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 .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     xpath_with_ns,
  20     unsmuggle_url,
  21     int_or_none,
  22     url_basename,
  23     float_or_none,
  24 )
  25
  26 default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  27 _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  28
  29
  30 class ThePlatformBaseIE(InfoExtractor):
  31     def _extract_theplatform_smil(self, smil_url, video_id, note='Downloading SMIL data'):
  32         meta = self._download_xml(smil_url, video_id, note=note)
  33         try:
  34             error_msg = next(
  35                 n.attrib['abstract']
  36                 for n in meta.findall(_x('.//smil:ref'))
  37                 if n.attrib.get('title') == 'Geographic Restriction' or n.attrib.get('title') == 'Expired')
  38         except StopIteration:
  39             pass
  40         else:
  41             raise ExtractorError(error_msg, expected=True)
  42
  43         formats = self._parse_smil_formats(
  44             meta, smil_url, video_id, namespace=default_ns,
  45             # the parameters are from syfy.com, other sites may use others,
  46             # they also work for nbc.com
  47             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  48             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  49
  50         for _format in formats:
  51             ext = determine_ext(_format['url'])
  52             if ext == 'once':
  53                 _format['ext'] = 'mp4'
  54
  55         self._sort_formats(formats)
  56
  57         subtitles = self._parse_smil_subtitles(meta, default_ns)
  58
  59         return formats, subtitles
  60
  61     def get_metadata(self, path, video_id):
  62         info_url = 'http://link.theplatform.com/s/%s?format=preview' % path
  63         info = self._download_json(info_url, video_id)
  64
  65         subtitles = {}
  66         captions = info.get('captions')
  67         if isinstance(captions, list):
  68             for caption in captions:
  69                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  70                 subtitles[lang] = [{
  71                     'ext': 'srt' if mime == 'text/srt' else 'ttml',
  72                     'url': src,
  73                 }]
  74
  75         return {
  76             'title': info['title'],
  77             'subtitles': subtitles,
  78             'description': info['description'],
  79             'thumbnail': info['defaultThumbnailUrl'],
  80             'duration': int_or_none(info.get('duration'), 1000),
  81         }
  82
  83
  84 class ThePlatformIE(ThePlatformBaseIE):
  85     _VALID_URL = r'''(?x)
  86         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
  87            (?:(?P<media>(?:[^/]+/)+select/media/)|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
  88          |theplatform:)(?P<id>[^/\?&]+)'''
  89
  90     _TESTS = [{
  91         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
  92         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
  93         'info_dict': {
  94             'id': 'e9I_cZgTgIPd',
  95             'ext': 'flv',
  96             'title': 'Blackberry\'s big, bold Z30',
  97             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
  98             'duration': 247,
  99         },
 100         'params': {
 101             # rtmp download
 102             'skip_download': True,
 103         },
 104     }, {
 105         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 106         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 107         'info_dict': {
 108             'id': '22d_qsQ6MIRT',
 109             'ext': 'flv',
 110             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 111             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 112         },
 113         'params': {
 114             # rtmp download
 115             'skip_download': True,
 116         }
 117     }, {
 118         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 119         'info_dict': {
 120             'id': 'yMBg9E8KFxZD',
 121             'ext': 'mp4',
 122             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 123             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 124         }
 125     }, {
 126         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 127         'only_matching': True,
 128     }, {
 129         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 130         'md5': '734f3790fb5fc4903da391beeebc4836',
 131         'info_dict': {
 132             'id': 'tdy_or_siri_150701',
 133             'ext': 'mp4',
 134             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 135             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 136             'duration': 83.0,
 137             'thumbnail': 're:^https?://.*\.jpg$',
 138             'timestamp': 1435752600,
 139             'upload_date': '20150701',
 140             'categories': ['Today/Shows/Orange Room', 'Today/Sections/Money', 'Today/Topics/Tech', "Today/Topics/Editor's picks"],
 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             webpage = self._download_webpage(url, video_id)
 208             smil_url = self._search_regex(
 209                 r'<link[^>]+href=(["\'])(?P<url>.+?)\1[^>]+type=["\']application/smil\+xml',
 210                 webpage, 'smil url', group='url')
 211             path = self._search_regex(
 212                 r'link\.theplatform\.com/s/((?:[^/?#&]+/)+[^/?#&]+)', smil_url, 'path')
 213             smil_url += '?' if '?' not in smil_url else '&' + 'formats=m3u,mpeg4&format=SMIL'
 214         elif mobj.group('config'):
 215             config_url = url + '&form=json'
 216             config_url = config_url.replace('swf/', 'config/')
 217             config_url = config_url.replace('onsite/', 'onsite/config/')
 218             config = self._download_json(config_url, video_id, 'Downloading config')
 219             if 'releaseUrl' in config:
 220                 release_url = config['releaseUrl']
 221             else:
 222                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 223             smil_url = release_url + '&format=SMIL&formats=MPEG4&manifest=f4m'
 224         else:
 225             smil_url = 'http://link.theplatform.com/s/%s/meta.smil?format=smil&mbr=true' % path
 226
 227         sig = smuggled_data.get('sig')
 228         if sig:
 229             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 230
 231         formats, subtitles = self._extract_theplatform_smil(smil_url, video_id)
 232
 233         ret = self.get_metadata(path, video_id)
 234         combined_subtitles = self._merge_subtitles(ret.get('subtitles', {}), subtitles)
 235         ret.update({
 236             'id': video_id,
 237             'formats': formats,
 238             'subtitles': combined_subtitles,
 239         })
 240
 241         return ret
 242
 243
 244 class ThePlatformFeedIE(ThePlatformBaseIE):
 245     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&byGuid=%s'
 246     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*byGuid=(?P<id>[a-zA-Z0-9_]+)'
 247     _TEST = {
 248         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 249         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 250         'md5': '22d2b84f058d3586efcd99e57d59d314',
 251         'info_dict': {
 252             'id': 'n_hardball_5biden_140207',
 253             'ext': 'mp4',
 254             'title': 'The Biden factor: will Joe run in 2016?',
 255             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 256             'thumbnail': 're:^https?://.*\.jpg$',
 257             'upload_date': '20140208',
 258             'timestamp': 1391824260,
 259             'duration': 467.0,
 260             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 261         },
 262     }
 263
 264     def _real_extract(self, url):
 265         mobj = re.match(self._VALID_URL, url)
 266
 267         video_id = mobj.group('id')
 268         provider_id = mobj.group('provider_id')
 269         feed_id = mobj.group('feed_id')
 270
 271         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, video_id)
 272         feed = self._download_json(real_url, video_id)
 273         entry = feed['entries'][0]
 274
 275         formats = []
 276         subtitles = {}
 277         first_video_id = None
 278         duration = None
 279         for item in entry['media$content']:
 280             smil_url = item['plfile$url'] + '&format=SMIL&Tracking=true&Embedded=true&formats=MPEG4,F4M'
 281             cur_video_id = url_basename(smil_url)
 282             if first_video_id is None:
 283                 first_video_id = cur_video_id
 284                 duration = float_or_none(item.get('plfile$duration'))
 285             cur_formats, cur_subtitles = self._extract_theplatform_smil(smil_url, video_id, 'Downloading SMIL data for %s' % cur_video_id)
 286             formats.extend(cur_formats)
 287             subtitles = self._merge_subtitles(subtitles, cur_subtitles)
 288
 289         self._sort_formats(formats)
 290
 291         thumbnails = [{
 292             'url': thumbnail['plfile$url'],
 293             'width': int_or_none(thumbnail.get('plfile$width')),
 294             'height': int_or_none(thumbnail.get('plfile$height')),
 295         } for thumbnail in entry.get('media$thumbnails', [])]
 296
 297         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 298         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 299
 300         ret = self.get_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 301         subtitles = self._merge_subtitles(subtitles, ret['subtitles'])
 302         ret.update({
 303             'id': video_id,
 304             'formats': formats,
 305             'subtitles': subtitles,
 306             'thumbnails': thumbnails,
 307             'duration': duration,
 308             'timestamp': timestamp,
 309             'categories': categories,
 310         })
 311
 312         return ret