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