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_str,
  14 )
  15 from ..utils import (
  16     determine_ext,
  17     ExtractorError,
  18     xpath_with_ns,
  19     unsmuggle_url,
  20 )
  21
  22 _x = lambda p: xpath_with_ns(p, {'smil': 'http://www.w3.org/2005/SMIL21/Language'})
  23
  24
  25 class ThePlatformIE(InfoExtractor):
  26     _VALID_URL = r'''(?x)
  27         (?:https?://(?:link|player)\.theplatform\.com/[sp]/(?P<provider_id>[^/]+)/
  28            (?P<config>(?:[^/\?]+/(?:swf|config)|onsite)/select/)?
  29          |theplatform:)(?P<id>[^/\?&]+)'''
  30
  31     _TESTS = [{
  32         # from http://www.metacafe.com/watch/cb-e9I_cZgTgIPd/blackberrys_big_bold_z30/
  33         'url': 'http://link.theplatform.com/s/dJ5BDC/e9I_cZgTgIPd/meta.smil?format=smil&Tracking=true&mbr=true',
  34         'info_dict': {
  35             'id': 'e9I_cZgTgIPd',
  36             'ext': 'flv',
  37             'title': 'Blackberry\'s big, bold Z30',
  38             'description': 'The Z30 is Blackberry\'s biggest, baddest mobile messaging device yet.',
  39             'duration': 247,
  40         },
  41         'params': {
  42             # rtmp download
  43             'skip_download': True,
  44         },
  45         # from http://www.cnet.com/videos/tesla-model-s-a-second-step-towards-a-cleaner-motoring-future/
  46     }, {
  47         'url': 'http://link.theplatform.com/s/kYEXFC/22d_qsQ6MIRT',
  48         'info_dict': {
  49             'id': '22d_qsQ6MIRT',
  50             'ext': 'flv',
  51             'description': 'md5:ac330c9258c04f9d7512cf26b9595409',
  52             'title': 'Tesla Model S: A second step towards a cleaner motoring future',
  53         },
  54         'params': {
  55             # rtmp download
  56             'skip_download': True,
  57         }
  58     }]
  59
  60     @staticmethod
  61     def _sign_url(url, sig_key, sig_secret, life=600, include_qs=False):
  62         flags = '10' if include_qs else '00'
  63         expiration_date = '%x' % (int(time.time()) + life)
  64
  65         def str_to_hex(str):
  66             return binascii.b2a_hex(str.encode('ascii')).decode('ascii')
  67
  68         def hex_to_str(hex):
  69             return binascii.a2b_hex(hex)
  70
  71         relative_path = url.split('http://link.theplatform.com/s/')[1].split('?')[0]
  72         clear_text = hex_to_str(flags + expiration_date + str_to_hex(relative_path))
  73         checksum = hmac.new(sig_key.encode('ascii'), clear_text, hashlib.sha1).hexdigest()
  74         sig = flags + expiration_date + checksum + str_to_hex(sig_secret)
  75         return '%s&sig=%s' % (url, sig)
  76
  77     def _real_extract(self, url):
  78         url, smuggled_data = unsmuggle_url(url, {})
  79
  80         mobj = re.match(self._VALID_URL, url)
  81         provider_id = mobj.group('provider_id')
  82         video_id = mobj.group('id')
  83
  84         if not provider_id:
  85             provider_id = 'dJ5BDC'
  86
  87         if smuggled_data.get('force_smil_url', False):
  88             smil_url = url
  89         elif mobj.group('config'):
  90             config_url = url + '&form=json'
  91             config_url = config_url.replace('swf/', 'config/')
  92             config_url = config_url.replace('onsite/', 'onsite/config/')
  93             config = self._download_json(config_url, video_id, 'Downloading config')
  94             smil_url = config['releaseUrl'] + '&format=SMIL&formats=MPEG4&manifest=f4m'
  95         else:
  96             smil_url = ('http://link.theplatform.com/s/{0}/{1}/meta.smil?'
  97                         'format=smil&mbr=true'.format(provider_id, video_id))
  98
  99         sig = smuggled_data.get('sig')
 100         if sig:
 101             smil_url = self._sign_url(smil_url, sig['key'], sig['secret'])
 102
 103         meta = self._download_xml(smil_url, video_id)
 104         try:
 105             error_msg = next(
 106                 n.attrib['abstract']
 107                 for n in meta.findall(_x('.//smil:ref'))
 108                 if n.attrib.get('title') == 'Geographic Restriction' or n.attrib.get('title') == 'Expired')
 109         except StopIteration:
 110             pass
 111         else:
 112             raise ExtractorError(error_msg, expected=True)
 113
 114         info_url = 'http://link.theplatform.com/s/{0}/{1}?format=preview'.format(provider_id, video_id)
 115         info_json = self._download_webpage(info_url, video_id)
 116         info = json.loads(info_json)
 117
 118         subtitles = {}
 119         captions = info.get('captions')
 120         if isinstance(captions, list):
 121             for caption in captions:
 122                 lang, src, mime = caption.get('lang', 'en'), caption.get('src'), caption.get('type')
 123                 subtitles[lang] = [{
 124                     'ext': 'srt' if mime == 'text/srt' else 'ttml',
 125                     'url': src,
 126                 }]
 127
 128         head = meta.find(_x('smil:head'))
 129         body = meta.find(_x('smil:body'))
 130
 131         f4m_node = body.find(_x('smil:seq//smil:video'))
 132         if f4m_node is not None and '.f4m' in f4m_node.attrib['src']:
 133             f4m_url = f4m_node.attrib['src']
 134             if 'manifest.f4m?' not in f4m_url:
 135                 f4m_url += '?'
 136             # the parameters are from syfy.com, other sites may use others,
 137             # they also work for nbc.com
 138             f4m_url += '&g=UXWGVKRWHFSP&hdcore=3.0.3'
 139             formats = self._extract_f4m_formats(f4m_url, video_id)
 140         else:
 141             formats = []
 142             switch = body.find(_x('smil:switch'))
 143             if switch is None:
 144                 switch = body.find(_x('smil:par//smil:switch'))
 145             if switch is None:
 146                 switch = body.find(_x('smil:par'))
 147             if switch is not None:
 148                 base_url = head.find(_x('smil:meta')).attrib['base']
 149                 for f in switch.findall(_x('smil:video')):
 150                     attr = f.attrib
 151                     width = int(attr['width'])
 152                     height = int(attr['height'])
 153                     vbr = int(attr['system-bitrate']) // 1000
 154                     format_id = '%dx%d_%dk' % (width, height, vbr)
 155                     formats.append({
 156                         'format_id': format_id,
 157                         'url': base_url,
 158                         'play_path': 'mp4:' + attr['src'],
 159                         'ext': 'flv',
 160                         'width': width,
 161                         'height': height,
 162                         'vbr': vbr,
 163                     })
 164             else:
 165                 switch = body.find(_x('smil:seq//smil:switch'))
 166                 for f in switch.findall(_x('smil:video')):
 167                     attr = f.attrib
 168                     vbr = int(attr['system-bitrate']) // 1000
 169                     ext = determine_ext(attr['src'])
 170                     if ext == 'once':
 171                         ext = 'mp4'
 172                     formats.append({
 173                         'format_id': compat_str(vbr),
 174                         'url': attr['src'],
 175                         'vbr': vbr,
 176                         'ext': ext,
 177                     })
 178             self._sort_formats(formats)
 179
 180         return {
 181             'id': video_id,
 182             'title': info['title'],
 183             'subtitles': subtitles,
 184             'formats': formats,
 185             'description': info['description'],
 186             'thumbnail': info['defaultThumbnailUrl'],
 187             'duration': info['duration'] // 1000,
 188         }