yt_dlp/extractor/theplatform.py

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