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 ..compat import (
  13     compat_parse_qs,
  14     compat_urllib_parse_urlparse,
  15 )
  16 from ..utils import (
  17     determine_ext,
  18     ExtractorError,
  19     float_or_none,
  20     int_or_none,
  21     sanitized_Request,
  22     unsmuggle_url,
  23     update_url_query,
  24     xpath_with_ns,
  25     mimetype2ext,
  26     find_xpath_attr,
  27 )
  28
  29 default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  30 _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  31
  32
  33 class ThePlatformBaseIE(OnceIE):
  34     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  35         meta = self._download_xml(smil_url, video_id, note=note, query={'format': 'SMIL'})
  36         error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src')
  37         if error_element is not None and error_element.attrib['src'].startswith(
  38                 'http://link.theplatform.com/s/errorFiles/Unavailable.'):
  39             raise ExtractorError(error_element.attrib['abstract'], expected=True)
  40
  41         smil_formats = self._parse_smil_formats(
  42             meta, smil_url, video_id, namespace=default_ns,
  43             # the parameters are from syfy.com, other sites may use others,
  44             # they also work for nbc.com
  45             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  46             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  47
  48         formats = []
  49         for _format in smil_formats:
  50             if OnceIE.suitable(_format['url']):
  51                 formats.extend(self._extract_once_formats(_format['url']))
  52             else:
  53                 media_url = _format['url']
  54                 if determine_ext(media_url) == 'm3u8':
  55                     hdnea2 = self._get_cookies(media_url).get('hdnea2')
  56                     if hdnea2:
  57                         _format['url'] = update_url_query(media_url, {'hdnea3': hdnea2.value})
  58
  59                 formats.append(_format)
  60
  61         subtitles = self._parse_smil_subtitles(meta, default_ns)
  62
  63         return formats, subtitles
  64
  65     def get_metadata(self, path, video_id):
  66         info_url = 'http://link.theplatform.com/s/%s?format=preview' % path
  67         info = self._download_json(info_url, video_id)
  68
  69         subtitles = {}
  70         captions = info.get('captions')
  71         if isinstance(captions, list):
  72             for caption in captions:
  73                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  74                 subtitles[lang] = [{
  75                     'ext': mimetype2ext(mime),
  76                     'url': src,
  77                 }]
  78
  79         return {
  80             'title': info['title'],
  81             'subtitles': subtitles,
  82             'description': info['description'],
  83             'thumbnail': info['defaultThumbnailUrl'],
  84             'duration': int_or_none(info.get('duration'), 1000),
  85             'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
  86             'uploader': info.get('billingCode'),
  87         }
  88
  89
  90 class ThePlatformIE(ThePlatformBaseIE):
  91     _VALID_URL = r'''(?x)
  92         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
  93            (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
  94          |theplatform:)(?P<id>[^/\?&]+)'''
  95
  96     _TESTS = [{
  97         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
  98         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
  99         'info_dict': {
 100             'id': 'e9I_cZgTgIPd',
 101             'ext': 'flv',
 102             'title': 'Blackberry\'s big, bold Z30',
 103             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
 104             'duration': 247,
 105             'timestamp': 1383239700,
 106             'upload_date': '20131031',
 107             'uploader': 'CBSI-NEW',
 108         },
 109         'params': {
 110             # rtmp download
 111             'skip_download': True,
 112         },
 113     }, {
 114         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 115         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 116         'info_dict': {
 117             'id': '22d_qsQ6MIRT',
 118             'ext': 'flv',
 119             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 120             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 121             'timestamp': 1426176191,
 122             'upload_date': '20150312',
 123             'uploader': 'CBSI-NEW',
 124         },
 125         'params': {
 126             # rtmp download
 127             'skip_download': True,
 128         }
 129     }, {
 130         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 131         'info_dict': {
 132             'id': 'yMBg9E8KFxZD',
 133             'ext': 'mp4',
 134             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 135             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 136             'uploader': 'EGSM',
 137         }
 138     }, {
 139         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 140         'only_matching': True,
 141     }, {
 142         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 143         'md5': 'fb96bb3d85118930a5b055783a3bd992',
 144         'info_dict': {
 145             'id': 'tdy_or_siri_150701',
 146             'ext': 'mp4',
 147             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 148             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 149             'duration': 83.0,
 150             'thumbnail': 're:^https?://.*\.jpg$',
 151             'timestamp': 1435752600,
 152             'upload_date': '20150701',
 153             'uploader': 'NBCU-NEWS',
 154         },
 155     }, {
 156         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 157         # geo-restricted (US), HLS encrypted with AES-128
 158         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 159         'only_matching': True,
 160     }]
 161
 162     @classmethod
 163     def _extract_urls(cls, webpage):
 164         m = re.search(
 165             r'''(?x)
 166                     <meta\s+
 167                         property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
 168                         content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2
 169             ''', webpage)
 170         if m:
 171             return [m.group('url')]
 172
 173         matches = re.findall(
 174             r'<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage)
 175         if matches:
 176             return list(zip(*matches))[1]
 177
 178     @staticmethod
 179     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 180         flags = '10' if include_qs else '00'
 181         expiration_date = '%x' % (int(time.time()) + life)
 182
 183         def str_to_hex(str):
 184             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 185
 186         def hex_to_bytes(hex):
 187             return binascii.a2b_hex(hex.encode('ascii'))
 188
 189         relative_path = re.match(r'https?://link.theplatform.com/s/([^?]+)', url).group(1)
 190         clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path))
 191         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 192         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 193         return '%s&sig=%s' % (url, sig)
 194
 195     def _real_extract(self, url):
 196         url, smuggled_data = unsmuggle_url(url, {})
 197
 198         mobj = re.match(self._VALID_URL, url)
 199         provider_id = mobj.group('provider_id')
 200         video_id = mobj.group('id')
 201
 202         if not provider_id:
 203             provider_id = 'dJ5BDC'
 204
 205         path = provider_id + '/'
 206         if mobj.group('media'):
 207             path += mobj.group('media')
 208         path += video_id
 209
 210         qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
 211         if 'guid' in qs_dict:
 212             webpage = self._download_webpage(url, video_id)
 213             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 214             feed_id = None
 215             # feed id usually locates in the last script.
 216             # Seems there's no pattern for the interested script filename, so
 217             # I try one by one
 218             for script in reversed(scripts):
 219                 feed_script = self._download_webpage(
 220                     self._proto_relative_url(script, 'http:'),
 221                     video_id, 'Downloading feed script')
 222                 feed_id = self._search_regex(
 223                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 224                     'default feed id', default=None)
 225                 if feed_id is not None:
 226                     break
 227             if feed_id is None:
 228                 raise ExtractorError('Unable to find feed id')
 229             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 230                 provider_id, feed_id, qs_dict['guid'][0]))
 231
 232         if smuggled_data.get('force_smil_url', False):
 233             smil_url = url
 234         # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385)
 235         elif '/guid/' in url:
 236             headers = {}
 237             source_url = smuggled_data.get('source_url')
 238             if source_url:
 239                 headers['Referer'] = source_url
 240             request = sanitized_Request(url, headers=headers)
 241             webpage = self._download_webpage(request, video_id)
 242             smil_url = self._search_regex(
 243                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 244                 webpage, 'smil url', group='url')
 245             path = self._search_regex(
 246                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 247             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4'
 248         elif mobj.group('config'):
 249             config_url = url + '&form=json'
 250             config_url = config_url.replace('swf/', 'config/')
 251             config_url = config_url.replace('onsite/', 'onsite/config/')
 252             config = self._download_json(config_url, video_id, 'Downloading config')
 253             if 'releaseUrl' in config:
 254                 release_url = config['releaseUrl']
 255             else:
 256                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 257             smil_url = release_url + '&formats=MPEG4&manifest=f4m'
 258         else:
 259             smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 260
 261         sig = smuggled_data.get('sig')
 262         if sig:
 263             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 264
 265         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 266         self._sort_formats(formats)
 267
 268         ret = self.get_metadata(path, video_id)
 269         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 270         ret.update({
 271             'id': video_id,
 272             'formats': formats,
 273             'subtitles': combined_subtitles,
 274         })
 275
 276         return ret
 277
 278
 279 class ThePlatformFeedIE(ThePlatformBaseIE):
 280     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
 281     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[\w-]+))'
 282     _TESTS = [{
 283         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 284         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 285         'md5': '6e32495b5073ab414471b615c5ded394',
 286         'info_dict': {
 287             'id': 'n_hardball_5biden_140207',
 288             'ext': 'mp4',
 289             'title': 'The Biden factor: will Joe run in 2016?',
 290             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 291             'thumbnail': 're:^https?://.*\.jpg$',
 292             'upload_date': '20140208',
 293             'timestamp': 1391824260,
 294             'duration': 467.0,
 295             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 296             'uploader': 'NBCU-NEWS',
 297         },
 298     }]
 299
 300     def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}):
 301         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
 302         entry = self._download_json(real_url, video_id)['entries'][0]
 303
 304         formats = []
 305         subtitles = {}
 306         first_video_id = None
 307         duration = None
 308         asset_types = []
 309         for item in entry['media$content']:
 310             smil_url = item['plfile$url']
 311             cur_video_id = ThePlatformIE._match_id(smil_url)
 312             if first_video_id is None:
 313                 first_video_id = cur_video_id
 314                 duration = float_or_none(item.get('plfile$duration'))
 315             for asset_type in item['plfile$assetTypes']:
 316                 if asset_type in asset_types:
 317                     continue
 318                 asset_types.append(asset_type)
 319                 query = {
 320                     'mbr': 'true',
 321                     'formats': item['plfile$format'],
 322                     'assetTypes': asset_type,
 323                 }
 324                 if asset_type in asset_types_query:
 325                     query.update(asset_types_query[asset_type])
 326                 cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
 327                     smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
 328                 formats.extend(cur_formats)
 329                 subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 330
 331         self._sort_formats(formats)
 332
 333         thumbnails = [{
 334             'url': thumbnail['plfile$url'],
 335             'width': int_or_none(thumbnail.get('plfile$width')),
 336             'height': int_or_none(thumbnail.get('plfile$height')),
 337         } for thumbnail in entry.get('media$thumbnails', [])]
 338
 339         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 340         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 341
 342         ret = self.get_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 343         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 344         ret.update({
 345             'id': video_id,
 346             'formats': formats,
 347             'subtitles': subtitles,
 348             'thumbnails': thumbnails,
 349             'duration': duration,
 350             'timestamp': timestamp,
 351             'categories': categories,
 352         })
 353         if custom_fields:
 354             ret.update(custom_fields(entry))
 355
 356         return ret
 357
 358     def _real_extract(self, url):
 359         mobj = re.match(self._VALID_URL, url)
 360
 361         video_id = mobj.group('id')
 362         provider_id = mobj.group('provider_id')
 363         feed_id = mobj.group('feed_id')
 364         filter_query = mobj.group('filter')
 365
 366         return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)