Raphaël G. Git Repositories - youtubedl/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 .adobepass import AdobePassIE
  13 from ..compat import (
  14     compat_parse_qs,
  15     compat_urllib_parse_urlparse,
  16 )
  17 from ..utils import (
  18     determine_ext,
  19     ExtractorError,
  20     float_or_none,
  21     int_or_none,
  22     sanitized_Request,
  23     unsmuggle_url,
  24     update_url_query,
  25     xpath_with_ns,
  26     mimetype2ext,
  27     find_xpath_attr,
  28 )
  29
  30 default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  31 _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  32
  33
  34 class ThePlatformBaseIE(OnceIE):
  35     _TP_TLD = 'com'
  36
  37     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  38         meta = self._download_xml(
  39             smil_url, video_id, note=note, query={'format': 'SMIL'},
  40             headers=self.geo_verification_headers())
  41         error_element = find_xpath_attr(meta, _x('.//smil:ref'), 'src')
  42         if error_element is not None and error_element.attrib['src'].startswith(
  43                 'http://link.theplatform.%s/s/errorFiles/Unavailable.' % self._TP_TLD):
  44             raise ExtractorError(error_element.attrib['abstract'], expected=True)
  45
  46         smil_formats = self._parse_smil_formats(
  47             meta, smil_url, video_id, namespace=default_ns,
  48             # the parameters are from syfy.com, other sites may use others,
  49             # they also work for nbc.com
  50             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  51             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  52
  53         formats = []
  54         for _format in smil_formats:
  55             if OnceIE.suitable(_format['url']):
  56                 formats.extend(self._extract_once_formats(_format['url']))
  57             else:
  58                 media_url = _format['url']
  59                 if determine_ext(media_url) == 'm3u8':
  60                     hdnea2 = self._get_cookies(media_url).get('hdnea2')
  61                     if hdnea2:
  62                         _format['url'] = update_url_query(media_url, {'hdnea3': hdnea2.value})
  63
  64                 formats.append(_format)
  65
  66         subtitles = self._parse_smil_subtitles(meta, default_ns)
  67
  68         return formats, subtitles
  69
  70     def _download_theplatform_metadata(self, path, video_id):
  71         info_url = 'http://link.theplatform.%s/s/%s?format=preview' % (self._TP_TLD, path)
  72         return self._download_json(info_url, video_id)
  73
  74     def _parse_theplatform_metadata(self, info):
  75         subtitles = {}
  76         captions = info.get('captions')
  77         if isinstance(captions, list):
  78             for caption in captions:
  79                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  80                 subtitles.setdefault(lang, []).append({
  81                     'ext': mimetype2ext(mime),
  82                     'url': src,
  83                 })
  84
  85         duration = info.get('duration')
  86         tp_chapters = info.get('chapters', [])
  87         chapters = []
  88         if tp_chapters:
  89             def _add_chapter(start_time, end_time):
  90                 start_time = float_or_none(start_time, 1000)
  91                 end_time = float_or_none(end_time, 1000)
  92                 if start_time is None or end_time is None:
  93                     return
  94                 chapters.append({
  95                     'start_time': start_time,
  96                     'end_time': end_time,
  97                 })
  98
  99             for chapter in tp_chapters[:-1]:
 100                 _add_chapter(chapter.get('startTime'), chapter.get('endTime'))
 101             _add_chapter(tp_chapters[-1].get('startTime'), tp_chapters[-1].get('endTime') or duration)
 102
 103         return {
 104             'title': info['title'],
 105             'subtitles': subtitles,
 106             'description': info['description'],
 107             'thumbnail': info['defaultThumbnailUrl'],
 108             'duration': float_or_none(duration, 1000),
 109             'timestamp': int_or_none(info.get('pubDate'), 1000) or None,
 110             'uploader': info.get('billingCode'),
 111             'chapters': chapters,
 112         }
 113
 114     def _extract_theplatform_metadata(self, path, video_id):
 115         info = self._download_theplatform_metadata(path, video_id)
 116         return self._parse_theplatform_metadata(info)
 117
 118
 119 class ThePlatformIE(ThePlatformBaseIE, AdobePassIE):
 120     _VALID_URL = r'''(?x)
 121         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
 122            (?:(?:(?:[^/]+/)+select/)?(?P<media>media/(?:guid/\d+/)?)?|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
 123          |theplatform:)(?P<id>[^/\?&]+)'''
 124
 125     _TESTS = [{
 126         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
 127         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
 128         'info_dict': {
 129             'id': 'e9I_cZgTgIPd',
 130             'ext': 'flv',
 131             'title': 'Blackberry\'s big, bold Z30',
 132             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
 133             'duration': 247,
 134             'timestamp': 1383239700,
 135             'upload_date': '20131031',
 136             'uploader': 'CBSI-NEW',
 137         },
 138         'params': {
 139             # rtmp download
 140             'skip_download': True,
 141         },
 142         'skip': '404 Not Found',
 143     }, {
 144         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 145         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 146         'info_dict': {
 147             'id': '22d_qsQ6MIRT',
 148             'ext': 'flv',
 149             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 150             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 151             'timestamp': 1426176191,
 152             'upload_date': '20150312',
 153             'uploader': 'CBSI-NEW',
 154         },
 155         'params': {
 156             # rtmp download
 157             'skip_download': True,
 158         }
 159     }, {
 160         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 161         'info_dict': {
 162             'id': 'yMBg9E8KFxZD',
 163             'ext': 'mp4',
 164             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 165             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 166             'uploader': 'EGSM',
 167         }
 168     }, {
 169         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 170         'only_matching': True,
 171     }, {
 172         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 173         'md5': 'fb96bb3d85118930a5b055783a3bd992',
 174         'info_dict': {
 175             'id': 'tdy_or_siri_150701',
 176             'ext': 'mp4',
 177             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 178             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 179             'duration': 83.0,
 180             'thumbnail': r're:^https?://.*\.jpg$',
 181             'timestamp': 1435752600,
 182             'upload_date': '20150701',
 183             'uploader': 'NBCU-NEWS',
 184         },
 185     }, {
 186         # From http://www.nbc.com/the-blacklist/video/sir-crispin-crandall/2928790?onid=137781#vc137781=1
 187         # geo-restricted (US), HLS encrypted with AES-128
 188         'url': 'http://player.theplatform.com/p/NnzsPC/onsite_universal/select/media/guid/2410887629/2928790?fwsitesection=nbc_the_blacklist_video_library&autoPlay=true&carouselID=137781',
 189         'only_matching': True,
 190     }]
 191
 192     @classmethod
 193     def _extract_urls(cls, webpage):
 194         m = re.search(
 195             r'''(?x)
 196                     <meta\s+
 197                         property=(["'])(?:og:video(?::(?:secure_)?url)?|twitter:player)\1\s+
 198                         content=(["'])(?P<url>https?://player\.theplatform\.com/p/.+?)\2
 199             ''', webpage)
 200         if m:
 201             return [m.group('url')]
 202
 203         # Are whitesapces ignored in URLs?
 204         # https://github.com/rg3/youtube-dl/issues/12044
 205         matches = re.findall(
 206             r'(?s)<(?:iframe|script)[^>]+src=(["\'])((?:https?:)?//player\.theplatform\.com/p/.+?)\1', webpage)
 207         if matches:
 208             return [re.sub(r'\s', '', list(zip(*matches))[1][0])]
 209
 210     @staticmethod
 211     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 212         flags = '10' if include_qs else '00'
 213         expiration_date = '%x' % (int(time.time()) + life)
 214
 215         def str_to_hex(str):
 216             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 217
 218         def hex_to_bytes(hex):
 219             return binascii.a2b_hex(hex.encode('ascii'))
 220
 221         relative_path = re.match(r'https?://link\.theplatform\.com/s/([^?]+)', url).group(1)
 222         clear_text = hex_to_bytes(flags + expiration_date + str_to_hex(relative_path))
 223         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 224         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 225         return '%s&sig=%s' % (url, sig)
 226
 227     def _real_extract(self, url):
 228         url, smuggled_data = unsmuggle_url(url, {})
 229
 230         mobj = re.match(self._VALID_URL, url)
 231         provider_id = mobj.group('provider_id')
 232         video_id = mobj.group('id')
 233
 234         if not provider_id:
 235             provider_id = 'dJ5BDC'
 236
 237         path = provider_id + '/'
 238         if mobj.group('media'):
 239             path += mobj.group('media')
 240         path += video_id
 241
 242         qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
 243         if 'guid' in qs_dict:
 244             webpage = self._download_webpage(url, video_id)
 245             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 246             feed_id = None
 247             # feed id usually locates in the last script.
 248             # Seems there's no pattern for the interested script filename, so
 249             # I try one by one
 250             for script in reversed(scripts):
 251                 feed_script = self._download_webpage(
 252                     self._proto_relative_url(script, 'http:'),
 253                     video_id, 'Downloading feed script')
 254                 feed_id = self._search_regex(
 255                     r'defaultFeedId\s*:\s*"([^"]+)"', feed_script,
 256                     'default feed id', default=None)
 257                 if feed_id is not None:
 258                     break
 259             if feed_id is None:
 260                 raise ExtractorError('Unable to find feed id')
 261             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 262                 provider_id, feed_id, qs_dict['guid'][0]))
 263
 264         if smuggled_data.get('force_smil_url', False):
 265             smil_url = url
 266         # Explicitly specified SMIL (see https://github.com/rg3/youtube-dl/issues/7385)
 267         elif '/guid/' in url:
 268             headers = {}
 269             source_url = smuggled_data.get('source_url')
 270             if source_url:
 271                 headers['Referer'] = source_url
 272             request = sanitized_Request(url, headers=headers)
 273             webpage = self._download_webpage(request, video_id)
 274             smil_url = self._search_regex(
 275                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 276                 webpage, 'smil url', group='url')
 277             path = self._search_regex(
 278                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 279             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4'
 280         elif mobj.group('config'):
 281             config_url = url + '&form=json'
 282             config_url = config_url.replace('swf/', 'config/')
 283             config_url = config_url.replace('onsite/', 'onsite/config/')
 284             config = self._download_json(config_url, video_id, 'Downloading config')
 285             if 'releaseUrl' in config:
 286                 release_url = config['releaseUrl']
 287             else:
 288                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 289             smil_url = release_url + '&formats=MPEG4&manifest=f4m'
 290         else:
 291             smil_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 292
 293         sig = smuggled_data.get('sig')
 294         if sig:
 295             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 296
 297         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 298         self._sort_formats(formats)
 299
 300         ret = self._extract_theplatform_metadata(path, video_id)
 301         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 302         ret.update({
 303             'id': video_id,
 304             'formats': formats,
 305             'subtitles': combined_subtitles,
 306         })
 307
 308         return ret
 309
 310
 311 class ThePlatformFeedIE(ThePlatformBaseIE):
 312     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&%s'
 313     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*(?P<filter>by(?:Gui|I)d=(?P<id>[^&]+))'
 314     _TESTS = [{
 315         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 316         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 317         'md5': '6e32495b5073ab414471b615c5ded394',
 318         'info_dict': {
 319             'id': 'n_hardball_5biden_140207',
 320             'ext': 'mp4',
 321             'title': 'The Biden factor: will Joe run in 2016?',
 322             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 323             'thumbnail': r're:^https?://.*\.jpg$',
 324             'upload_date': '20140208',
 325             'timestamp': 1391824260,
 326             'duration': 467.0,
 327             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 328             'uploader': 'NBCU-NEWS',
 329         },
 330     }, {
 331         'url': 'http://feed.theplatform.com/f/2E2eJC/nnd_NBCNews?byGuid=nn_netcast_180306.Copy.01',
 332         'only_matching': True,
 333     }]
 334
 335     def _extract_feed_info(self, provider_id, feed_id, filter_query, video_id, custom_fields=None, asset_types_query={}, account_id=None):
 336         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, filter_query)
 337         entry = self._download_json(real_url, video_id)['entries'][0]
 338         main_smil_url = 'http://link.theplatform.com/s/%s/media/guid/%d/%s' % (provider_id, account_id, entry['guid']) if account_id else None
 339
 340         formats = []
 341         subtitles = {}
 342         first_video_id = None
 343         duration = None
 344         asset_types = []
 345         for item in entry['media$content']:
 346             smil_url = item['plfile$url']
 347             cur_video_id = ThePlatformIE._match_id(smil_url)
 348             if first_video_id is None:
 349                 first_video_id = cur_video_id
 350                 duration = float_or_none(item.get('plfile$duration'))
 351             for asset_type in item['plfile$assetTypes']:
 352                 if asset_type in asset_types:
 353                     continue
 354                 asset_types.append(asset_type)
 355                 query = {
 356                     'mbr': 'true',
 357                     'formats': item['plfile$format'],
 358                     'assetTypes': asset_type,
 359                 }
 360                 if asset_type in asset_types_query:
 361                     query.update(asset_types_query[asset_type])
 362                 cur_formats, cur_subtitles = self._extract_theplatform_smil(update_url_query(
 363                     main_smil_url or smil_url, query), video_id, 'Downloading SMIL data for %s' % asset_type)
 364                 formats.extend(cur_formats)
 365                 subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 366
 367         self._sort_formats(formats)
 368
 369         thumbnails = [{
 370             'url': thumbnail['plfile$url'],
 371             'width': int_or_none(thumbnail.get('plfile$width')),
 372             'height': int_or_none(thumbnail.get('plfile$height')),
 373         } for thumbnail in entry.get('media$thumbnails', [])]
 374
 375         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 376         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 377
 378         ret = self._extract_theplatform_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 379         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 380         ret.update({
 381             'id': video_id,
 382             'formats': formats,
 383             'subtitles': subtitles,
 384             'thumbnails': thumbnails,
 385             'duration': duration,
 386             'timestamp': timestamp,
 387             'categories': categories,
 388         })
 389         if custom_fields:
 390             ret.update(custom_fields(entry))
 391
 392         return ret
 393
 394     def _real_extract(self, url):
 395         mobj = re.match(self._VALID_URL, url)
 396
 397         video_id = mobj.group('id')
 398         provider_id = mobj.group('provider_id')
 399         feed_id = mobj.group('feed_id')
 400         filter_query = mobj.group('filter')
 401
 402         return self._extract_feed_info(provider_id, feed_id, filter_query, video_id)