[bilibili] Fix and improve extraction (closes #15048, closes #15430, closes #15622, closes #15863)

This commit is contained in:
Sergey M․ 2018-03-16 00:19:17 +07:00
parent 8e70c1bfac
commit 3526c3043b
No known key found for this signature in database
GPG Key ID: 2C393E0F18A9236D
1 changed files with 93 additions and 52 deletions

View File

@ -27,14 +27,14 @@ class BiliBiliIE(InfoExtractor):
_TESTS = [{ _TESTS = [{
'url': 'http://www.bilibili.tv/video/av1074402/', 'url': 'http://www.bilibili.tv/video/av1074402/',
'md5': '9fa226fe2b8a9a4d5a69b4c6a183417e', 'md5': '5f7d29e1a2872f3df0cf76b1f87d3788',
'info_dict': { 'info_dict': {
'id': '1074402', 'id': '1074402',
'ext': 'mp4', 'ext': 'flv',
'title': '【金坷垃】金泡沫', 'title': '【金坷垃】金泡沫',
'description': 'md5:ce18c2a2d2193f0df2917d270f2e5923', 'description': 'md5:ce18c2a2d2193f0df2917d270f2e5923',
'duration': 308.315, 'duration': 308.067,
'timestamp': 1398012660, 'timestamp': 1398012678,
'upload_date': '20140420', 'upload_date': '20140420',
'thumbnail': r're:^https?://.+\.jpg', 'thumbnail': r're:^https?://.+\.jpg',
'uploader': '菊子桑', 'uploader': '菊子桑',
@ -59,17 +59,38 @@ class BiliBiliIE(InfoExtractor):
'url': 'http://www.bilibili.com/video/av8903802/', 'url': 'http://www.bilibili.com/video/av8903802/',
'info_dict': { 'info_dict': {
'id': '8903802', 'id': '8903802',
'ext': 'mp4',
'title': '阿滴英文|英文歌分享#6 "Closer', 'title': '阿滴英文|英文歌分享#6 "Closer',
'description': '滴妹今天唱Closer給你聽! 有史以来,被推最多次也是最久的歌曲,其实歌词跟我原本想像差蛮多的,不过还是好听! 微博@阿滴英文', 'description': '滴妹今天唱Closer給你聽! 有史以来,被推最多次也是最久的歌曲,其实歌词跟我原本想像差蛮多的,不过还是好听! 微博@阿滴英文',
'uploader': '阿滴英文',
'uploader_id': '65880958',
'timestamp': 1488382620,
'upload_date': '20170301',
},
'params': {
'skip_download': True, # Test metadata only
}, },
'playlist': [{
'info_dict': {
'id': '8903802_part1',
'ext': 'flv',
'title': '阿滴英文|英文歌分享#6 "Closer',
'description': 'md5:3b1b9e25b78da4ef87e9b548b88ee76a',
'uploader': '阿滴英文',
'uploader_id': '65880958',
'timestamp': 1488382634,
'upload_date': '20170301',
},
'params': {
'skip_download': True, # Test metadata only
},
}, {
'info_dict': {
'id': '8903802_part2',
'ext': 'flv',
'title': '阿滴英文|英文歌分享#6 "Closer',
'description': 'md5:3b1b9e25b78da4ef87e9b548b88ee76a',
'uploader': '阿滴英文',
'uploader_id': '65880958',
'timestamp': 1488382634,
'upload_date': '20170301',
},
'params': {
'skip_download': True, # Test metadata only
},
}]
}] }]
_APP_KEY = '84956560bc028eb7' _APP_KEY = '84956560bc028eb7'
@ -92,9 +113,13 @@ class BiliBiliIE(InfoExtractor):
webpage = self._download_webpage(url, video_id) webpage = self._download_webpage(url, video_id)
if 'anime/' not in url: if 'anime/' not in url:
cid = compat_parse_qs(self._search_regex( cid = self._search_regex(
[r'EmbedPlayer\([^)]+,\s*"([^"]+)"\)', r'cid(?:["\']:|=)(\d+)', webpage, 'cid',
r'<iframe[^>]+src="https://secure\.bilibili\.com/secure,([^"]+)"'], default=None
) or compat_parse_qs(self._search_regex(
[r'1EmbedPlayer\([^)]+,\s*"([^"]+)"\)',
r'1EmbedPlayer\([^)]+,\s*\\"([^"]+)\\"\)',
r'1<iframe[^>]+src="https://secure\.bilibili\.com/secure,([^"]+)"'],
webpage, 'player parameters'))['cid'][0] webpage, 'player parameters'))['cid'][0]
else: else:
if 'no_bangumi_tip' not in smuggled_data: if 'no_bangumi_tip' not in smuggled_data:
@ -114,53 +139,66 @@ class BiliBiliIE(InfoExtractor):
self._report_error(js) self._report_error(js)
cid = js['result']['cid'] cid = js['result']['cid']
payload = 'appkey=%s&cid=%s&otype=json&quality=2&type=mp4' % (self._APP_KEY, cid)
sign = hashlib.md5((payload + self._BILIBILI_KEY).encode('utf-8')).hexdigest()
headers = { headers = {
'Referer': url 'Referer': url
} }
headers.update(self.geo_verification_headers()) headers.update(self.geo_verification_headers())
video_info = self._download_json(
'http://interface.bilibili.com/playurl?%s&sign=%s' % (payload, sign),
video_id, note='Downloading video info page',
headers=headers)
if 'durl' not in video_info:
self._report_error(video_info)
entries = [] entries = []
for idx, durl in enumerate(video_info['durl']): RENDITIONS = ('qn=80&quality=80&type=', 'quality=2&type=mp4')
formats = [{ for num, rendition in enumerate(RENDITIONS, start=1):
'url': durl['url'], payload = 'appkey=%s&cid=%s&otype=json&%s' % (self._APP_KEY, cid, rendition)
'filesize': int_or_none(durl['size']), sign = hashlib.md5((payload + self._BILIBILI_KEY).encode('utf-8')).hexdigest()
}]
for backup_url in durl.get('backup_url', []): video_info = self._download_json(
formats.append({ 'http://interface.bilibili.com/playurl?%s&sign=%s' % (payload, sign),
'url': backup_url, video_id, note='Downloading video info page',
# backup URLs have lower priorities headers=headers, fatal=num == len(RENDITIONS))
'preference': -2 if 'hd.mp4' in backup_url else -3,
if not video_info:
continue
if 'durl' not in video_info:
if num < len(RENDITIONS):
continue
self._report_error(video_info)
for idx, durl in enumerate(video_info['durl']):
formats = [{
'url': durl['url'],
'filesize': int_or_none(durl['size']),
}]
for backup_url in durl.get('backup_url', []):
formats.append({
'url': backup_url,
# backup URLs have lower priorities
'preference': -2 if 'hd.mp4' in backup_url else -3,
})
for a_format in formats:
a_format.setdefault('http_headers', {}).update({
'Referer': url,
})
self._sort_formats(formats)
entries.append({
'id': '%s_part%s' % (video_id, idx),
'duration': float_or_none(durl.get('length'), 1000),
'formats': formats,
}) })
break
for a_format in formats: title = self._html_search_regex(
a_format.setdefault('http_headers', {}).update({ ('<h1[^>]+\btitle=(["\'])(?P<title>(?:(?!\1).)+)\1',
'Referer': url, '(?s)<h1[^>]*>(?P<title>.+?)</h1>'), webpage, 'title',
}) group='title')
self._sort_formats(formats)
entries.append({
'id': '%s_part%s' % (video_id, idx),
'duration': float_or_none(durl.get('length'), 1000),
'formats': formats,
})
title = self._html_search_regex('<h1[^>]*>([^<]+)</h1>', webpage, 'title')
description = self._html_search_meta('description', webpage) description = self._html_search_meta('description', webpage)
timestamp = unified_timestamp(self._html_search_regex( timestamp = unified_timestamp(self._html_search_regex(
r'<time[^>]+datetime="([^"]+)"', webpage, 'upload time', default=None)) r'<time[^>]+datetime="([^"]+)"', webpage, 'upload time',
default=None) or self._html_search_meta(
'uploadDate', webpage, 'timestamp', default=None))
thumbnail = self._html_search_meta(['og:image', 'thumbnailUrl'], webpage) thumbnail = self._html_search_meta(['og:image', 'thumbnailUrl'], webpage)
# TODO 'view_count' requires deobfuscating Javascript # TODO 'view_count' requires deobfuscating Javascript
@ -174,13 +212,16 @@ class BiliBiliIE(InfoExtractor):
} }
uploader_mobj = re.search( uploader_mobj = re.search(
r'<a[^>]+href="(?:https?:)?//space\.bilibili\.com/(?P<id>\d+)"[^>]+title="(?P<name>[^"]+)"', r'<a[^>]+href="(?:https?:)?//space\.bilibili\.com/(?P<id>\d+)"[^>]*>(?P<name>[^<]+)',
webpage) webpage)
if uploader_mobj: if uploader_mobj:
info.update({ info.update({
'uploader': uploader_mobj.group('name'), 'uploader': uploader_mobj.group('name'),
'uploader_id': uploader_mobj.group('id'), 'uploader_id': uploader_mobj.group('id'),
}) })
if not info.get('uploader'):
info['uploader'] = self._html_search_meta(
'author', webpage, 'uploader', default=None)
for entry in entries: for entry in entries:
entry.update(info) entry.update(info)