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