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