_ 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 ..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_formats(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         return formats
  58
  59     def get_metadata(self, path, video_id):
  60         info_url = 'http://link.theplatform.com/s/%s?format=preview' % path
  61         info_json = self._download_webpage(info_url, video_id)
  62         info = json.loads(info_json)
  63
  64         subtitles = {}
  65         captions = info.get('captions')
  66         if isinstance(captions, list):
  67             for caption in captions:
  68                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
  69                 subtitles[lang] = [{
  70                     'ext': 'srt' if mime == 'text/srt' else 'ttml',
  71                     'url': src,
  72                 }]
  73
  74         return {
  75             'title': info['title'],
  76             'subtitles': subtitles,
  77             'description': info['description'],
  78             'thumbnail': info['defaultThumbnailUrl'],
  79             'duration': int_or_none(info.get('duration'), 1000),
  80         }
  81
  82
  83 class ThePlatformIE(ThePlatformBaseIE):
  84     _VALID_URL = r'''(?x)
  85         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
  86            (?:(?P<media>(?:[^/]+/)+select/media/)|(?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/))?
  87          |theplatform:)(?P<id>[^/\?&]+)'''
  88
  89     _TESTS = [{
  90         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
  91         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
  92         'info_dict': {
  93             'id': 'e9I_cZgTgIPd',
  94             'ext': 'flv',
  95             'title': 'Blackberry\'s big, bold Z30',
  96             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
  97             'duration': 247,
  98         },
  99         'params': {
 100             # rtmp download
 101             'skip_download': True,
 102         },
 103     }, {
 104         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
 105         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
 106         'info_dict': {
 107             'id': '22d_qsQ6MIRT',
 108             'ext': 'flv',
 109             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
 110             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
 111         },
 112         'params': {
 113             # rtmp download
 114             'skip_download': True,
 115         }
 116     }, {
 117         'url': 'https://player.theplatform.com/p/D6x-PC/pulse_preview/embed/select/media/yMBg9E8KFxZD',
 118         'info_dict': {
 119             'id': 'yMBg9E8KFxZD',
 120             'ext': 'mp4',
 121             'description': 'md5:644ad9188d655b742f942bf2e06b002d',
 122             'title': 'HIGHLIGHTS: USA bag first ever series Cup win',
 123         }
 124     }, {
 125         'url': 'http://player.theplatform.com/p/NnzsPC/widget/select/media/4Y0TlYUr_ZT7',
 126         'only_matching': True,
 127     }, {
 128         'url': 'http://player.theplatform.com/p/2E2eJC/nbcNewsOffsite?guid=tdy_or_siri_150701',
 129         'md5': '734f3790fb5fc4903da391beeebc4836',
 130         'info_dict': {
 131             'id': 'tdy_or_siri_150701',
 132             'ext': 'mp4',
 133             'title': 'iPhone Siri’s sassy response to a math question has people talking',
 134             'description': 'md5:a565d1deadd5086f3331d57298ec6333',
 135             'duration': 83.0,
 136             'thumbnail': 're:^https?://.*\.jpg$',
 137             'timestamp': 1435752600,
 138             'upload_date': '20150701',
 139             'categories': ['Today/Shows/Orange Room', 'Today/Sections/Money', 'Today/Topics/Tech', "Today/Topics/Editor's picks"],
 140         },
 141     }]
 142
 143     @staticmethod
 144     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
 145         flags = '10' if include_qs else '00'
 146         expiration_date = '%x' % (int(time.time()) + life)
 147
 148         def str_to_hex(str):
 149             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
 150
 151         def hex_to_str(hex):
 152             return binascii.a2b_hex(hex)
 153
 154         relative_path = url.split('http://link.theplatform.com/s/')[1].split('?')[0]
 155         clear_text = hex_to_str(flags + expiration_date + str_to_hex(relative_path))
 156         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
 157         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
 158         return '%s&sig=%s' % (url, sig)
 159
 160     def _real_extract(self, url):
 161         url, smuggled_data = unsmuggle_url(url, {})
 162
 163         mobj = re.match(self._VALID_URL, url)
 164         provider_id = mobj.group('provider_id')
 165         video_id = mobj.group('id')
 166
 167         if not provider_id:
 168             provider_id = 'dJ5BDC'
 169
 170         path = provider_id
 171         if mobj.group('media'):
 172             path += '/media'
 173         path += '/' + video_id
 174
 175         qs_dict = compat_parse_qs(compat_urllib_parse_urlparse(url).query)
 176         if 'guid' in qs_dict:
 177             webpage = self._download_webpage(url, video_id)
 178             scripts = re.findall(r'<script[^>]+src="([^"]+)"', webpage)
 179             feed_id = None
 180             # feed id usually locates in the last script.
 181             # Seems there's no pattern for the interested script filename, so
 182             # I try one by one
 183             for script in reversed(scripts):
 184                 feed_script = self._download_webpage(script, video_id, 'Downloading feed script')
 185                 feed_id = self._search_regex(r'defaultFeedId\s*:\s*"([^"]+)"', feed_script, 'default feed id', default=None)
 186                 if feed_id is not None:
 187                     break
 188             if feed_id is None:
 189                 raise ExtractorError('Unable to find feed id')
 190             return self.url_result('http://feed.theplatform.com/f/%s/%s?byGuid=%s' % (
 191                 provider_id, feed_id, qs_dict['guid'][0]))
 192
 193         if smuggled_data.get('force_smil_url', False):
 194             smil_url = url
 195         elif mobj.group('config'):
 196             config_url = url + '&form=json'
 197             config_url = config_url.replace('swf/', 'config/')
 198             config_url = config_url.replace('onsite/', 'onsite/config/')
 199             config = self._download_json(config_url, video_id, 'Downloading config')
 200             if 'releaseUrl' in config:
 201                 release_url = config['releaseUrl']
 202             else:
 203                 release_url = 'http://link.theplatform.com/s/%s?mbr=true' % path
 204             smil_url = release_url + '&format=SMIL&formats=MPEG4&manifest=f4m'
 205         else:
 206             smil_url = 'http://link.theplatform.com/s/%s/meta.smil?format=smil&mbr=true' % path
 207
 208         sig = smuggled_data.get('sig')
 209         if sig:
 210             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 211
 212         formats = self._extract_theplatform_smil_formats(smil_url, video_id)
 213
 214         ret = self.get_metadata(path, video_id)
 215         ret.update({
 216             'id': video_id,
 217             'formats': formats,
 218         })
 219
 220         return ret
 221
 222
 223 class ThePlatformFeedIE(ThePlatformBaseIE):
 224     _URL_TEMPLATE = '%s//feed.theplatform.com/f/%s/%s?form=json&byGuid=%s'
 225     _VALID_URL = r'https?://feed\.theplatform\.com/f/(?P<provider_id>[^/]+)/(?P<feed_id>[^?/]+)\?(?:[^&]+&)*byGuid=(?P<id>[a-zA-Z0-9_]+)'
 226     _TEST = {
 227         # From http://player.theplatform.com/p/7wvmTC/MSNBCEmbeddedOffSite?guid=n_hardball_5biden_140207
 228         'url': 'http://feed.theplatform.com/f/7wvmTC/msnbc_video-p-test?form=json&pretty=true&range=-40&byGuid=n_hardball_5biden_140207',
 229         'md5': '22d2b84f058d3586efcd99e57d59d314',
 230         'info_dict': {
 231             'id': 'n_hardball_5biden_140207',
 232             'ext': 'mp4',
 233             'title': 'The Biden factor: will Joe run in 2016?',
 234             'description': 'Could Vice President Joe Biden be preparing a 2016 campaign? Mark Halperin and Sam Stein weigh in.',
 235             'thumbnail': 're:^https?://.*\.jpg$',
 236             'upload_date': '20140208',
 237             'timestamp': 1391824260,
 238             'duration': 467.0,
 239             'categories': ['MSNBC/Issues/Democrats', 'MSNBC/Issues/Elections/Election 2016'],
 240         },
 241     }
 242
 243     def _real_extract(self, url):
 244         mobj = re.match(self._VALID_URL, url)
 245
 246         video_id = mobj.group('id')
 247         provider_id = mobj.group('provider_id')
 248         feed_id = mobj.group('feed_id')
 249
 250         real_url = self._URL_TEMPLATE % (self.http_scheme(), provider_id, feed_id, video_id)
 251         feed = self._download_json(real_url, video_id)
 252         entry = feed['entries'][0]
 253
 254         formats = []
 255         first_video_id = None
 256         duration = None
 257         for item in entry['media$content']:
 258             smil_url = item['plfile$url'] + '&format=SMIL&Tracking=true&Embedded=true&formats=MPEG4,F4M'
 259             cur_video_id = url_basename(smil_url)
 260             if first_video_id is None:
 261                 first_video_id = cur_video_id
 262                 duration = float_or_none(item.get('plfile$duration'))
 263             formats.extend(self._extract_theplatform_smil_formats(smil_url, video_id, 'Downloading SMIL data for %s' % cur_video_id))
 264
 265         self._sort_formats(formats)
 266
 267         thumbnails = [{
 268             'url': thumbnail['plfile$url'],
 269             'width': int_or_none(thumbnail.get('plfile$width')),
 270             'height': int_or_none(thumbnail.get('plfile$height')),
 271         } for thumbnail in entry.get('media$thumbnails', [])]
 272
 273         timestamp = int_or_none(entry.get('media$availableDate'), scale=1000)
 274         categories = [item['media$name'] for item in entry.get('media$categories', [])]
 275
 276         ret = self.get_metadata('%s/%s' % (provider_id, first_video_id), video_id)
 277         ret.update({
 278             'id': video_id,
 279             'formats': formats,
 280             'thumbnails': thumbnails,
 281             'duration': duration,
 282             'timestamp': timestamp,
 283             'categories': categories,
 284         })
 285
 286         return ret