672 字
3 分钟
百度百科X-Bk-Token逆向

地址: 百度百科_全球领先的中文百科全书

JS: https://bkssl.bdimg.com/resource/pages/pc-lemma-search_5755c937.js

API: https://baike.baidu.com/lemma/api/search?enc=utf8&word=双休

具有明确名称的参数(非一个单一字母, 如v, x), 可以先试试全局搜索

image-20250625134912563

没找到就查看发起程序的调用堆栈, 一点点定位; 然后这个函数名就比较”明显”的意图. 很容易就找到了

image-20250625135507947

断点调试逻辑, createApiTokenHeaders函数大概逻辑如下

i = {
"pageToken": "RiVzODzhWR_mcU5VnDcW5g",
"pageTokenExpireTime": 1750647249,
"pageParams": "0"
}
p = new URL("https://baike.baidu.com/search?enc=utf8&word=%E5%8F%8C%E4%BC%91")
x = JSON.stringify({
pageUrl: Number(i.pageParams) ? p.pathname : p.pathname + p.search, // "/search?enc=utf8&word=%E5%8F%8C%E4%BC%91"
expireTime: i.pageTokenExpireTime
})
k = o.HmacSHA256("/lemma/api/searchenc=utf8&word=%E5%8F%8C%E4%BC%91", i.pageToken).toString(o.enc.Base64url)
S = o.enc.Base64url.stringify(o.enc.Utf8.parse(x))
X-Bk-Token = "".concat(S, ".").concat(k)

i是通过函数y返回的, 继续的还原i的生成逻辑, y函数中获取了html中的两个值

image-20250625140108988

调试分析逻辑, 大致逻辑如下

<meta name="visit-id" content="05451938112522025242564761020552325305301355" data-use-parmas="0">
<meta name="visit-time" content="1750649085">
pageTokenExpireTime = Number(<visit-time>)
pageParams = <data-use-parmas>
p = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_-'
l = pageTokenExpireTime % p.length
c = p.slice(l) + p.slice(0,l)
// 原逻辑使用对象做映射, 比较麻烦
// s = {}
// c.split("").map((x,i) => {
// key = i.toString().padStart(2, '0')
// s[key] = x
// })
// // 每两位分割
// f = visit_id.match(/.{1,2}/g)
// pageToken = f.map(x => s[x]).join("")
// 更简洁的写法, 以数组索引作为映射
s = c.split("")
f = visit_id.match(/.{1,2}/g)
pageToken = f.map(x => s[Number(x)]).join("")

转为Python后, 一步一步与网页断点生成的值校验对比是否一致;

经过对比一致, 表明算法无问题, 但是生成的X-Bk-Token无法使用;

实际整体请求经过两次请求,

第一次: https://baike.baidu.com/search 返回 visit-id 相关信息

第二次: https://baike.baidu.com/lemma/api/search 返回最终数据

关键点在于 两个请求需要携带一个共同的cookie BAIDUID_BFESS串联, 否则生成的请求无法使用

完整代码如下

import re
from bs4 import BeautifulSoup
import requests
import json
import urllib.parse
import hmac
import hashlib
import base64
cookies = {
'BAIDUID_BFESS': 'D6142F2E6FF25FF7CF112FED00000B0F:FG=1',
}
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/137.0.0.0 Safari/537.36 Edg/137.0.0.0',
}
params = {
'enc': 'utf8',
'word': '双休',
}
lemma_url = 'https://baike.baidu.com/lemma/api/search'
response = requests.get('https://baike.baidu.com/search', params=params, cookies=cookies, headers=headers)
print(response.status_code, response.url)
soup = BeautifulSoup(response.text, 'html.parser')
visit_id = soup.find('meta', attrs={'name': 'visit-id'}).get('content')
pageTokenExpireTime = int(soup.find('meta', attrs={'name': 'visit-time'}).get('content'))
pageParams = soup.find('meta', attrs={'name': 'visit-id'}).get('data-use-parmas')
# visit_id = '47023738254900123118194151080063185628602011'
# pageTokenExpireTime = 1750672217
# pageParams = "0"
character_set = '0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ_-'
l = pageTokenExpireTime % len(character_set)
s = character_set[l:] + character_set[:l]
print("s: ", s)
s = list(s)
# 每两位分割
f = re.findall(r'.{1,2}', visit_id)
# 映射并生成 pageToken
pageToken = ''.join(map(lambda x : s[int(x)], f))
print(visit_id, pageTokenExpireTime, pageParams)
print("pageToken: ", pageToken)
p = urllib.parse.urlparse(response.url)
page_url = p.path if int(pageParams) else p.path + "?" + p.query
# 生成 JSON 字符串 x
x = json.dumps({
"pageUrl": page_url,
"expireTime": pageTokenExpireTime
}, separators=(",",":"))
print(x)
# '{"pageUrl":"/search?enc=utf8&word=%E5%8F%8C%E4%BC%91","expireTime":1750668767}'
# 生成 HMAC-SHA256 签名 k
key = pageToken.encode('utf-8')
lemma_path = urllib.parse.urlparse(lemma_url).path
lemma_params = urllib.parse.urlencode(params)
message = (lemma_path+lemma_params).encode('utf-8')
print(message)
hmac_hash = hmac.new(key, message, hashlib.sha256).digest()
k = base64.urlsafe_b64encode(hmac_hash).decode('utf-8').rstrip('=')
print("k: ", k)
# 2gKbcIfD7n8XRsnYL9GY_skWswAaUk0-dy51U8LQmEc
# 生成 Base64 URL 编码的字符串 S
S = base64.urlsafe_b64encode(x.encode('utf-8')).decode('utf-8').rstrip('=')
print("S: ", S)
# eyJwYWdlVXJsIjoiL3NlYXJjaD9lbmM9dXRmOCZ3b3JkPSVFNSU4RiU4QyVFNCVCQyU5MSIsImV4cGlyZVRpbWUiOjE3NTA2Njg3Njd9
# 生成 X-Bk-Token
X_Bk_Token = f"{S}.{k}"
print(f"X-Bk-Token: {X_Bk_Token}")
headers.update({
'X-Bk-Token': X_Bk_Token
})
response = requests.get(lemma_url, params=params, cookies=cookies, headers=headers)
print(response.status_code, response.url)
print(response.text)

image-20250625141627563

百度百科X-Bk-Token逆向
https://blog.tiamix.cn/posts/百度百科x-bk-token逆向/
作者
Tiam
发布于
2025-06-25
许可协议
CC BY-NC-SA 4.0