_ 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 ..compat import (
  13     compat_parse_qs,
  14     compat_urllib_parse_urlparse,
  15 )
  16 from ..utils import (
  17     ExtractorError,
  18     float_or_none,
  19     int_or_none,
  20     sanitized_Request,
  21     unsmuggle_url,
  22     xpath_with_ns,
  23     mimetype2ext,
  24     find_xpath_attr,
  25 )
  26
  27 default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  28 _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  29
  30
  31 class ThePlatformBaseIE(OnceIE):
  32     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  33         meta = self._download_xml(smil_url, video_id, note=note, query={'format': 'SMIL'})
  34         error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src')
  35         if error_element is not None and error_element.attrib['src'].startswith(
  36                 'http://link.theplatform.com/s/errorFiles/Unavailable.'):
  37             raise ExtractorError(error_element.attrib['abstract'], expected=True)
  38
  39         smil_formats = self._parse_smil_formats(
  40             meta, smil_url, video_id, namespace=default_ns,
  41             # the parameters are from syfy.com, other sites may use others,
  42             # they also work for nbc.com
  43             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  44             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  45
  46         formats = []
  47         for _format in smil_formats:
  48             if OnceIE.suitable(_format['url']):
  49                 formats.extend(self._extract_once_formats(_format['url']))
  50             else:
  51                 formats.append(_format)
  52
  53         self._sort_formats(formats)
  54
  55         subtitles = self._parse_smil_subtitles(meta, default_ns)
  56
  57         return formats, subtitles
  58
  59     def get_metadata(self, path, video_id):
  60         info_url = 'http://link.theplatform.com/s/%s?format=preview' % path
  61         info = self._download_json(info_url, video_id)
  62
  63         subtitles = {}
  64         captions = info.get('captions')
  65         if isinstance(captions, list):
  66             for caption in captions:
  67                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  68                 subtitles[lang] = [{
  69                     'ext': mimetype2ext(mime),
  70                     'url': src,
  71                 }]
  72
  73         return {
  74             'title': info['title'],
  75             'subtitles': subtitles,
  76             'description': info['description'],
  77             'thumbnail': info['defaultThumbnailUrl'],
  78             'duration': int_or_none(info.get('duration'), 1000),
  79             'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
  80             'uploader': info.get('billingCode'),
  81         }
  82
  83
  84 class ThePlatformIE(ThePlatformBaseIE):
  85     _VALID_URL = r'''(?x)
  86         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
  87            (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
  88          |theplatform:)(?P<id>[^/\?&]+)'''
  89
  90     _TESTS = [{
  91         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
  92         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
  93         'info_dict': {
  94             'id': 'e9I_cZgTgIPd',
  95             'ext': 'flv',
  96             'title': 'Blackberry\'s big, bold Z30',
  97             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
  98             'duration': 247,
  99             'timestamp': 1383239700,
 100             'upload_date': '20131031',
 101             'uploader': 'CBSI-NEW',
 102         },
 103         'params': {
 104             # rtmp download
 105             'skip_download': True,
 106         },
 107     }, {
 108         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 109         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 110         'info_dict': {
 111             'id': '22d_qsQ6MIRT',
 112             'ext': 'flv',
 113             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 114             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 115             'timestamp': 1426176191,
 116             'upload_date': '20150312',
 117             'uploader': 'CBSI-NEW',
 118         },
 119         'params': {
 120             # rtmp download
 121             'skip_download': True,
 122         }
 123     }, {
 124         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 125         'info_dict': {
 126             'id': 'yMBg9E8KFxZD',
 127             'ext': 'mp4',
 128             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 129             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 130             'uploader': 'EGSM',
 131         }
 132     }, {
 133         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 134         'only_matching': True,
 135     }, {
 136         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 137         'md5': 'fb96bb3d85118930a5b055783a3bd992',
 138         'info_dict': {
 139             'id': 'tdy_or_siri_150701',
 140             'ext': 'mp4',
 141             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 142             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 143             'duration': 83.0,
 144             'thumbnail': 're:^https?://.*\.jpg$',
 145             'timestamp': 1435752600,
 146             'upload_date': '20150701',
 147             'uploader': 'NBCU-NEWS',
 148         },
 149     }, {
 150         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 151         # geo-restricted (US), HLS encrypted with AES-128
 152         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 153         'only_matching': True,
 154     }]
 155
 156     @staticmethod
 157     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 158         flags = '10' if include_qs else '00'
 159         expiration_date = '%x' % (int(time.time()) + life)
 160
 161         def str_to_hex(str):
 162             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 163
 164         def hex_to_str(hex):
 165             return binascii.a2b_hex(hex)
 166
 167         relative_path = re.match(r'https?://link.theplatform.com/s/([^?]+)', url).group(1)
 168         clear_text = hex_to_str(flags + expiration_date + str_to_hex(relative_path))
 169         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 170         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 171         return '%s&sig=%s' % (url, sig)
 172
 173     def _real_extract(self, url):
 174         url, smuggled_data = unsmuggle_url(url, {})
 175
 176         mobj = re.match(self._VALID_URL, url)
 177         provider_id = mobj.group('provider_id')
 178         video_id = mobj.group('id')
 179
 180         if not provider_id:
 181             provider_id = 'dJ5BDC'
 182
 183         path = provider_id + '/'
 184         if mobj.group('media'):
 185             path += mobj.group('media')
 186         path += video_id
 187
 188         qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
 189         if 'guid' in qs_dict:
 190             webpage = self._download_webpage(url, video_id)
 191             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 192             feed_id = None
 193             # feed id usually locates in the last script.
 194             # Seems there's no pattern for the interested script filename, so
 195             # I try one by one
 196             for script in reversed(scripts):
 197                 feed_script = self._download_webpage(
 198                     self._proto_relative_url(script, 'http:'),
 199                     video_id, 'Downloading feed script')
 200                 feed_id = self._search_regex(
 201                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 202                     'default feed id', default=None)
 203                 if feed_id is not None:
 204                     break
 205             if feed_id is None:
 206                 raise ExtractorError('Unable to find feed id')
 207             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 208                 provider_id, feed_id, qs_dict['guid'][0]))
 209
 210         if smuggled_data.get('force_smil_url', False):
 211             smil_url = url
 212         # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385)
 213         elif '/guid/' in url:
 214             headers = {}
 215             source_url = smuggled_data.get('source_url')
 216             if source_url:
 217                 headers['Referer'] = source_url
 218             request = sanitized_Request(url, headers=headers)
 219             webpage = self._download_webpage(request, video_id)
 220             smil_url = self._search_regex(
 221                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 222                 webpage, 'smil url', group='url')
 223             path = self._search_regex(
 224                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 225             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4'
 226         elif mobj.group('config'):
 227             config_url = url + '&form=json'
 228             config_url = config_url.replace('swf/', 'config/')
 229             config_url = config_url.replace('onsite/', 'onsite/config/')
 230             config = self._download_json(config_url, video_id, 'Downloading config')
 231             if 'releaseUrl' in config:
 232                 release_url = config['releaseUrl']
 233             else:
 234                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 235             smil_url = release_url + '&formats=MPEG4&manifest=f4m'
 236         else:
 237             smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 238
 239         sig = smuggled_data.get('sig')
 240         if sig:
 241             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 242
 243         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 244
 245         ret = self.get_metadata(path, video_id)
 246         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 247         ret.update({
 248             'id': video_id,
 249             'formats': formats,
 250             'subtitles': combined_subtitles,
 251         })
 252
 253         return ret
 254
 255
 256 class ThePlatformFeedIE(ThePlatformBaseIE):
 257     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&byGuid=%s'
 258     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*byGuid=(?P<id>[a-zA-Z0-9_]+)'
 259     _TEST = {
 260         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 261         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 262         'md5': '6e32495b5073ab414471b615c5ded394',
 263         'info_dict': {
 264             'id': 'n_hardball_5biden_140207',
 265             'ext': 'mp4',
 266             'title': 'The Biden factor: will Joe run in 2016?',
 267             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 268             'thumbnail': 're:^https?://.*\.jpg$',
 269             'upload_date': '20140208',
 270             'timestamp': 1391824260,
 271             'duration': 467.0,
 272             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 273         },
 274     }
 275
 276     def _real_extract(self, url):
 277         mobj = re.match(self._VALID_URL, url)
 278
 279         video_id = mobj.group('id')
 280         provider_id = mobj.group('provider_id')
 281         feed_id = mobj.group('feed_id')
 282
 283         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, video_id)
 284         feed = self._download_json(real_url, video_id)
 285         entry = feed['entries'][0]
 286
 287         formats = []
 288         subtitles = {}
 289         first_video_id = None
 290         duration = None
 291         for item in entry['media$content']:
 292             smil_url = item['plfile$url'] + '&mbr=true'
 293             cur_video_id = ThePlatformIE._match_id(smil_url)
 294             if first_video_id is None:
 295                 first_video_id = cur_video_id
 296                 duration = float_or_none(item.get('plfile$duration'))
 297             cur_formats, cur_subtitles = self._extract_theplatform_smil(smil_url, video_id, 'Downloading SMIL data for %s' % cur_video_id)
 298             formats.extend(cur_formats)
 299             subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 300
 301         self._sort_formats(formats)
 302
 303         thumbnails = [{
 304             'url': thumbnail['plfile$url'],
 305             'width': int_or_none(thumbnail.get('plfile$width')),
 306             'height': int_or_none(thumbnail.get('plfile$height')),
 307         } for thumbnail in entry.get('media$thumbnails', [])]
 308
 309         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 310         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 311
 312         ret = self.get_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 313         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 314         ret.update({
 315             'id': video_id,
 316             'formats': formats,
 317             'subtitles': subtitles,
 318             'thumbnails': thumbnails,
 319             'duration': duration,
 320             'timestamp': timestamp,
 321             'categories': categories,
 322         })
 323
 324         return ret