yt_dlp/extractor/theplatform.py

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import re
   5 import time
   6 import hmac
   7 import binascii
   8 import hashlib
   9
  10
  11 from .once import OnceIE
  12 from .adobepass import AdobePassIE
  13 from ..compat import (
  14     compat_parse_qs,
  15     compat_urllib_parse_urlparse,
  16 )
  17 from ..utils import (
  18     determine_ext,
  19     ExtractorError,
  20     float_or_none,
  21     int_or_none,
  22     sanitized_Request,
  23     unsmuggle_url,
  24     update_url_query,
  25     xpath_with_ns,
  26     mimetype2ext,
  27     find_xpath_attr,
  28 )
  29
  30 default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  31 _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  32
  33
  34 class ThePlatformBaseIE(OnceIE):
  35     _TP_TLD = 'com'
  36
  37     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  38         meta = self._download_xml(
  39             smil_url, video_id, note=note, query={'format': 'SMIL'},
  40             headers=self.geo_verification_headers())
  41         error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src')
  42         if error_element is not None:
  43             exception = find_xpath_attr(
  44                 error_element, _x('.//smil:param'), 'name', 'exception')
  45             if exception is not None:
  46                 if exception.get('value') == 'GeoLocationBlocked':
  47                     self.raise_geo_restricted(error_element.attrib['abstract'])
  48                 elif error_element.attrib['src'].startswith(
  49                         'http://link.theplatform.%s/s/errorFiles/Unavailable.'
  50                         % self._TP_TLD):
  51                     raise ExtractorError(
  52                         error_element.attrib['abstract'], expected=True)
  53
  54         smil_formats = self._parse_smil_formats(
  55             meta, smil_url, video_id, namespace=default_ns,
  56             # the parameters are from syfy.com, other sites may use others,
  57             # they also work for nbc.com
  58             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  59             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  60
  61         formats = []
  62         for _format in smil_formats:
  63             if OnceIE.suitable(_format['url']):
  64                 formats.extend(self._extract_once_formats(_format['url']))
  65             else:
  66                 media_url = _format['url']
  67                 if determine_ext(media_url) == 'm3u8':
  68                     hdnea2 = self._get_cookies(media_url).get('hdnea2')
  69                     if hdnea2:
  70                         _format['url'] = update_url_query(media_url, {'hdnea3': hdnea2.value})
  71
  72                 formats.append(_format)
  73
  74         subtitles = self._parse_smil_subtitles(meta, default_ns)
  75
  76         return formats, subtitles
  77
  78     def _download_theplatform_metadata(self, path, video_id):
  79         info_url = 'http://link.theplatform.%s/s/%s?format=preview' % (self._TP_TLD, path)
  80         return self._download_json(info_url, video_id)
  81
  82     def _parse_theplatform_metadata(self, info):
  83         subtitles = {}
  84         captions = info.get('captions')
  85         if isinstance(captions, list):
  86             for caption in captions:
  87                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  88                 subtitles.setdefault(lang, []).append({
  89                     'ext': mimetype2ext(mime),
  90                     'url': src,
  91                 })
  92
  93         duration = info.get('duration')
  94         tp_chapters = info.get('chapters', [])
  95         chapters = []
  96         if tp_chapters:
  97             def _add_chapter(start_time, end_time):
  98                 start_time = float_or_none(start_time, 1000)
  99                 end_time = float_or_none(end_time, 1000)
 100                 if start_time is None or end_time is None:
 101                     return
 102                 chapters.append({
 103                     'start_time': start_time,
 104                     'end_time': end_time,
 105                 })
 106
 107             for chapter in tp_chapters[:-1]:
 108                 _add_chapter(chapter.get('startTime'), chapter.get('endTime'))
 109             _add_chapter(tp_chapters[-1].get('startTime'), tp_chapters[-1].get('endTime') or duration)
 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         }
 121
 122     def _extract_theplatform_metadata(self, path, video_id):
 123         info = self._download_theplatform_metadata(path, video_id)
 124         return self._parse_theplatform_metadata(info)
 125
 126
 127 class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
 128     _VALID_URL = r'''(?x)
 129         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
 130            (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)?|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
 131          |theplatform:)(?P<id>[^/\?&]+)'''
 132
 133     _TESTS = [{
 134         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
 135         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
 136         'info_dict': {
 137             'id': 'e9I_cZgTgIPd',
 138             'ext': 'flv',
 139             'title': 'Blackberry\'s big, bold Z30',
 140             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
 141             'duration': 247,
 142             'timestamp': 1383239700,
 143             'upload_date': '20131031',
 144             'uploader': 'CBSI-NEW',
 145         },
 146         'params': {
 147             # rtmp download
 148             'skip_download': True,
 149         },
 150         'skip': '404 Not Found',
 151     }, {
 152         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 153         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 154         'info_dict': {
 155             'id': '22d_qsQ6MIRT',
 156             'ext': 'flv',
 157             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 158             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 159             'timestamp': 1426176191,
 160             'upload_date': '20150312',
 161             'uploader': 'CBSI-NEW',
 162         },
 163         'params': {
 164             # rtmp download
 165             'skip_download': True,
 166         }
 167     }, {
 168         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 169         'info_dict': {
 170             'id': 'yMBg9E8KFxZD',
 171             'ext': 'mp4',
 172             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 173             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 174             'uploader': 'EGSM',
 175         }
 176     }, {
 177         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 178         'only_matching': True,
 179     }, {
 180         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 181         'md5': 'fb96bb3d85118930a5b055783a3bd992',
 182         'info_dict': {
 183             'id': 'tdy_or_siri_150701',
 184             'ext': 'mp4',
 185             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 186             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 187             'duration': 83.0,
 188             'thumbnail': r're:^https?://.*\.jpg$',
 189             'timestamp': 1435752600,
 190             'upload_date': '20150701',
 191             'uploader': 'NBCU-NEWS',
 192         },
 193     }, {
 194         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 195         # geo-restricted (US), HLS encrypted with AES-128
 196         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 197         'only_matching': True,
 198     }]
 199
 200     @classmethod
 201     def _extract_urls(cls, webpage):
 202         m = re.search(
 203             r'''(?x)
 204                     <meta\s+
 205                         property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
 206                         content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2
 207             ''', webpage)
 208         if m:
 209             return [m.group('url')]
 210
 211         # Are whitespaces ignored in URLs?
 212         # https://github.com/ytdl-org/youtube-dl/issues/12044
 213         matches = re.findall(
 214             r'(?s)<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage)
 215         if matches:
 216             return [re.sub(r'\s', '', list(zip(*matches))[1][0])]
 217
 218     @staticmethod
 219     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 220         flags = '10' if include_qs else '00'
 221         expiration_date = '%x' % (int(time.time()) + life)
 222
 223         def str_to_hex(str):
 224             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 225
 226         def hex_to_bytes(hex):
 227             return binascii.a2b_hex(hex.encode('ascii'))
 228
 229         relative_path = re.match(r'https?://link\.theplatform\.com/s/([^?]+)', url).group(1)
 230         clear_text = hex_to_bytes(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 '%s&sig=%s' % (url, 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 = re.match(self._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 = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
 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/%s/%s?byGuid=%s' % (
 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 = sanitized_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             if 'releaseUrl' in config:
 297                 release_url = config['releaseUrl']
 298             else:
 299                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 300             smil_url = release_url + '&formats=MPEG4&manifest=f4m'
 301         else:
 302             smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 303
 304         sig = smuggled_data.get('sig')
 305         if sig:
 306             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 307
 308         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 309         self._sort_formats(formats)
 310
 311         ret = self._extract_theplatform_metadata(path, video_id)
 312         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 313         ret.update({
 314             'id': video_id,
 315             'formats': formats,
 316             'subtitles': combined_subtitles,
 317         })
 318
 319         return ret
 320
 321
 322 class ThePlatformFeedIE(ThePlatformBaseIE):
 323     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
 324     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[^&]+))'
 325     _TESTS = [{
 326         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 327         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 328         'md5': '6e32495b5073ab414471b615c5ded394',
 329         'info_dict': {
 330             'id': 'n_hardball_5biden_140207',
 331             'ext': 'mp4',
 332             'title': 'The Biden factor: will Joe run in 2016?',
 333             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 334             'thumbnail': r're:^https?://.*\.jpg$',
 335             'upload_date': '20140208',
 336             'timestamp': 1391824260,
 337             'duration': 467.0,
 338             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 339             'uploader': 'NBCU-NEWS',
 340         },
 341     }, {
 342         'url': 'http://feed.theplatform.com/f/2E2eJC/nnd_NBCNews?byGuid=nn_netcast_180306.Copy.01',
 343         'only_matching': True,
 344     }]
 345
 346     def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}, account_id=None):
 347         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
 348         entry = self._download_json(real_url, video_id)['entries'][0]
 349         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$publicUrl')
 350
 351         formats = []
 352         subtitles = {}
 353         first_video_id = None
 354         duration = None
 355         asset_types = []
 356         for item in entry['media$content']:
 357             smil_url = item['plfile$url']
 358             cur_video_id = ThePlatformIE._match_id(smil_url)
 359             if first_video_id is None:
 360                 first_video_id = cur_video_id
 361                 duration = float_or_none(item.get('plfile$duration'))
 362             file_asset_types = item.get('plfile$assetTypes') or compat_parse_qs(compat_urllib_parse_urlparse(smil_url).query)['assetTypes']
 363             for asset_type in file_asset_types:
 364                 if asset_type in asset_types:
 365                     continue
 366                 asset_types.append(asset_type)
 367                 query = {
 368                     'mbr': 'true',
 369                     'formats': item['plfile$format'],
 370                     'assetTypes': asset_type,
 371                 }
 372                 if asset_type in asset_types_query:
 373                     query.update(asset_types_query[asset_type])
 374                 cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
 375                     main_smil_url or smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
 376                 formats.extend(cur_formats)
 377                 subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 378
 379         self._sort_formats(formats)
 380
 381         thumbnails = [{
 382             'url': thumbnail['plfile$url'],
 383             'width': int_or_none(thumbnail.get('plfile$width')),
 384             'height': int_or_none(thumbnail.get('plfile$height')),
 385         } for thumbnail in entry.get('media$thumbnails', [])]
 386
 387         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 388         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 389
 390         ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 391         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 392         ret.update({
 393             'id': video_id,
 394             'formats': formats,
 395             'subtitles': subtitles,
 396             'thumbnails': thumbnails,
 397             'duration': duration,
 398             'timestamp': timestamp,
 399             'categories': categories,
 400         })
 401         if custom_fields:
 402             ret.update(custom_fields(entry))
 403
 404         return ret
 405
 406     def _real_extract(self, url):
 407         mobj = re.match(self._VALID_URL, url)
 408
 409         video_id = mobj.group('id')
 410         provider_id = mobj.group('provider_id')
 411         feed_id = mobj.group('feed_id')
 412         filter_query = mobj.group('filter')
 413
 414         return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)