_ 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     _WORKING = False
  25
  26     _VALID_URL = r'https?://www\.bilibili\.(?:tv|com)/video/av(?P<id>\d+)'
  27
  28     _TESTS = [{
  29         'url': 'http://www.bilibili.tv/video/av1074402/',
  30         'md5': '9fa226fe2b8a9a4d5a69b4c6a183417e',
  31         'info_dict': {
  32             'id': '1554319',
  33             'ext': 'mp4',
  34             'title': '【金坷垃】金泡沫',
  35             'description': 'md5:ce18c2a2d2193f0df2917d270f2e5923',
  36             'duration': 308.315,
  37             'timestamp': 1398012660,
  38             'upload_date': '20140420',
  39             'thumbnail': 're:^https?://.+\.jpg',
  40             'uploader': '菊子桑',
  41             'uploader_id': '156160',
  42         },
  43     }, {
  44         'url': 'http://www.bilibili.com/video/av1041170/',
  45         'info_dict': {
  46             'id': '1507019',
  47             'ext': 'mp4',
  48             'title': '【BD1080P】刀语【诸神&异域】',
  49             'description': '这是个神奇的故事~每个人不留弹幕不给走哦~切利哦！~',
  50             'timestamp': 1396530060,
  51             'upload_date': '20140403',
  52             'uploader': '枫叶逝去',
  53             'uploader_id': '520116',
  54         },
  55     }, {
  56         'url': 'http://www.bilibili.com/video/av4808130/',
  57         'info_dict': {
  58             'id': '7802182',
  59             'ext': 'mp4',
  60             'title': '【长篇】哆啦A梦443【钉铛】',
  61             'description': '(2016.05.27)来组合客人的脸吧&amp;amp;寻母六千里锭 抱歉，又轮到周日上班现在才到家 封面www.pixiv.net/member_illust.php?mode=medium&amp;amp;illust_id=56912929',
  62             'timestamp': 1464564180,
  63             'upload_date': '20160529',
  64             'uploader': '喜欢拉面',
  65             'uploader_id': '151066',
  66         },
  67     }, {
  68         # Missing upload time
  69         'url': 'http://www.bilibili.com/video/av1867637/',
  70         'info_dict': {
  71             'id': '2880301',
  72             'ext': 'mp4',
  73             'title': '【HDTV】【喜剧】岳父岳母真难当 （2014）【法国票房冠军】',
  74             'description': '一个信奉天主教的法国旧式传统资产阶级家庭中有四个女儿。三个女儿却分别找了阿拉伯、犹太、中国丈夫，老夫老妻唯独期盼剩下未嫁的小女儿能找一个信奉天主教的法国白人，结果没想到小女儿找了一位非裔黑人……【这次应该不会跳帧了】',
  75             'uploader': '黑夜为猫',
  76             'uploader_id': '610729',
  77         },
  78         'params': {
  79             # Just to test metadata extraction
  80             'skip_download': True,
  81         },
  82         'expected_warnings': ['upload time'],
  83     }]
  84
  85     # BiliBili blocks keys from time to time. The current key is extracted from
  86     # the Android client
  87     # TODO: find the sign algorithm used in the flash player
  88     _APP_KEY = '86385cdc024c0f6c'
  89
  90     def _real_extract(self, url):
  91         mobj = re.match(self._VALID_URL, url)
  92         video_id = mobj.group('id')
  93
  94         webpage = self._download_webpage(url, video_id)
  95
  96         params = compat_parse_qs(self._search_regex(
  97             [r'EmbedPlayer\([^)]+,\s*"([^"]+)"\)',
  98              r'<iframe[^>]+src="https://secure\.bilibili\.com/secure,([^"]+)"'],
  99             webpage, 'player parameters'))
 100         cid = params['cid'][0]
 101
 102         info_xml_str = self._download_webpage(
 103             'http://interface.bilibili.com/v_cdn_play',
 104             cid, query={'appkey': self._APP_KEY, 'cid': cid},
 105             note='Downloading video info page')
 106
 107         err_msg = None
 108         durls = None
 109         info_xml = None
 110         try:
 111             info_xml = compat_etree_fromstring(info_xml_str.encode('utf-8'))
 112         except compat_xml_parse_error:
 113             info_json = self._parse_json(info_xml_str, video_id, fatal=False)
 114             err_msg = (info_json or {}).get('error_text')
 115         else:
 116             err_msg = xpath_text(info_xml, './message')
 117
 118         if info_xml is not None:
 119             durls = info_xml.findall('./durl')
 120         if not durls:
 121             if err_msg:
 122                 raise ExtractorError('%s said: %s' % (self.IE_NAME, err_msg), expected=True)
 123             else:
 124                 raise ExtractorError('No videos found!')
 125
 126         entries = []
 127
 128         for durl in durls:
 129             size = xpath_text(durl, ['./filesize', './size'])
 130             formats = [{
 131                 'url': durl.find('./url').text,
 132                 'filesize': int_or_none(size),
 133             }]
 134             for backup_url in durl.findall('./backup_url/url'):
 135                 formats.append({
 136                     'url': backup_url.text,
 137                     # backup URLs have lower priorities
 138                     'preference': -2 if 'hd.mp4' in backup_url.text else -3,
 139                 })
 140
 141             self._sort_formats(formats)
 142
 143             entries.append({
 144                 'id': '%s_part%s' % (cid, xpath_text(durl, './order')),
 145                 'duration': int_or_none(xpath_text(durl, './length'), 1000),
 146                 'formats': formats,
 147             })
 148
 149         title = self._html_search_regex('<h1[^>]+title="([^"]+)">', webpage, 'title')
 150         description = self._html_search_meta('description', webpage)
 151         datetime_str = self._html_search_regex(
 152             r'<time[^>]+datetime="([^"]+)"', webpage, 'upload time', fatal=False)
 153         timestamp = None
 154         if datetime_str:
 155             timestamp = calendar.timegm(datetime.datetime.strptime(datetime_str, '%Y-%m-%dT%H:%M').timetuple())
 156
 157         # TODO 'view_count' requires deobfuscating Javascript
 158         info = {
 159             'id': compat_str(cid),
 160             'title': title,
 161             'description': description,
 162             'timestamp': timestamp,
 163             'thumbnail': self._html_search_meta('thumbnailUrl', webpage),
 164             'duration': float_or_none(xpath_text(info_xml, './timelength'), scale=1000),
 165         }
 166
 167         uploader_mobj = re.search(
 168             r'<a[^>]+href="https?://space\.bilibili\.com/(?P<id>\d+)"[^>]+title="(?P<name>[^"]+)"',
 169             webpage)
 170         if uploader_mobj:
 171             info.update({
 172                 'uploader': uploader_mobj.group('name'),
 173                 'uploader_id': uploader_mobj.group('id'),
 174             })
 175
 176         for entry in entries:
 177             entry.update(info)
 178
 179         if len(entries) == 1:
 180             return entries[0]
 181         else:
 182             for idx, entry in enumerate(entries):
 183                 entry['id'] = '%s_part%d' % (video_id, (idx + 1))
 184
 185             return {
 186                 '_type': 'multi_video',
 187                 'id': video_id,
 188                 'title': title,
 189                 'description': description,
 190                 'entries': entries,
 191             }