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

   1 # coding: utf-8
   2 from __future__ import unicode_literals
   3
   4 import calendar
   5 import datetime
   6 import re
   7
   8 from .common import InfoExtractor
   9 from ..compat import (
  10     compat_etree_fromstring,
  11     compat_str,
  12     compat_parse_qs,
  13     compat_xml_parse_error,
  14 )
  15 from ..utils import (
  16     ExtractorError,
  17     int_or_none,
  18     float_or_none,
  19     xpath_text,
  20 )
  21
  22
  23 class BiliBiliIE(InfoExtractor):
  24     _VALID_URL = r'https?://www\.bilibili\.(?:tv|com)/video/av(?P<id>\d+)'
  25
  26     _TESTS = [{
  27         'url': 'http://www.bilibili.tv/video/av1074402/',
  28         'md5': '5f7d29e1a2872f3df0cf76b1f87d3788',
  29         'info_dict': {
  30             'id': '1554319',
  31             'ext': 'flv',
  32             'title': '【金坷垃】金泡沫',
  33             'description': 'md5:ce18c2a2d2193f0df2917d270f2e5923',
  34             'duration': 308.067,
  35             'timestamp': 1398012660,
  36             'upload_date': '20140420',
  37             'thumbnail': 're:^https?://.+\.jpg',
  38             'uploader': '菊子桑',
  39             'uploader_id': '156160',
  40         },
  41     }, {
  42         'url': 'http://www.bilibili.com/video/av1041170/',
  43         'info_dict': {
  44             'id': '1041170',
  45             'title': '【BD1080P】刀语【诸神&异域】',
  46             'description': '这是个神奇的故事~每个人不留弹幕不给走哦~切利哦！~',
  47         },
  48         'playlist_count': 9,
  49     }]
  50
  51     # BiliBili blocks keys from time to time. The current key is extracted from
  52     # the Android client
  53     # TODO: find the sign algorithm used in the flash player
  54     _APP_KEY = '86385cdc024c0f6c'
  55
  56     def _real_extract(self, url):
  57         mobj = re.match(self._VALID_URL, url)
  58         video_id = mobj.group('id')
  59
  60         webpage = self._download_webpage(url, video_id)
  61
  62         params = compat_parse_qs(self._search_regex(
  63             [r'EmbedPlayer\([^)]+,\s*"([^"]+)"\)',
  64              r'<iframe[^>]+src="https://secure\.bilibili\.com/secure,([^"]+)"'],
  65             webpage, 'player parameters'))
  66         cid = params['cid'][0]
  67
  68         info_xml_str = self._download_webpage(
  69             'http://interface.bilibili.com/v_cdn_play',
  70             cid, query={'appkey': self._APP_KEY, 'cid': cid},
  71             note='Downloading video info page')
  72
  73         err_msg = None
  74         durls = None
  75         info_xml = None
  76         try:
  77             info_xml = compat_etree_fromstring(info_xml_str.encode('utf-8'))
  78         except compat_xml_parse_error:
  79             info_json = self._parse_json(info_xml_str, video_id, fatal=False)
  80             err_msg = (info_json or {}).get('error_text')
  81         else:
  82             err_msg = xpath_text(info_xml, './message')
  83
  84         if info_xml is not None:
  85             durls = info_xml.findall('./durl')
  86         if not durls:
  87             if err_msg:
  88                 raise ExtractorError('%s said: %s' % (self.IE_NAME, err_msg), expected=True)
  89             else:
  90                 raise ExtractorError('No videos found!')
  91
  92         entries = []
  93
  94         for durl in durls:
  95             size = xpath_text(durl, ['./filesize', './size'])
  96             formats = [{
  97                 'url': durl.find('./url').text,
  98                 'filesize': int_or_none(size),
  99             }]
 100             for backup_url in durl.findall('./backup_url/url'):
 101                 formats.append({
 102                     'url': backup_url.text,
 103                     # backup URLs have lower priorities
 104                     'preference': -2 if 'hd.mp4' in backup_url.text else -3,
 105                 })
 106
 107             self._sort_formats(formats)
 108
 109             entries.append({
 110                 'id': '%s_part%s' % (cid, xpath_text(durl, './order')),
 111                 'duration': int_or_none(xpath_text(durl, './length'), 1000),
 112                 'formats': formats,
 113             })
 114
 115         title = self._html_search_regex('<h1[^>]+title="([^"]+)">', webpage, 'title')
 116         description = self._html_search_meta('description', webpage)
 117         datetime_str = self._html_search_regex(
 118             r'<time[^>]+datetime="([^"]+)"', webpage, 'upload time', fatal=False)
 119         if datetime_str:
 120             timestamp = calendar.timegm(datetime.datetime.strptime(datetime_str, '%Y-%m-%dT%H:%M').timetuple())
 121
 122         # TODO 'view_count' requires deobfuscating Javascript
 123         info = {
 124             'id': compat_str(cid),
 125             'title': title,
 126             'description': description,
 127             'timestamp': timestamp,
 128             'thumbnail': self._html_search_meta('thumbnailUrl', webpage),
 129             'duration': float_or_none(xpath_text(info_xml, './timelength'), scale=1000),
 130         }
 131
 132         uploader_mobj = re.search(
 133             r'<a[^>]+href="https?://space\.bilibili\.com/(?P<id>\d+)"[^>]+title="(?P<name>[^"]+)"',
 134             webpage)
 135         if uploader_mobj:
 136             info.update({
 137                 'uploader': uploader_mobj.group('name'),
 138                 'uploader_id': uploader_mobj.group('id'),
 139             })
 140
 141         for entry in entries:
 142             entry.update(info)
 143
 144         if len(entries) == 1:
 145             return entries[0]
 146         else:
 147             return {
 148                 '_type': 'multi_video',
 149                 'id': video_id,
 150                 'title': title,
 151                 'description': description,
 152                 'entries': entries,
 153             }