJ5BDC'
 241
 242         path = provider_id + '/'
 243         if mobj.group('media'):
 244             path += mobj.group('media')
 245         path += video_id
 246
 247         qs_dict = parse_qs(url)
 248         if 'guid' in qs_dict:
 249             webpage = self._download_webpage(url, video_id)
 250             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 251             feed_id = None
 252             # feed id usually locates in the last script.
 253             # Seems there's no pattern for the interested script filename, so
 254             # I try one by one
 255             for script in reversed(scripts):
 256                 feed_script = self._download_webpage(
 257                     self._proto_relative_url(script, 'http:'),
 258                     video_id, 'Downloading feed script')
 259                 feed_id = self._search_regex(
 260                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 261                     'default feed id', default=None)
 262                 if feed_id is not None:
 263                     break
 264             if feed_id is None:
 265                 raise ExtractorError('Unable to find feed id')
 266             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 267                 provider_id, feed_id, qs_dict['guid'][0]))
 268
 269         if smuggled_data.get('force_smil_url', False):
 270             smil_url = url
 271         # Explicitly specified SMIL (see https://github.com/ytdl-org/youtube-dl/issues/7385)
 272         elif '/guid/' in url:
 273             headers = {}
 274             source_url = smuggled_data.get('source_url')
 275             if source_url:
 276                 headers['Referer'] = source_url
 277             request = sanitized_Request(url, headers=headers)
 278             webpage = self._download_webpage(request, video_id)
 279             smil_url = self._search_regex(
 280                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 281                 webpage, 'smil url', group='url')
 282             path = self._search_regex(
 283                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 284             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4'
 285         elif mobj.group('config'):
 286             config_url = url + '&form=json'
 287             config_url = config_url.replace('swf/', 'config/')
 288             config_url = config_url.replace('onsite/', 'onsite/config/')
 289             config = self._download_json(config_url, video_id, 'Downloading config')
 290             if 'releaseUrl' in config:
 291                 release_url = config['releaseUrl']
 292             else:
 293                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 294             smil_url = release_url + '&formats=MPEG4&manifest=f4m'
 295         else:
 296             smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 297
 298         sig = smuggled_data.get('sig')
 299         if sig:
 300             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 301
 302         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 303         self._sort_formats(formats)
 304
 305         ret = self._extract_theplatform_metadata(path, video_id)
 306         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 307         ret.update({
 308             'id': video_id,
 309             'formats': formats,
 310             'subtitles': combined_subtitles,
 311         })
 312
 313         return ret
 314
 315
 316 class ThePlatformFeedIE(ThePlatformBaseIE):
 317     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
 318     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[^&]+))'
 319     _TESTS = [{
 320         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 321         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 322         'md5': '6e32495b5073ab414471b615c5ded394',
 323         'info_dict': {
 324             'id': 'n_hardball_5biden_140207',
 325             'ext': 'mp4',
 326             'title': 'The Biden factor: will Joe run in 2016?',
 327             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 328             'thumbnail': r're:^https?://.*\.jpg$',
 329             'upload_date': '20140208',
 330             'timestamp': 1391824260,
 331             'duration': 467.0,
 332             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 333             'uploader': 'NBCU-NEWS',
 334         },
 335     }, {
 336         'url': 'http://feed.theplatform.com/f/2E2eJC/nnd_NBCNews?byGuid=nn_netcast_180306.Copy.01',
 337         'only_matching': True,
 338     }]
 339
 340     def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}, account_id=None):
 341         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
 342         entry = self._download_json(real_url, video_id)['entries'][0]
 343         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$publicUrl')
 344
 345         formats = []
 346         subtitles = {}
 347         first_video_id = None
 348         duration = None
 349         asset_types = []
 350         for item in entry['media$content']:
 351             smil_url = item['plfile$url']
 352             cur_video_id = ThePlatformIE._match_id(smil_url)
 353             if first_video_id is None:
 354                 first_video_id = cur_video_id
 355                 duration = float_or_none(item.get('plfile$duration'))
 356             file_asset_types = item.get('plfile$assetTypes') or parse_qs(smil_url)['assetTypes']
 357             for asset_type in file_asset_types:
 358                 if asset_type in asset_types:
 359                     continue
 360                 asset_types.append(asset_type)
 361                 query = {
 362                     'mbr': 'true',
 363                     'formats': item['plfile$format'],
 364                     'assetTypes': asset_type,
 365                 }
 366                 if asset_type in asset_types_query:
 367                     query.update(asset_types_query[asset_type])
 368                 cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
 369                     main_smil_url or smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
 370                 formats.extend(cur_formats)
 371                 subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 372
 373         self._sort_formats(formats)
 374
 375         thumbnails = [{
 376             'url': thumbnail['plfile$url'],
 377             'width': int_or_none(thumbnail.get('plfile$width')),
 378             'height': int_or_none(thumbnail.get('plfile$height')),
 379         } for thumbnail in entry.get('media$thumbnails', [])]
 380
 381         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 382         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 383
 384         ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 385         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 386         ret.update({
 387             'id': video_id,
 388             'formats': formats,
 389             'subtitles': subtitles,
 390             'thumbnails': thumbnails,
 391             'duration': duration,
 392             'timestamp': timestamp,
 393             'categories': categories,
 394         })
 395         if custom_fields:
 396             ret.update(custom_fields(entry))
 397
 398         return ret
 399
 400     def _real_extract(self, url):
 401         mobj = self._match_valid_url(url)
 402
 403         video_id = mobj.group('id')
 404         provider_id = mobj.group('provider_id')
 405         feed_id = mobj.group('feed_id')
 406         filter_query = mobj.group('filter')
 407
 408         return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)