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