async def get_page_info_ep(self, url):
async with aiohttp.ClientSession() as session:
async with session.request("get", url, headers=self.headers) as r:
data = json.loads(match1(await r.text(), "__INITIAL_STATE__=({.+?});"))
title = data.get("h1Title") or match1(html, "<title>(.+?)_番剧_bilibili_哔哩哔哩<")
cid = data["epInfo"]["cid"]
bvid = data["epInfo"]["bvid"]
mediaInfo = data["mediaInfo"]
season_type = mediaInfo.get("season_type") or mediaInfo.get("ssType")
upInfo = mediaInfo.get("upInfo")
artist = upInfo and upInfo.get("name")
return bvid, cid, title, artist, season_type
async def get_play_info(self):
ret = dict()
title = ""
artist = ""
async with aiohttp.ClientSession() as sess:
async with sess.request("get", self.room_url,
headers=self.headers) as resp:
html = await resp.text()
self.rid = match1(
html,
"\$ROOM\.room_id\s*=\s*(\d+)",
"room_id\s*=\s*(\d+)",
'"room_id.?":(\d+)',
"data-onlineid=(\d+)",
)
title = match1(html, 'Title-head\w*">([^<]+)<')
artist = match1(html, 'Title-anchorName\w*" title="([^"]+)"')
if not title or not artist:
async with sess.request("get",
self.api_url2 + self.rid,
headers=self.headers) as resp1:
room_data = await resp1.json()
if room_data["error"] == 0:
room_data = room_data["data"]
title = room_data["room_name"]
artist = room_data["owner_name"]
ret["title"] = u"{} - {}".format(title, artist)
js_enc = await self.get_h5enc(html)
params = {"cdn": "", "iar": 0, "ive": 0}
await self.ub98484234(js_enc, params)
params["rate"] = 0
live_data = None
async with aiohttp.ClientSession() as sess:
async with sess.request("post",
self.api_url3 + self.rid,
data=params,
headers=self.headers) as resp:
live_data = await resp.json()
if live_data["error"]:
print(live_data["msg"])
live_data = live_data["data"]
# print(live_data)
ret["play_url"] = "{}/{}".format(live_data["rtmp_url"],
live_data["rtmp_live"])
# print(ret)
return ret
async def get_page_info_ep(self, url):
async with aiohttp.ClientSession() as session:
async with session.request('get', url, headers=self.headers) as r:
data = json.loads(
match1(await r.text(), '__INITIAL_STATE__=({.+?});'))
title = data.get('h1Title') or match1(
html, '<title>(.+?)_番剧_bilibili_哔哩哔哩<')
cid = data['epInfo']['cid']
bvid = data['epInfo']['bvid']
mediaInfo = data['mediaInfo']
season_type = mediaInfo.get('season_type') or mediaInfo.get('ssType')
upInfo = mediaInfo.get('upInfo')
artist = upInfo and upInfo.get('name')
return bvid, cid, title, artist, season_type
async def get_play_info(self):
ret = dict()
title = ''
artist = ''
async with aiohttp.ClientSession() as sess:
async with sess.request('get', self.room_url,
headers=self.headers) as resp:
h = await resp.text()
json_stream = match1(h, '"stream": "([a-zA-Z0-9+=/]+)"')
assert json_stream, "live video is offline"
data = json.loads(base64.b64decode(json_stream).decode())
assert data['status'] == 200, data['msg']
room_info = data['data'][0]['gameLiveInfo']
ret['title'] = u'{}「{} - {}」'.format(room_info['roomName'],
room_info['nick'],
room_info['introduction'])
stream_info = random.choice(
data['data'][0]['gameStreamInfoList'])
sStreamName = stream_info['sStreamName']
for sType in ('flv', 'hls'):
sType = sType.title()
sUrl = stream_info['s{}Url'.format(sType)]
sUrlSuffix = stream_info['s{}UrlSuffix'.format(sType)]
sAntiCode = stream_info['s{}AntiCode'.format(sType)]
ret['play_url'] = html.unescape(u'{}/{}.{}?{}'.format(
sUrl, sStreamName, sUrlSuffix, sAntiCode))
break
# print(ret)
return ret
async def get_play_info(self):
ret = dict()
title = ""
artist = ""
async with aiohttp.ClientSession() as sess:
async with sess.request("get", self.room_url,
headers=self.headers) as resp:
h = await resp.text()
json_stream = match1(h, '"stream": "([a-zA-Z0-9+=/]+)"')
assert json_stream, "live video is offline"
data = json.loads(base64.b64decode(json_stream).decode())
assert data["status"] == 200, data["msg"]
room_info = data["data"][0]["gameLiveInfo"]
ret["title"] = u"{}「{} - {}」".format(room_info["roomName"],
room_info["nick"],
room_info["introduction"])
stream_info = random.choice(
data["data"][0]["gameStreamInfoList"])
sStreamName = stream_info["sStreamName"]
for sType in ("flv", "hls"):
sType = sType.title()
sUrl = stream_info["s{}Url".format(sType)]
sUrlSuffix = stream_info["s{}UrlSuffix".format(sType)]
sAntiCode = stream_info["s{}AntiCode".format(sType)]
ret["play_url"] = html.unescape(u"{}/{}.{}?{}".format(
sUrl, sStreamName, sUrlSuffix, sAntiCode))
break
# print(ret)
return ret
async def get_play_info(self):
ret = dict()
title = ''
artist = ''
async with aiohttp.ClientSession() as sess:
async with sess.request('get', self.room_url, headers=self.headers) as resp:
html = await resp.text()
self.rid = match1(html, '\$ROOM\.room_id\s*=\s*(\d+)',
'room_id\s*=\s*(\d+)',
'"room_id.?":(\d+)',
'data-onlineid=(\d+)')
title = match1(html, 'Title-head\w*">([^<]+)<')
artist = match1(html, 'Title-anchorName\w*" title="([^"]+)"')
if not title or not artist:
async with sess.request('get', self.api_url2 + self.rid, headers=self.headers) as resp1:
room_data = await resp1.json()
if room_data['error'] == 0:
room_data = room_data['data']
title = room_data['room_name']
artist = room_data['owner_name']
ret['title'] = u'{} - {}'.format(title, artist)
js_enc = await self.get_h5enc(html)
params = {
'cdn': '',
'iar': 0,
'ive': 0
}
await self.ub98484234(js_enc, params)
params['rate'] = 0
live_data = None
async with aiohttp.ClientSession() as sess:
async with sess.request('post', self.api_url3 + self.rid, data=params, headers=self.headers) as resp:
live_data = await resp.json()
if live_data['error']:
print(live_data['msg'])
live_data = live_data["data"]
# print(live_data)
ret['play_url'] = '{}/{}'.format(live_data['rtmp_url'], live_data['rtmp_live'])
# print(ret)
return ret
async def get_h5enc(self, html):
js_enc = match1(html, '(var vdwdae325w_64we =[\s\S]+?)\s*</script>')
if js_enc is None or 'ub98484234(' not in js_enc:
params = {
'rids': self.rid
}
async with aiohttp.ClientSession() as sess:
async with sess.request('get', self.api_url1, params=params, headers=self.headers) as resp:
data = await resp.json()
assert data['error'] == 0, data['msg']
js_enc = data['data']['room' + self.rid]
return js_enc
async def get_page_info(self, url):
page_index = match1(url, "\?p=(\d+)", "index_(\d+)\.") or "1"
data = None
async with aiohttp.ClientSession() as session:
async with session.request("get", url, headers=self.headers) as r:
data = json.loads(match1(await r.text(), "__INITIAL_STATE__=({.+?});"))["videoData"]
bvid = data["bvid"]
title = data["title"]
artist = data["owner"]["name"]
pages = data["pages"]
for page in pages:
index = str(page["page"])
subtitle = page["part"]
if index == page_index:
cid = page["cid"]
if len(pages) > 1:
title = u"{} - {} - {}".format(title, index, subtitle)
elif subtitle and subtitle != title:
title = u"{} - {}".format(title, subtitle)
break
return bvid, cid, title, artist
async def get_h5enc(self, html):
js_enc = match1(html, "(var vdwdae325w_64we =[\s\S]+?)\s*</script>")
if js_enc is None or "ub98484234(" not in js_enc:
params = {"rids": self.rid}
async with aiohttp.ClientSession() as sess:
async with sess.request("get",
self.api_url1,
params=params,
headers=self.headers) as resp:
data = await resp.json()
assert data["error"] == 0, data["msg"]
js_enc = data["data"]["room" + self.rid]
return js_enc
async def get_page_info(self, url):
page_index = match1(url, '\?p=(\d+)', 'index_(\d+)\.') or '1'
data = None
async with aiohttp.ClientSession() as session:
async with session.request('get', url, headers=self.headers) as r:
data = json.loads(
match1(await r.text(),
'__INITIAL_STATE__=({.+?});'))['videoData']
bvid = data['bvid']
title = data['title']
artist = data['owner']['name']
pages = data['pages']
for page in pages:
index = str(page['page'])
subtitle = page['part']
if index == page_index:
cid = page['cid']
if len(pages) > 1:
title = u'{} - {} - {}'.format(title, index, subtitle)
elif subtitle and subtitle != title:
title = u'{} - {}'.format(title, subtitle)
break
return bvid, cid, title, artist
async def ub98484234(self, js_enc, params):
names_dict = {
'debugMessages': get_random_name(8),
'decryptedCodes': get_random_name(8),
'resoult': get_random_name(8),
'_ub98484234': get_random_name(8),
'workflow': match1(js_enc, 'function ub98484234\(.+?\Weval\((\w+)\);'),
}
js_dom = '''
{debugMessages} = {{{decryptedCodes}: []}};
if (!this.window) {{window = {{}};}}
if (!this.document) {{document = {{}};}}
'''.format(**names_dict)
js_patch = '''
{debugMessages}.{decryptedCodes}.push({workflow});
var patchCode = function(workflow) {{
var testVari = /(\w+)=(\w+)\([\w\+]+\);.*?(\w+)="\w+";/.exec(workflow);
if (testVari && testVari[1] == testVari[2]) {{
{workflow} += testVari[1] + "[" + testVari[3] + "] = function() {{return true;}};";
}}
}};
patchCode({workflow});
var subWorkflow = /(?:\w+=)?eval\((\w+)\)/.exec({workflow});
if (subWorkflow) {{
var subPatch = `
{debugMessages}.{decryptedCodes}.push('sub workflow: ' + subWorkflow);
patchCode(subWorkflow);
`.replace(/subWorkflow/g, subWorkflow[1]) + subWorkflow[0];
{workflow} = {workflow}.replace(subWorkflow[0], subPatch);
}}
eval({workflow});
'''.format(**names_dict)
js_debug = '''
var {_ub98484234} = ub98484234;
ub98484234 = function(p1, p2, p3) {{
try {{
var resoult = {_ub98484234}(p1, p2, p3);
{debugMessages}.{resoult} = resoult;
}} catch(e) {{
{debugMessages}.{resoult} = e.message;
}}
return {debugMessages};
}};
'''.format(**names_dict)
js_enc = js_enc.replace('eval({workflow});'.format(**names_dict), js_patch)
js_ctx = JSEngine()
js_ctx.append(self.js_md5)
js_ctx.append(js_dom)
js_ctx.append(js_enc)
js_ctx.append(js_debug)
did = uuid.uuid4().hex
tt = str(int(time.time()))
ub98484234 = js_ctx.call('ub98484234', self.rid, did, tt)
# print(ub98484234)
# extractor.logger.debug('ub98484234: %s', ub98484234)
ub98484234 = ub98484234[names_dict['resoult']]
params.update({
'v': match1(ub98484234, 'v=(\d+)'),
'did': did,
'tt': tt,
'sign': match1(ub98484234, 'sign=(\w{32})')
})