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