_ 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     url_basename,
  24     xpath_with_ns,
  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(InfoExtractor):
  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)
  34         try:
  35             error_msg = next(
  36                 n.attrib['abstract']
  37                 for n in meta.findall(_x('.//smil:ref'))
  38                 if n.attrib.get('title') == 'Geographic Restriction' or n.attrib.get('title') == 'Expired')
  39         except StopIteration:
  40             pass
  41         else:
  42             raise ExtractorError(error_msg, expected=True)
  43
  44         formats = self._parse_smil_formats(
  45             meta, smil_url, video_id, namespace=default_ns,
  46             # the parameters are from syfy.com, other sites may use others,
  47             # they also work for nbc.com
  48             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  49             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  50
  51         for _format in formats:
  52             ext = determine_ext(_format['url'])
  53             if ext == 'once':
  54                 _format['ext'] = 'mp4'
  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': 'srt' if mime == 'text/srt' else 'ttml',
  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': '734f3790fb5fc4903da391beeebc4836',
 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             'categories': ['Today/Shows/Orange Room', 'Today/Sections/Money', 'Today/Topics/Tech', "Today/Topics/Editor's picks"],
 142         },
 143     }, {
 144         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 145         # geo-restricted (US), HLS encrypted with AES-128
 146         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 147         'only_matching': True,
 148     }]
 149
 150     @staticmethod
 151     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 152         flags = '10' if include_qs else '00'
 153         expiration_date = '%x' % (int(time.time()) + life)
 154
 155         def str_to_hex(str):
 156             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 157
 158         def hex_to_str(hex):
 159             return binascii.a2b_hex(hex)
 160
 161         relative_path = url.split('http://link.theplatform.com/s/')[1].split('?')[0]
 162         clear_text = hex_to_str(flags + expiration_date + str_to_hex(relative_path))
 163         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 164         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 165         return '%s&sig=%s' % (url, sig)
 166
 167     def _real_extract(self, url):
 168         url, smuggled_data = unsmuggle_url(url, {})
 169
 170         mobj = re.match(self._VALID_URL, url)
 171         provider_id = mobj.group('provider_id')
 172         video_id = mobj.group('id')
 173
 174         if not provider_id:
 175             provider_id = 'dJ5BDC'
 176
 177         path = provider_id
 178         if mobj.group('media'):
 179             path += '/media'
 180         path += '/' + video_id
 181
 182         qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
 183         if 'guid' in qs_dict:
 184             webpage = self._download_webpage(url, video_id)
 185             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 186             feed_id = None
 187             # feed id usually locates in the last script.
 188             # Seems there's no pattern for the interested script filename, so
 189             # I try one by one
 190             for script in reversed(scripts):
 191                 feed_script = self._download_webpage(
 192                     self._proto_relative_url(script, 'http:'),
 193                     video_id, 'Downloading feed script')
 194                 feed_id = self._search_regex(
 195                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 196                     'default feed id', default=None)
 197                 if feed_id is not None:
 198                     break
 199             if feed_id is None:
 200                 raise ExtractorError('Unable to find feed id')
 201             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 202                 provider_id, feed_id, qs_dict['guid'][0]))
 203
 204         if smuggled_data.get('force_smil_url', False):
 205             smil_url = url
 206         # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385)
 207         elif '/guid/' in url:
 208             headers = {}
 209             source_url = smuggled_data.get('source_url')
 210             if source_url:
 211                 headers['Referer'] = source_url
 212             request = sanitized_Request(url, headers=headers)
 213             webpage = self._download_webpage(request, video_id)
 214             smil_url = self._search_regex(
 215                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 216                 webpage, 'smil url', group='url')
 217             path = self._search_regex(
 218                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 219             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4&format=SMIL'
 220         elif mobj.group('config'):
 221             config_url = url + '&form=json'
 222             config_url = config_url.replace('swf/', 'config/')
 223             config_url = config_url.replace('onsite/', 'onsite/config/')
 224             config = self._download_json(config_url, video_id, 'Downloading config')
 225             if 'releaseUrl' in config:
 226                 release_url = config['releaseUrl']
 227             else:
 228                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 229             smil_url = release_url + '&format=SMIL&formats=MPEG4&manifest=f4m'
 230         else:
 231             smil_url = 'http://link.theplatform.com/s/%s/meta.smil?format=smil&mbr=true' % path
 232
 233         sig = smuggled_data.get('sig')
 234         if sig:
 235             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 236
 237         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 238
 239         ret = self.get_metadata(path, video_id)
 240         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 241         ret.update({
 242             'id': video_id,
 243             'formats': formats,
 244             'subtitles': combined_subtitles,
 245         })
 246
 247         return ret
 248
 249
 250 class ThePlatformFeedIE(ThePlatformBaseIE):
 251     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&byGuid=%s'
 252     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*byGuid=(?P<id>[a-zA-Z0-9_]+)'
 253     _TEST = {
 254         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 255         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 256         'md5': '22d2b84f058d3586efcd99e57d59d314',
 257         'info_dict': {
 258             'id': 'n_hardball_5biden_140207',
 259             'ext': 'mp4',
 260             'title': 'The Biden factor: will Joe run in 2016?',
 261             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 262             'thumbnail': 're:^https?://.*\.jpg$',
 263             'upload_date': '20140208',
 264             'timestamp': 1391824260,
 265             'duration': 467.0,
 266             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 267         },
 268     }
 269
 270     def _real_extract(self, url):
 271         mobj = re.match(self._VALID_URL, url)
 272
 273         video_id = mobj.group('id')
 274         provider_id = mobj.group('provider_id')
 275         feed_id = mobj.group('feed_id')
 276
 277         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, video_id)
 278         feed = self._download_json(real_url, video_id)
 279         entry = feed['entries'][0]
 280
 281         formats = []
 282         subtitles = {}
 283         first_video_id = None
 284         duration = None
 285         for item in entry['media$content']:
 286             smil_url = item['plfile$url'] + '&format=SMIL&Tracking=true&Embedded=true&formats=MPEG4,F4M'
 287             cur_video_id = url_basename(smil_url)
 288             if first_video_id is None:
 289                 first_video_id = cur_video_id
 290                 duration = float_or_none(item.get('plfile$duration'))
 291             cur_formats, cur_subtitles = self._extract_theplatform_smil(smil_url, video_id, 'Downloading SMIL data for %s' % cur_video_id)
 292             formats.extend(cur_formats)
 293             subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 294
 295         self._sort_formats(formats)
 296
 297         thumbnails = [{
 298             'url': thumbnail['plfile$url'],
 299             'width': int_or_none(thumbnail.get('plfile$width')),
 300             'height': int_or_none(thumbnail.get('plfile$height')),
 301         } for thumbnail in entry.get('media$thumbnails', [])]
 302
 303         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 304         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 305
 306         ret = self.get_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 307         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 308         ret.update({
 309             'id': video_id,
 310             'formats': formats,
 311             'subtitles': subtitles,
 312             'thumbnails': thumbnails,
 313             'duration': duration,
 314             'timestamp': timestamp,
 315             'categories': categories,
 316         })
 317
 318         return ret