_ Git - youtube-dl/blob - youtube_dl/extractor/theplatform.py

   1 from __future__ import unicode_literals
   2
   3 import re
   4 import json
   5 import time
   6 import hmac
   7 import binascii
   8 import hashlib
   9
  10
  11 from .common import InfoExtractor
  12 from ..utils import (
  13     determine_ext,
  14     ExtractorError,
  15     xpath_with_ns,
  16     unsmuggle_url,
  17     int_or_none,
  18     url_basename,
  19     float_or_none,
  20 )
  21
  22 default_ns = 'http://www.w3.org/2005/SMIL21/Language'
  23 _x = lambda p: xpath_with_ns(p, {'smil': default_ns})
  24
  25
  26 class ThePlatformBaseIE(InfoExtractor):
  27     def _extract_theplatform_smil_formats(self, smil_url, video_id, note='Downloading SMIL data'):
  28         meta = self._download_xml(smil_url, video_id, note=note)
  29         try:
  30             error_msg = next(
  31                 n.attrib['abstract']
  32                 for n in meta.findall(_x('.//smil:ref'))
  33                 if n.attrib.get('title') == 'Geographic Restriction' or n.attrib.get('title') == 'Expired')
  34         except StopIteration:
  35             pass
  36         else:
  37             raise ExtractorError(error_msg, expected=True)
  38
  39         formats = self._parse_smil_formats(
  40             meta, smil_url, video_id, namespace=default_ns,
  41             # the parameters are from syfy.com, other sites may use others,
  42             # they also work for nbc.com
  43             f4m_params={'g': 'UXWGVKRWHFSP', 'hdcore': '3.0.3'},
  44             transform_rtmp_url=lambda streamer, src: (streamer, 'mp4:' + src))
  45
  46         for _format in formats:
  47             ext = determine_ext(_format['url'])
  48             if ext == 'once':
  49                 _format['ext'] = 'mp4'
  50
  51         self._sort_formats(formats)
  52
  53         return formats
  54
  55     def get_metadata(self, path, video_id):
  56         info_url = 'http://link.theplatform.com/s/%s?format=preview' % path
  57         info_json = self._download_webpage(info_url, video_id)
  58         info = json.loads(info_json)
  59
  60         subtitles = {}
  61         captions = info.get('captions')
  62         if isinstance(captions, list):
  63             for caption in captions:
  64                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  65                 subtitles[lang] = [{
  66                     'ext': 'srt' if mime == 'text/srt' else 'ttml',
  67                     'url': src,
  68                 }]
  69
  70         return {
  71             'title': info['title'],
  72             'subtitles': subtitles,
  73             'description': info['description'],
  74             'thumbnail': info['defaultThumbnailUrl'],
  75             'duration': int_or_none(info.get('duration'), 1000),
  76         }
  77
  78
  79 class ThePlatformIE(ThePlatformBaseIE):
  80     _VALID_URL = r'''(?x)
  81         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
  82            (?:(?P<media>(?:[^/]+/)+select/media/)|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
  83          |theplatform:)(?P<id>[^/\?&]+)'''
  84
  85     _TESTS = [{
  86         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
  87         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
  88         'info_dict': {
  89             'id': 'e9I_cZgTgIPd',
  90             'ext': 'flv',
  91             'title': 'Blackberry\'s big, bold Z30',
  92             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
  93             'duration': 247,
  94         },
  95         'params': {
  96             # rtmp download
  97             'skip_download': True,
  98         },
  99     }, {
 100         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 101         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 102         'info_dict': {
 103             'id': '22d_qsQ6MIRT',
 104             'ext': 'flv',
 105             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 106             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 107         },
 108         'params': {
 109             # rtmp download
 110             'skip_download': True,
 111         }
 112     }, {
 113         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 114         'info_dict': {
 115             'id': 'yMBg9E8KFxZD',
 116             'ext': 'mp4',
 117             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 118             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 119         }
 120     }, {
 121         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 122         'only_matching': True,
 123     }]
 124
 125     @staticmethod
 126     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 127         flags = '10' if include_qs else '00'
 128         expiration_date = '%x' % (int(time.time()) + life)
 129
 130         def str_to_hex(str):
 131             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 132
 133         def hex_to_str(hex):
 134             return binascii.a2b_hex(hex)
 135
 136         relative_path = url.split('http://link.theplatform.com/s/')[1].split('?')[0]
 137         clear_text = hex_to_str(flags + expiration_date + str_to_hex(relative_path))
 138         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 139         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 140         return '%s&sig=%s' % (url, sig)
 141
 142     def _real_extract(self, url):
 143         url, smuggled_data = unsmuggle_url(url, {})
 144
 145         mobj = re.match(self._VALID_URL, url)
 146         provider_id = mobj.group('provider_id')
 147         video_id = mobj.group('id')
 148
 149         if not provider_id:
 150             provider_id = 'dJ5BDC'
 151
 152         path = provider_id
 153         if mobj.group('media'):
 154             path += '/media'
 155         path += '/' + video_id
 156
 157         if smuggled_data.get('force_smil_url', False):
 158             smil_url = url
 159         elif mobj.group('config'):
 160             config_url = url + '&form=json'
 161             config_url = config_url.replace('swf/', 'config/')
 162             config_url = config_url.replace('onsite/', 'onsite/config/')
 163             config = self._download_json(config_url, video_id, 'Downloading config')
 164             if 'releaseUrl' in config:
 165                 release_url = config['releaseUrl']
 166             else:
 167                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 168             smil_url = release_url + '&format=SMIL&formats=MPEG4&manifest=f4m'
 169         else:
 170             smil_url = 'http://link.theplatform.com/s/%s/meta.smil?format=smil&mbr=true' % path
 171
 172         sig = smuggled_data.get('sig')
 173         if sig:
 174             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 175
 176         formats = self._extract_theplatform_smil_formats(smil_url, video_id)
 177
 178         ret = self.get_metadata(path, video_id)
 179         ret.update({
 180             'id': video_id,
 181             'formats': formats,
 182         })
 183
 184         return ret
 185
 186
 187 class ThePlatformFeedIE(ThePlatformBaseIE):
 188     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&byGuid=%s'
 189     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*byGuid=(?P<id>[a-zA-Z0-9_]+)'
 190     _TEST = {
 191         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 192         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 193         'md5': '22d2b84f058d3586efcd99e57d59d314',
 194         'info_dict': {
 195             'id': 'n_hardball_5biden_140207',
 196             'ext': 'mp4',
 197             'title': 'The Biden factor: will Joe run in 2016?',
 198             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 199             'thumbnail': 're:^https?://.*\.jpg$',
 200             'upload_date': '20140208',
 201             'timestamp': 1391824260,
 202             'duration': 467.0,
 203             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 204         },
 205     }
 206
 207     def _real_extract(self, url):
 208         mobj = re.match(self._VALID_URL, url)
 209
 210         video_id = mobj.group('id')
 211         provider_id = mobj.group('provider_id')
 212         feed_id = mobj.group('feed_id')
 213
 214         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, video_id)
 215         feed = self._download_json(real_url, video_id)
 216         entry = feed['entries'][0]
 217
 218         formats = []
 219         first_video_id = None
 220         duration = None
 221         for item in entry['media$content']:
 222             smil_url = item['plfile$url'] + '&format=SMIL&Tracking=true&Embedded=true&formats=MPEG4,F4M'
 223             cur_video_id = url_basename(smil_url)
 224             if first_video_id is None:
 225                 first_video_id = cur_video_id
 226                 duration = float_or_none(item.get('plfile$duration'))
 227             formats.extend(self._extract_theplatform_smil_formats(smil_url, video_id, 'Downloading SMIL data for %s' % cur_video_id))
 228
 229         self._sort_formats(formats)
 230
 231         thumbnails = [{
 232             'url': thumbnail['plfile$url'],
 233             'width': int_or_none(thumbnail.get('plfile$width')),
 234             'height': int_or_none(thumbnail.get('plfile$height')),
 235         } for thumbnail in entry.get('media$thumbnails', [])]
 236
 237         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 238         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 239
 240         ret = self.get_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 241         ret.update({
 242             'id': video_id,
 243             'formats': formats,
 244             'thumbnails': thumbnails,
 245             'duration': duration,
 246             'timestamp': timestamp,
 247             'categories': categories,
 248         })
 249
 250         return ret