_ 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 import netrc
  10
  11
  12 from .once import OnceIE
  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     unescapeHTML,
  29     urlencode_postdata,
  30     unified_timestamp,
  31 )
  32
  33 default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  34 _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  35
  36
  37 class ThePlatformBaseIE(OnceIE):
  38     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  39         meta = self._download_xml(smil_url, video_id, note=note, query={'format': 'SMIL'})
  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.com/s/errorFiles/Unavailable.'):
  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.com/s/%s?format=preview' % 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[lang] = [{
  80                     'ext': mimetype2ext(mime),
  81                     'url': src,
  82                 }]
  83
  84         return {
  85             'title': info['title'],
  86             'subtitles': subtitles,
  87             'description': info['description'],
  88             'thumbnail': info['defaultThumbnailUrl'],
  89             'duration': int_or_none(info.get('duration'), 1000),
  90             'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
  91             'uploader': info.get('billingCode'),
  92         }
  93
  94     def _extract_theplatform_metadata(self, path, video_id):
  95         info = self._download_theplatform_metadata(path, video_id)
  96         return self._parse_theplatform_metadata(info)
  97
  98
  99 class ThePlatformIE(ThePlatformBaseIE):
 100     _VALID_URL = r'''(?x)
 101         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
 102            (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
 103          |theplatform:)(?P<id>[^/\?&]+)'''
 104
 105     _TESTS = [{
 106         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
 107         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
 108         'info_dict': {
 109             'id': 'e9I_cZgTgIPd',
 110             'ext': 'flv',
 111             'title': 'Blackberry\'s big, bold Z30',
 112             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
 113             'duration': 247,
 114             'timestamp': 1383239700,
 115             'upload_date': '20131031',
 116             'uploader': 'CBSI-NEW',
 117         },
 118         'params': {
 119             # rtmp download
 120             'skip_download': True,
 121         },
 122     }, {
 123         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 124         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 125         'info_dict': {
 126             'id': '22d_qsQ6MIRT',
 127             'ext': 'flv',
 128             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 129             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 130             'timestamp': 1426176191,
 131             'upload_date': '20150312',
 132             'uploader': 'CBSI-NEW',
 133         },
 134         'params': {
 135             # rtmp download
 136             'skip_download': True,
 137         }
 138     }, {
 139         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 140         'info_dict': {
 141             'id': 'yMBg9E8KFxZD',
 142             'ext': 'mp4',
 143             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 144             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 145             'uploader': 'EGSM',
 146         }
 147     }, {
 148         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 149         'only_matching': True,
 150     }, {
 151         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 152         'md5': 'fb96bb3d85118930a5b055783a3bd992',
 153         'info_dict': {
 154             'id': 'tdy_or_siri_150701',
 155             'ext': 'mp4',
 156             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 157             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 158             'duration': 83.0,
 159             'thumbnail': 're:^https?://.*\.jpg$',
 160             'timestamp': 1435752600,
 161             'upload_date': '20150701',
 162             'uploader': 'NBCU-NEWS',
 163         },
 164     }, {
 165         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 166         # geo-restricted (US), HLS encrypted with AES-128
 167         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 168         'only_matching': True,
 169     }]
 170     _SERVICE_PROVIDER_TEMPLATE = 'https://sp.auth.adobe.com/adobe-services/%s'
 171
 172     @classmethod
 173     def _extract_urls(cls, webpage):
 174         m = re.search(
 175             r'''(?x)
 176                     <meta\s+
 177                         property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
 178                         content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2
 179             ''', webpage)
 180         if m:
 181             return [m.group('url')]
 182
 183         matches = re.findall(
 184             r'<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage)
 185         if matches:
 186             return list(zip(*matches))[1]
 187
 188     @staticmethod
 189     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 190         flags = '10' if include_qs else '00'
 191         expiration_date = '%x' % (int(time.time()) + life)
 192
 193         def str_to_hex(str):
 194             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 195
 196         def hex_to_bytes(hex):
 197             return binascii.a2b_hex(hex.encode('ascii'))
 198
 199         relative_path = re.match(r'https?://link.theplatform.com/s/([^?]+)', url).group(1)
 200         clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path))
 201         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 202         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 203         return '%s&sig=%s' % (url, sig)
 204
 205     def _extract_mvpd_auth(self, url, video_id, requestor_id, resource):
 206         def xml_text(xml_str, tag):
 207             return self._search_regex(
 208                 '<%s>(.+?)</%s>' % (tag, tag), xml_str, tag)
 209
 210         mvpd_headers = {
 211             'ap_42': 'anonymous',
 212             'ap_11': 'Linux i686',
 213             'ap_z': 'Mozilla/5.0 (X11; Linux i686; rv:47.0) Gecko/20100101 Firefox/47.0',
 214             'User-Agent': 'Mozilla/5.0 (X11; Linux i686; rv:47.0) Gecko/20100101 Firefox/47.0',
 215         }
 216
 217         guid = xml_text(resource, 'guid')
 218         requestor_info = self._downloader.cache.load('mvpd', requestor_id) or {}
 219         authn_token = requestor_info.get('authn_token')
 220         if authn_token:
 221             token_expires = unified_timestamp(re.sub(r'[_ ]GMT', '', xml_text(authn_token, 'simpleTokenExpires')))
 222             if token_expires and token_expires <= int(time.time()):
 223                 authn_token = None
 224                 requestor_info = {}
 225         if not authn_token:
 226             # TODO add support for other TV Providers
 227             mso_id = 'DTV'
 228             username, password = self._get_netrc_login_info(mso_id)
 229             if not username or not password:
 230                 return ''
 231
 232             def post_form(form_page, note, data={}):
 233                 post_url = self._html_search_regex(r'<form[^>]+action=(["\'])(?P<url>.+?)\1', form_page, 'post url', group='url')
 234                 return self._download_webpage(
 235                     post_url, video_id, note, data=urlencode_postdata(data or self._hidden_inputs(form_page)), headers={
 236                         'Content-Type': 'application/x-www-form-urlencoded',
 237                     })
 238
 239             provider_redirect_page = self._download_webpage(
 240                 self._SERVICE_PROVIDER_TEMPLATE % 'authenticate/saml', video_id,
 241                 'Downloading Provider Redirect Page', query={
 242                     'noflash': 'true',
 243                     'mso_id': mso_id,
 244                     'requestor_id': requestor_id,
 245                     'no_iframe': 'false',
 246                     'domain_name': 'adobe.com',
 247                     'redirect_url': url,
 248                 })
 249             provider_login_page = post_form(
 250                 provider_redirect_page, 'Downloading Provider Login Page')
 251             mvpd_confirm_page = post_form(provider_login_page, 'Logging in', {
 252                 'username': username,
 253                 'password': password,
 254             })
 255             post_form(mvpd_confirm_page, 'Confirming Login')
 256
 257             session = self._download_webpage(
 258                 self._SERVICE_PROVIDER_TEMPLATE % 'session', video_id,
 259                 'Retrieving Session', data=urlencode_postdata({
 260                     '_method': 'GET',
 261                     'requestor_id': requestor_id,
 262                 }), headers=mvpd_headers)
 263             authn_token = unescapeHTML(xml_text(session, 'authnToken'))
 264             requestor_info['authn_token'] = authn_token
 265             self._downloader.cache.store('mvpd', requestor_id, requestor_info)
 266
 267         authz_token = requestor_info.get(guid)
 268         if not authz_token:
 269             authorize = self._download_webpage(
 270                 self._SERVICE_PROVIDER_TEMPLATE % 'authorize', video_id,
 271                 'Retrieving Authorization Token', data=urlencode_postdata({
 272                     'resource_id': resource,
 273                     'requestor_id': requestor_id,
 274                     'authentication_token': authn_token,
 275                     'mso_id': xml_text(authn_token, 'simpleTokenMsoID'),
 276                     'userMeta': '1',
 277                 }), headers=mvpd_headers)
 278             authz_token = unescapeHTML(xml_text(authorize, 'authzToken'))
 279             requestor_info[guid] = authz_token
 280             self._downloader.cache.store('mvpd', requestor_id, requestor_info)
 281
 282         mvpd_headers.update({
 283             'ap_19': xml_text(authn_token, 'simpleSamlNameID'),
 284             'ap_23': xml_text(authn_token, 'simpleSamlSessionIndex'),
 285         })
 286
 287         return self._download_webpage(
 288             self._SERVICE_PROVIDER_TEMPLATE % 'shortAuthorize',
 289             video_id, 'Retrieving Media Token', data=urlencode_postdata({
 290                 'authz_token': authz_token,
 291                 'requestor_id': requestor_id,
 292                 'session_guid': xml_text(authn_token, 'simpleTokenAuthenticationGuid'),
 293                 'hashed_guid': 'false',
 294             }), headers=mvpd_headers)
 295
 296     def _real_extract(self, url):
 297         url, smuggled_data = unsmuggle_url(url, {})
 298
 299         mobj = re.match(self._VALID_URL, url)
 300         provider_id = mobj.group('provider_id')
 301         video_id = mobj.group('id')
 302
 303         if not provider_id:
 304             provider_id = 'dJ5BDC'
 305
 306         path = provider_id + '/'
 307         if mobj.group('media'):
 308             path += mobj.group('media')
 309         path += video_id
 310
 311         qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
 312         if 'guid' in qs_dict:
 313             webpage = self._download_webpage(url, video_id)
 314             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 315             feed_id = None
 316             # feed id usually locates in the last script.
 317             # Seems there's no pattern for the interested script filename, so
 318             # I try one by one
 319             for script in reversed(scripts):
 320                 feed_script = self._download_webpage(
 321                     self._proto_relative_url(script, 'http:'),
 322                     video_id, 'Downloading feed script')
 323                 feed_id = self._search_regex(
 324                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 325                     'default feed id', default=None)
 326                 if feed_id is not None:
 327                     break
 328             if feed_id is None:
 329                 raise ExtractorError('Unable to find feed id')
 330             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 331                 provider_id, feed_id, qs_dict['guid'][0]))
 332
 333         if smuggled_data.get('force_smil_url', False):
 334             smil_url = url
 335         # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385)
 336         elif '/guid/' in url:
 337             headers = {}
 338             source_url = smuggled_data.get('source_url')
 339             if source_url:
 340                 headers['Referer'] = source_url
 341             request = sanitized_Request(url, headers=headers)
 342             webpage = self._download_webpage(request, video_id)
 343             smil_url = self._search_regex(
 344                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 345                 webpage, 'smil url', group='url')
 346             path = self._search_regex(
 347                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 348             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4'
 349         elif mobj.group('config'):
 350             config_url = url + '&form=json'
 351             config_url = config_url.replace('swf/', 'config/')
 352             config_url = config_url.replace('onsite/', 'onsite/config/')
 353             config = self._download_json(config_url, video_id, 'Downloading config')
 354             if 'releaseUrl' in config:
 355                 release_url = config['releaseUrl']
 356             else:
 357                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 358             smil_url = release_url + '&formats=MPEG4&manifest=f4m'
 359         else:
 360             smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 361
 362         sig = smuggled_data.get('sig')
 363         if sig:
 364             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 365
 366         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 367         self._sort_formats(formats)
 368
 369         ret = self._extract_theplatform_metadata(path, video_id)
 370         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 371         ret.update({
 372             'id': video_id,
 373             'formats': formats,
 374             'subtitles': combined_subtitles,
 375         })
 376
 377         return ret
 378
 379
 380 class ThePlatformFeedIE(ThePlatformBaseIE):
 381     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
 382     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[\w-]+))'
 383     _TESTS = [{
 384         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 385         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 386         'md5': '6e32495b5073ab414471b615c5ded394',
 387         'info_dict': {
 388             'id': 'n_hardball_5biden_140207',
 389             'ext': 'mp4',
 390             'title': 'The Biden factor: will Joe run in 2016?',
 391             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 392             'thumbnail': 're:^https?://.*\.jpg$',
 393             'upload_date': '20140208',
 394             'timestamp': 1391824260,
 395             'duration': 467.0,
 396             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 397             'uploader': 'NBCU-NEWS',
 398         },
 399     }]
 400
 401     def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}):
 402         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
 403         entry = self._download_json(real_url, video_id)['entries'][0]
 404
 405         formats = []
 406         subtitles = {}
 407         first_video_id = None
 408         duration = None
 409         asset_types = []
 410         for item in entry['media$content']:
 411             smil_url = item['plfile$url']
 412             cur_video_id = ThePlatformIE._match_id(smil_url)
 413             if first_video_id is None:
 414                 first_video_id = cur_video_id
 415                 duration = float_or_none(item.get('plfile$duration'))
 416             for asset_type in item['plfile$assetTypes']:
 417                 if asset_type in asset_types:
 418                     continue
 419                 asset_types.append(asset_type)
 420                 query = {
 421                     'mbr': 'true',
 422                     'formats': item['plfile$format'],
 423                     'assetTypes': asset_type,
 424                 }
 425                 if asset_type in asset_types_query:
 426                     query.update(asset_types_query[asset_type])
 427                 cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
 428                     smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
 429                 formats.extend(cur_formats)
 430                 subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 431
 432         self._sort_formats(formats)
 433
 434         thumbnails = [{
 435             'url': thumbnail['plfile$url'],
 436             'width': int_or_none(thumbnail.get('plfile$width')),
 437             'height': int_or_none(thumbnail.get('plfile$height')),
 438         } for thumbnail in entry.get('media$thumbnails', [])]
 439
 440         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 441         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 442
 443         ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 444         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 445         ret.update({
 446             'id': video_id,
 447             'formats': formats,
 448             'subtitles': subtitles,
 449             'thumbnails': thumbnails,
 450             'duration': duration,
 451             'timestamp': timestamp,
 452             'categories': categories,
 453         })
 454         if custom_fields:
 455             ret.update(custom_fields(entry))
 456
 457         return ret
 458
 459     def _real_extract(self, url):
 460         mobj = re.match(self._VALID_URL, url)
 461
 462         video_id = mobj.group('id')
 463         provider_id = mobj.group('provider_id')
 464         feed_id = mobj.group('feed_id')
 465         filter_query = mobj.group('filter')
 466
 467         return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)