_ Git - youtube-dl/blob - youtube_dl/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     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  37         meta = self._download_xml(
  38             smil_url, video_id, note=note, query={'format': 'SMIL'},
  39             headers=self.geo_verification_headers())
  40         error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src')
  41         if error_element is not None and error_element.attrib['src'].startswith(
  42                 'http://link.theplatform.%s/s/errorFiles/Unavailable.' % self._TP_TLD):
  43             raise ExtractorError(error_element.attrib['abstract'], expected=True)
  44
  45         smil_formats = self._parse_smil_formats(
  46             meta, smil_url, video_id, namespace=default_ns,
  47             # the parameters are from syfy.com, other sites may use others,
  48             # they also work for nbc.com
  49             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  50             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  51
  52         formats = []
  53         for _format in smil_formats:
  54             if OnceIE.suitable(_format['url']):
  55                 formats.extend(self._extract_once_formats(_format['url']))
  56             else:
  57                 media_url = _format['url']
  58                 if determine_ext(media_url) == 'm3u8':
  59                     hdnea2 = self._get_cookies(media_url).get('hdnea2')
  60                     if hdnea2:
  61                         _format['url'] = update_url_query(media_url, {'hdnea3': hdnea2.value})
  62
  63                 formats.append(_format)
  64
  65         subtitles = self._parse_smil_subtitles(meta, default_ns)
  66
  67         return formats, subtitles
  68
  69     def _download_theplatform_metadata(self, path, video_id):
  70         info_url = 'http://link.theplatform.%s/s/%s?format=preview' % (self._TP_TLD, path)
  71         return self._download_json(info_url, video_id)
  72
  73     def _parse_theplatform_metadata(self, info):
  74         subtitles = {}
  75         captions = info.get('captions')
  76         if isinstance(captions, list):
  77             for caption in captions:
  78                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  79                 subtitles.setdefault(lang, []).append({
  80                     'ext': mimetype2ext(mime),
  81                     'url': src,
  82                 })
  83
  84         duration = info.get('duration')
  85         tp_chapters = info.get('chapters', [])
  86         chapters = []
  87         if tp_chapters:
  88             def _add_chapter(start_time, end_time):
  89                 start_time = float_or_none(start_time, 1000)
  90                 end_time = float_or_none(end_time, 1000)
  91                 if start_time is None or end_time is None:
  92                     return
  93                 chapters.append({
  94                     'start_time': start_time,
  95                     'end_time': end_time,
  96                 })
  97
  98             for chapter in tp_chapters[:-1]:
  99                 _add_chapter(chapter.get('startTime'), chapter.get('endTime'))
 100             _add_chapter(tp_chapters[-1].get('startTime'), tp_chapters[-1].get('endTime') or duration)
 101
 102         return {
 103             'title': info['title'],
 104             'subtitles': subtitles,
 105             'description': info['description'],
 106             'thumbnail': info['defaultThumbnailUrl'],
 107             'duration': float_or_none(duration, 1000),
 108             'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
 109             'uploader': info.get('billingCode'),
 110             'chapters': chapters,
 111         }
 112
 113     def _extract_theplatform_metadata(self, path, video_id):
 114         info = self._download_theplatform_metadata(path, video_id)
 115         return self._parse_theplatform_metadata(info)
 116
 117
 118 class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
 119     _VALID_URL = r'''(?x)
 120         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
 121            (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)?|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
 122          |theplatform:)(?P<id>[^/\?&]+)'''
 123
 124     _TESTS = [{
 125         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
 126         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
 127         'info_dict': {
 128             'id': 'e9I_cZgTgIPd',
 129             'ext': 'flv',
 130             'title': 'Blackberry\'s big, bold Z30',
 131             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
 132             'duration': 247,
 133             'timestamp': 1383239700,
 134             'upload_date': '20131031',
 135             'uploader': 'CBSI-NEW',
 136         },
 137         'params': {
 138             # rtmp download
 139             'skip_download': True,
 140         },
 141         'skip': '404 Not Found',
 142     }, {
 143         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 144         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 145         'info_dict': {
 146             'id': '22d_qsQ6MIRT',
 147             'ext': 'flv',
 148             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 149             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 150             'timestamp': 1426176191,
 151             'upload_date': '20150312',
 152             'uploader': 'CBSI-NEW',
 153         },
 154         'params': {
 155             # rtmp download
 156             'skip_download': True,
 157         }
 158     }, {
 159         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 160         'info_dict': {
 161             'id': 'yMBg9E8KFxZD',
 162             'ext': 'mp4',
 163             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 164             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 165             'uploader': 'EGSM',
 166         }
 167     }, {
 168         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 169         'only_matching': True,
 170     }, {
 171         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 172         'md5': 'fb96bb3d85118930a5b055783a3bd992',
 173         'info_dict': {
 174             'id': 'tdy_or_siri_150701',
 175             'ext': 'mp4',
 176             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 177             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 178             'duration': 83.0,
 179             'thumbnail': r're:^https?://.*\.jpg$',
 180             'timestamp': 1435752600,
 181             'upload_date': '20150701',
 182             'uploader': 'NBCU-NEWS',
 183         },
 184     }, {
 185         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 186         # geo-restricted (US), HLS encrypted with AES-128
 187         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 188         'only_matching': True,
 189     }]
 190
 191     @classmethod
 192     def _extract_urls(cls, webpage):
 193         m = re.search(
 194             r'''(?x)
 195                     <meta\s+
 196                         property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
 197                         content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2
 198             ''', webpage)
 199         if m:
 200             return [m.group('url')]
 201
 202         # Are whitesapces ignored in URLs?
 203         # https://github.com/rg3/youtube-dl/issues/12044
 204         matches = re.findall(
 205             r'(?s)<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage)
 206         if matches:
 207             return [re.sub(r'\s', '', list(zip(*matches))[1][0])]
 208
 209     @staticmethod
 210     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 211         flags = '10' if include_qs else '00'
 212         expiration_date = '%x' % (int(time.time()) + life)
 213
 214         def str_to_hex(str):
 215             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 216
 217         def hex_to_bytes(hex):
 218             return binascii.a2b_hex(hex.encode('ascii'))
 219
 220         relative_path = re.match(r'https?://link\.theplatform\.com/s/([^?]+)', url).group(1)
 221         clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path))
 222         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 223         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 224         return '%s&sig=%s' % (url, sig)
 225
 226     def _real_extract(self, url):
 227         url, smuggled_data = unsmuggle_url(url, {})
 228
 229         mobj = re.match(self._VALID_URL, url)
 230         provider_id = mobj.group('provider_id')
 231         video_id = mobj.group('id')
 232
 233         if not provider_id:
 234             provider_id = 'dJ5BDC'
 235
 236         path = provider_id + '/'
 237         if mobj.group('media'):
 238             path += mobj.group('media')
 239         path += video_id
 240
 241         qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
 242         if 'guid' in qs_dict:
 243             webpage = self._download_webpage(url, video_id)
 244             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 245             feed_id = None
 246             # feed id usually locates in the last script.
 247             # Seems there's no pattern for the interested script filename, so
 248             # I try one by one
 249             for script in reversed(scripts):
 250                 feed_script = self._download_webpage(
 251                     self._proto_relative_url(script, 'http:'),
 252                     video_id, 'Downloading feed script')
 253                 feed_id = self._search_regex(
 254                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 255                     'default feed id', default=None)
 256                 if feed_id is not None:
 257                     break
 258             if feed_id is None:
 259                 raise ExtractorError('Unable to find feed id')
 260             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 261                 provider_id, feed_id, qs_dict['guid'][0]))
 262
 263         if smuggled_data.get('force_smil_url', False):
 264             smil_url = url
 265         # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385)
 266         elif '/guid/' in url:
 267             headers = {}
 268             source_url = smuggled_data.get('source_url')
 269             if source_url:
 270                 headers['Referer'] = source_url
 271             request = sanitized_Request(url, headers=headers)
 272             webpage = self._download_webpage(request, video_id)
 273             smil_url = self._search_regex(
 274                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 275                 webpage, 'smil url', group='url')
 276             path = self._search_regex(
 277                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 278             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4'
 279         elif mobj.group('config'):
 280             config_url = url + '&form=json'
 281             config_url = config_url.replace('swf/', 'config/')
 282             config_url = config_url.replace('onsite/', 'onsite/config/')
 283             config = self._download_json(config_url, video_id, 'Downloading config')
 284             if 'releaseUrl' in config:
 285                 release_url = config['releaseUrl']
 286             else:
 287                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 288             smil_url = release_url + '&formats=MPEG4&manifest=f4m'
 289         else:
 290             smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 291
 292         sig = smuggled_data.get('sig')
 293         if sig:
 294             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 295
 296         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 297         self._sort_formats(formats)
 298
 299         ret = self._extract_theplatform_metadata(path, video_id)
 300         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 301         ret.update({
 302             'id': video_id,
 303             'formats': formats,
 304             'subtitles': combined_subtitles,
 305         })
 306
 307         return ret
 308
 309
 310 class ThePlatformFeedIE(ThePlatformBaseIE):
 311     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
 312     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[\w-]+))'
 313     _TESTS = [{
 314         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 315         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 316         'md5': '6e32495b5073ab414471b615c5ded394',
 317         'info_dict': {
 318             'id': 'n_hardball_5biden_140207',
 319             'ext': 'mp4',
 320             'title': 'The Biden factor: will Joe run in 2016?',
 321             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 322             'thumbnail': r're:^https?://.*\.jpg$',
 323             'upload_date': '20140208',
 324             'timestamp': 1391824260,
 325             'duration': 467.0,
 326             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 327             'uploader': 'NBCU-NEWS',
 328         },
 329     }]
 330
 331     def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}, account_id=None):
 332         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
 333         entry = self._download_json(real_url, video_id)['entries'][0]
 334         main_smil_url = 'http://link.theplatform.com/s/%s/media/guid/%d/%s' % (provider_id, account_id, entry['guid']) if account_id else None
 335
 336         formats = []
 337         subtitles = {}
 338         first_video_id = None
 339         duration = None
 340         asset_types = []
 341         for item in entry['media$content']:
 342             smil_url = item['plfile$url']
 343             cur_video_id = ThePlatformIE._match_id(smil_url)
 344             if first_video_id is None:
 345                 first_video_id = cur_video_id
 346                 duration = float_or_none(item.get('plfile$duration'))
 347             for asset_type in item['plfile$assetTypes']:
 348                 if asset_type in asset_types:
 349                     continue
 350                 asset_types.append(asset_type)
 351                 query = {
 352                     'mbr': 'true',
 353                     'formats': item['plfile$format'],
 354                     'assetTypes': asset_type,
 355                 }
 356                 if asset_type in asset_types_query:
 357                     query.update(asset_types_query[asset_type])
 358                 cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
 359                     main_smil_url or smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
 360                 formats.extend(cur_formats)
 361                 subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 362
 363         self._sort_formats(formats)
 364
 365         thumbnails = [{
 366             'url': thumbnail['plfile$url'],
 367             'width': int_or_none(thumbnail.get('plfile$width')),
 368             'height': int_or_none(thumbnail.get('plfile$height')),
 369         } for thumbnail in entry.get('media$thumbnails', [])]
 370
 371         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 372         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 373
 374         ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 375         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 376         ret.update({
 377             'id': video_id,
 378             'formats': formats,
 379             'subtitles': subtitles,
 380             'thumbnails': thumbnails,
 381             'duration': duration,
 382             'timestamp': timestamp,
 383             'categories': categories,
 384         })
 385         if custom_fields:
 386             ret.update(custom_fields(entry))
 387
 388         return ret
 389
 390     def _real_extract(self, url):
 391         mobj = re.match(self._VALID_URL, url)
 392
 393         video_id = mobj.group('id')
 394         provider_id = mobj.group('provider_id')
 395         feed_id = mobj.group('feed_id')
 396         filter_query = mobj.group('filter')
 397
 398         return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)