_ 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 .common import InfoExtractor
  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     xpath_with_ns,
  24     mimetype2ext,
  25     find_xpath_attr,
  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(InfoExtractor):
  33     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  34         meta = self._download_xml(smil_url, video_id, note=note)
  35         error_element = find_xpath_attr(
  36             meta, _x('.//smil:ref'), 'src',
  37             'http://link.theplatform.com/s/errorFiles/Unavailable.mp4')
  38         if error_element is not None:
  39             raise ExtractorError(error_element.attrib['abstract'], expected=True)
  40
  41         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         for _format in formats:
  49             ext = determine_ext(_format['url'])
  50             if ext == 'once':
  51                 _format['ext'] = 'mp4'
  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         }
  80
  81
  82 class ThePlatformIE(ThePlatformBaseIE):
  83     _VALID_URL = r'''(?x)
  84         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
  85            (?:(?P<media>(?:(?:[^/]+/)+select/)?media/)|(?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         },
  98         'params': {
  99             # rtmp download
 100             'skip_download': True,
 101         },
 102     }, {
 103         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 104         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 105         'info_dict': {
 106             'id': '22d_qsQ6MIRT',
 107             'ext': 'flv',
 108             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 109             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 110         },
 111         'params': {
 112             # rtmp download
 113             'skip_download': True,
 114         }
 115     }, {
 116         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 117         'info_dict': {
 118             'id': 'yMBg9E8KFxZD',
 119             'ext': 'mp4',
 120             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 121             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 122         }
 123     }, {
 124         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 125         'only_matching': True,
 126     }, {
 127         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 128         'md5': '734f3790fb5fc4903da391beeebc4836',
 129         'info_dict': {
 130             'id': 'tdy_or_siri_150701',
 131             'ext': 'mp4',
 132             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 133             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 134             'duration': 83.0,
 135             'thumbnail': 're:^https?://.*\.jpg$',
 136             'timestamp': 1435752600,
 137             'upload_date': '20150701',
 138             'categories': ['Today/Shows/Orange Room', 'Today/Sections/Money', 'Today/Topics/Tech', "Today/Topics/Editor's picks"],
 139         },
 140     }, {
 141         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 142         # geo-restricted (US), HLS encrypted with AES-128
 143         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 144         'only_matching': True,
 145     }]
 146
 147     @staticmethod
 148     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 149         flags = '10' if include_qs else '00'
 150         expiration_date = '%x' % (int(time.time()) + life)
 151
 152         def str_to_hex(str):
 153             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 154
 155         def hex_to_str(hex):
 156             return binascii.a2b_hex(hex)
 157
 158         relative_path = url.split('http://link.theplatform.com/s/')[1].split('?')[0]
 159         clear_text = hex_to_str(flags + expiration_date + str_to_hex(relative_path))
 160         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 161         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 162         return '%s&sig=%s' % (url, sig)
 163
 164     def _real_extract(self, url):
 165         url, smuggled_data = unsmuggle_url(url, {})
 166
 167         mobj = re.match(self._VALID_URL, url)
 168         provider_id = mobj.group('provider_id')
 169         video_id = mobj.group('id')
 170
 171         if not provider_id:
 172             provider_id = 'dJ5BDC'
 173
 174         path = provider_id
 175         if mobj.group('media'):
 176             path += '/media'
 177         path += '/' + video_id
 178
 179         qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
 180         if 'guid' in qs_dict:
 181             webpage = self._download_webpage(url, video_id)
 182             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 183             feed_id = None
 184             # feed id usually locates in the last script.
 185             # Seems there's no pattern for the interested script filename, so
 186             # I try one by one
 187             for script in reversed(scripts):
 188                 feed_script = self._download_webpage(
 189                     self._proto_relative_url(script, 'http:'),
 190                     video_id, 'Downloading feed script')
 191                 feed_id = self._search_regex(
 192                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 193                     'default feed id', default=None)
 194                 if feed_id is not None:
 195                     break
 196             if feed_id is None:
 197                 raise ExtractorError('Unable to find feed id')
 198             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 199                 provider_id, feed_id, qs_dict['guid'][0]))
 200
 201         if smuggled_data.get('force_smil_url', False):
 202             smil_url = url
 203         # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385)
 204         elif '/guid/' in url:
 205             headers = {}
 206             source_url = smuggled_data.get('source_url')
 207             if source_url:
 208                 headers['Referer'] = source_url
 209             request = sanitized_Request(url, headers=headers)
 210             webpage = self._download_webpage(request, video_id)
 211             smil_url = self._search_regex(
 212                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 213                 webpage, 'smil url', group='url')
 214             path = self._search_regex(
 215                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 216             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4&format=SMIL'
 217         elif mobj.group('config'):
 218             config_url = url + '&form=json'
 219             config_url = config_url.replace('swf/', 'config/')
 220             config_url = config_url.replace('onsite/', 'onsite/config/')
 221             config = self._download_json(config_url, video_id, 'Downloading config')
 222             if 'releaseUrl' in config:
 223                 release_url = config['releaseUrl']
 224             else:
 225                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 226             smil_url = release_url + '&format=SMIL&formats=MPEG4&manifest=f4m'
 227         else:
 228             smil_url = 'http://link.theplatform.com/s/%s/meta.smil?format=smil&mbr=true' % path
 229
 230         sig = smuggled_data.get('sig')
 231         if sig:
 232             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 233
 234         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 235
 236         ret = self.get_metadata(path, video_id)
 237         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 238         ret.update({
 239             'id': video_id,
 240             'formats': formats,
 241             'subtitles': combined_subtitles,
 242         })
 243
 244         return ret
 245
 246
 247 class ThePlatformFeedIE(ThePlatformBaseIE):
 248     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&byGuid=%s'
 249     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*byGuid=(?P<id>[a-zA-Z0-9_]+)'
 250     _TEST = {
 251         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 252         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 253         'md5': '22d2b84f058d3586efcd99e57d59d314',
 254         'info_dict': {
 255             'id': 'n_hardball_5biden_140207',
 256             'ext': 'mp4',
 257             'title': 'The Biden factor: will Joe run in 2016?',
 258             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 259             'thumbnail': 're:^https?://.*\.jpg$',
 260             'upload_date': '20140208',
 261             'timestamp': 1391824260,
 262             'duration': 467.0,
 263             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 264         },
 265     }
 266
 267     def _real_extract(self, url):
 268         mobj = re.match(self._VALID_URL, url)
 269
 270         video_id = mobj.group('id')
 271         provider_id = mobj.group('provider_id')
 272         feed_id = mobj.group('feed_id')
 273
 274         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, video_id)
 275         feed = self._download_json(real_url, video_id)
 276         entry = feed['entries'][0]
 277
 278         formats = []
 279         subtitles = {}
 280         first_video_id = None
 281         duration = None
 282         for item in entry['media$content']:
 283             smil_url = item['plfile$url'] + '&format=SMIL&mbr=true'
 284             cur_video_id = ThePlatformIE._match_id(smil_url)
 285             if first_video_id is None:
 286                 first_video_id = cur_video_id
 287                 duration = float_or_none(item.get('plfile$duration'))
 288             cur_formats, cur_subtitles = self._extract_theplatform_smil(smil_url, video_id, 'Downloading SMIL data for %s' % cur_video_id)
 289             formats.extend(cur_formats)
 290             subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 291
 292         self._sort_formats(formats)
 293
 294         thumbnails = [{
 295             'url': thumbnail['plfile$url'],
 296             'width': int_or_none(thumbnail.get('plfile$width')),
 297             'height': int_or_none(thumbnail.get('plfile$height')),
 298         } for thumbnail in entry.get('media$thumbnails', [])]
 299
 300         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 301         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 302
 303         ret = self.get_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 304         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 305         ret.update({
 306             'id': video_id,
 307             'formats': formats,
 308             'subtitles': subtitles,
 309             'thumbnails': thumbnails,
 310             'duration': duration,
 311             'timestamp': timestamp,
 312             'categories': categories,
 313         })
 314
 315         return ret