Raphaël G. Git Repositories - youtubedl/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     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  36         meta = self._download_xml(
  37             smil_url, video_id, note=note, query={'format': 'SMIL'},
  38             headers=self.geo_verification_headers())
  39         error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src')
  40         if error_element is not None and error_element.attrib['src'].startswith(
  41                 'http://link.theplatform.com/s/errorFiles/Unavailable.'):
  42             raise ExtractorError(error_element.attrib['abstract'], expected=True)
  43
  44         smil_formats = self._parse_smil_formats(
  45             meta, smil_url, video_id, namespace=default_ns,
  46             # the parameters are from syfy.com, other sites may use others,
  47             # they also work for nbc.com
  48             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  49             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  50
  51         formats = []
  52         for _format in smil_formats:
  53             if OnceIE.suitable(_format['url']):
  54                 formats.extend(self._extract_once_formats(_format['url']))
  55             else:
  56                 media_url = _format['url']
  57                 if determine_ext(media_url) == 'm3u8':
  58                     hdnea2 = self._get_cookies(media_url).get('hdnea2')
  59                     if hdnea2:
  60                         _format['url'] = update_url_query(media_url, {'hdnea3': hdnea2.value})
  61
  62                 formats.append(_format)
  63
  64         subtitles = self._parse_smil_subtitles(meta, default_ns)
  65
  66         return formats, subtitles
  67
  68     def _download_theplatform_metadata(self, path, video_id):
  69         info_url = 'http://link.theplatform.com/s/%s?format=preview' % path
  70         return self._download_json(info_url, video_id)
  71
  72     def _parse_theplatform_metadata(self, info):
  73         subtitles = {}
  74         captions = info.get('captions')
  75         if isinstance(captions, list):
  76             for caption in captions:
  77                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  78                 subtitles.setdefault(lang, []).append({
  79                     'ext': mimetype2ext(mime),
  80                     'url': src,
  81                 })
  82
  83         return {
  84             'title': info['title'],
  85             'subtitles': subtitles,
  86             'description': info['description'],
  87             'thumbnail': info['defaultThumbnailUrl'],
  88             'duration': int_or_none(info.get('duration'), 1000),
  89             'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
  90             'uploader': info.get('billingCode'),
  91         }
  92
  93     def _extract_theplatform_metadata(self, path, video_id):
  94         info = self._download_theplatform_metadata(path, video_id)
  95         return self._parse_theplatform_metadata(info)
  96
  97
  98 class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
  99     _VALID_URL = r'''(?x)
 100         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
 101            (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)?|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
 102          |theplatform:)(?P<id>[^/\?&]+)'''
 103
 104     _TESTS = [{
 105         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
 106         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
 107         'info_dict': {
 108             'id': 'e9I_cZgTgIPd',
 109             'ext': 'flv',
 110             'title': 'Blackberry\'s big, bold Z30',
 111             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
 112             'duration': 247,
 113             'timestamp': 1383239700,
 114             'upload_date': '20131031',
 115             'uploader': 'CBSI-NEW',
 116         },
 117         'params': {
 118             # rtmp download
 119             'skip_download': True,
 120         },
 121         'skip': '404 Not Found',
 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': r'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
 171     @classmethod
 172     def _extract_urls(cls, webpage):
 173         m = re.search(
 174             r'''(?x)
 175                     <meta\s+
 176                         property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
 177                         content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2
 178             ''', webpage)
 179         if m:
 180             return [m.group('url')]
 181
 182         matches = re.findall(
 183             r'<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage)
 184         if matches:
 185             return list(zip(*matches))[1]
 186
 187     @staticmethod
 188     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 189         flags = '10' if include_qs else '00'
 190         expiration_date = '%x' % (int(time.time()) + life)
 191
 192         def str_to_hex(str):
 193             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 194
 195         def hex_to_bytes(hex):
 196             return binascii.a2b_hex(hex.encode('ascii'))
 197
 198         relative_path = re.match(r'https?://link.theplatform.com/s/([^?]+)', url).group(1)
 199         clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path))
 200         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 201         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 202         return '%s&sig=%s' % (url, sig)
 203
 204     def _real_extract(self, url):
 205         url, smuggled_data = unsmuggle_url(url, {})
 206
 207         mobj = re.match(self._VALID_URL, url)
 208         provider_id = mobj.group('provider_id')
 209         video_id = mobj.group('id')
 210
 211         if not provider_id:
 212             provider_id = 'dJ5BDC'
 213
 214         path = provider_id + '/'
 215         if mobj.group('media'):
 216             path += mobj.group('media')
 217         path += video_id
 218
 219         qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
 220         if 'guid' in qs_dict:
 221             webpage = self._download_webpage(url, video_id)
 222             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 223             feed_id = None
 224             # feed id usually locates in the last script.
 225             # Seems there's no pattern for the interested script filename, so
 226             # I try one by one
 227             for script in reversed(scripts):
 228                 feed_script = self._download_webpage(
 229                     self._proto_relative_url(script, 'http:'),
 230                     video_id, 'Downloading feed script')
 231                 feed_id = self._search_regex(
 232                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 233                     'default feed id', default=None)
 234                 if feed_id is not None:
 235                     break
 236             if feed_id is None:
 237                 raise ExtractorError('Unable to find feed id')
 238             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 239                 provider_id, feed_id, qs_dict['guid'][0]))
 240
 241         if smuggled_data.get('force_smil_url', False):
 242             smil_url = url
 243         # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385)
 244         elif '/guid/' in url:
 245             headers = {}
 246             source_url = smuggled_data.get('source_url')
 247             if source_url:
 248                 headers['Referer'] = source_url
 249             request = sanitized_Request(url, headers=headers)
 250             webpage = self._download_webpage(request, video_id)
 251             smil_url = self._search_regex(
 252                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 253                 webpage, 'smil url', group='url')
 254             path = self._search_regex(
 255                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 256             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4'
 257         elif mobj.group('config'):
 258             config_url = url + '&form=json'
 259             config_url = config_url.replace('swf/', 'config/')
 260             config_url = config_url.replace('onsite/', 'onsite/config/')
 261             config = self._download_json(config_url, video_id, 'Downloading config')
 262             if 'releaseUrl' in config:
 263                 release_url = config['releaseUrl']
 264             else:
 265                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 266             smil_url = release_url + '&formats=MPEG4&manifest=f4m'
 267         else:
 268             smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 269
 270         sig = smuggled_data.get('sig')
 271         if sig:
 272             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 273
 274         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 275         self._sort_formats(formats)
 276
 277         ret = self._extract_theplatform_metadata(path, video_id)
 278         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 279         ret.update({
 280             'id': video_id,
 281             'formats': formats,
 282             'subtitles': combined_subtitles,
 283         })
 284
 285         return ret
 286
 287
 288 class ThePlatformFeedIE(ThePlatformBaseIE):
 289     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
 290     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[\w-]+))'
 291     _TESTS = [{
 292         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 293         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 294         'md5': '6e32495b5073ab414471b615c5ded394',
 295         'info_dict': {
 296             'id': 'n_hardball_5biden_140207',
 297             'ext': 'mp4',
 298             'title': 'The Biden factor: will Joe run in 2016?',
 299             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 300             'thumbnail': r're:^https?://.*\.jpg$',
 301             'upload_date': '20140208',
 302             'timestamp': 1391824260,
 303             'duration': 467.0,
 304             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 305             'uploader': 'NBCU-NEWS',
 306         },
 307     }]
 308
 309     def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}):
 310         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
 311         entry = self._download_json(real_url, video_id)['entries'][0]
 312
 313         formats = []
 314         subtitles = {}
 315         first_video_id = None
 316         duration = None
 317         asset_types = []
 318         for item in entry['media$content']:
 319             smil_url = item['plfile$url']
 320             cur_video_id = ThePlatformIE._match_id(smil_url)
 321             if first_video_id is None:
 322                 first_video_id = cur_video_id
 323                 duration = float_or_none(item.get('plfile$duration'))
 324             for asset_type in item['plfile$assetTypes']:
 325                 if asset_type in asset_types:
 326                     continue
 327                 asset_types.append(asset_type)
 328                 query = {
 329                     'mbr': 'true',
 330                     'formats': item['plfile$format'],
 331                     'assetTypes': asset_type,
 332                 }
 333                 if asset_type in asset_types_query:
 334                     query.update(asset_types_query[asset_type])
 335                 cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
 336                     smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
 337                 formats.extend(cur_formats)
 338                 subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 339
 340         self._sort_formats(formats)
 341
 342         thumbnails = [{
 343             'url': thumbnail['plfile$url'],
 344             'width': int_or_none(thumbnail.get('plfile$width')),
 345             'height': int_or_none(thumbnail.get('plfile$height')),
 346         } for thumbnail in entry.get('media$thumbnails', [])]
 347
 348         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 349         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 350
 351         ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 352         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 353         ret.update({
 354             'id': video_id,
 355             'formats': formats,
 356             'subtitles': subtitles,
 357             'thumbnails': thumbnails,
 358             'duration': duration,
 359             'timestamp': timestamp,
 360             'categories': categories,
 361         })
 362         if custom_fields:
 363             ret.update(custom_fields(entry))
 364
 365         return ret
 366
 367     def _real_extract(self, url):
 368         mobj = re.match(self._VALID_URL, url)
 369
 370         video_id = mobj.group('id')
 371         provider_id = mobj.group('provider_id')
 372         feed_id = mobj.group('feed_id')
 373         filter_query = mobj.group('filter')
 374
 375         return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)