Python 爬虫验证码实战：AES 加密报文 + 每页算术验证码（CH-009）

1
{
2
  "code": "0",
3
  "data": {
4
    "list": [
5
      {"projectName": "六安特安新能源…光伏电站项目",
6
       "projectCode": "2610-341504-04-01-309092",
7
       "itemName": "固定资产投资项目备案（内资项目）",
8
       "unitName": "六安市叶集区发展和改革委员会…",
9
       "approveResultName": "通过",
10
       "itemHandleTime": "2026/10/01 11:01:15"}
11
    ],
12
    "pageNumber": 1, "pageSize": 10,
13
    "totalPage": 10, "totalRow": 100,
14
    "firstPage": true, "lastPage": false
15
  },
16
  "message": "查询成功",
17
  "status": 200
18
}

1
"""CH-009：答案 = 10 页共 100 条批复文号「最后一段数字」之和
2

3
三道坎：
4
  1. 报文是 AES-ECB(PKCS7) 加密后的 hex：请求体自己加密，响应体自己解
5
  2. 请求头要带 Adam-AppKey
6
  3. 每翻一页都得先解一道算术验证码（一次性）
7
"""
8
import base64
9
import json
10
import re
11
import time
12

13
import requests
14
from Crypto.Cipher import AES
15
from Crypto.Util.Padding import pad, unpad
16

17
BASE = "https://spider.jsnote.top"
18
CAPTCHA_URL = f"{BASE}/api/challenges/CH-009/captcha"
19
LIST_URL = f"{BASE}/api/challenges/CH-009/list"
20
APP_KEY = "3b08c9a1-924b-3876-9129-cca5976edfa9"     # 前端写死的
21
AES_KEY = b"Nr2bNQtm5o9A0qqz"                        # 前端写死的
22
PAGE_SIZE = 10
23
MAX_PAGES = 10
24
MAX_RETRY = 6
25

26

27
def encrypt(payload: dict) -> str:
28
    """JSON -> AES-ECB(PKCS7) -> hex 文本（body 里放的就是这段 hex，不是 JSON）"""
29
    raw = json.dumps(payload, ensure_ascii=False, separators=(",", ":")).encode("utf-8")
30
    return AES.new(AES_KEY, AES.MODE_ECB).encrypt(pad(raw, AES.block_size)).hex()
31

32

33
def decrypt(hex_text: str) -> dict:
34
    """hex 文本 -> AES-ECB -> JSON"""
35
    clean = re.sub(r"\s+", "", str(hex_text))
36
    plain = unpad(AES.new(AES_KEY, AES.MODE_ECB).decrypt(bytes.fromhex(clean)),
37
                  AES.block_size)
38
    return json.loads(plain.decode("utf-8"))
39

40

41
session = requests.Session()
42
session.headers.update({
43
    "Accept": "*/*",
44
    "Adam-AppKey": APP_KEY,
45
    "Content-Type": "application/json;charset=utf-8",
46
    "Referer": f"{BASE}/projects?challenge=CH-009",
47
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) Chrome/147.0.0.0 Safari/537.36",
48
    "X-Requested-With": "XMLHttpRequest",
49
})
50
session.cookies.set("session", "<把浏览器里的 session cookie 粘到这里>",
51
                    domain="spider.jsnote.top")
52

53

54
def call(url, payload) -> dict:
55
    response = session.post(url, data=encrypt(payload), timeout=20)
56
    if response.status_code == 401:
57
        raise SystemExit("401：需要登录态（Cookie 没填对？）")
58
    if not response.ok:
59
        raise SystemExit(f"{response.status_code}：{response.text[:200]}")
60
    return decrypt(response.text)
61

62

63
def solve_captcha(image_bytes: bytes) -> int:
64
    """识别 `8+2=?` 并算答案。"""
65
    import ddddocr                                            # pip install ddddocr
66

67
    text = ddddocr.DdddOcr(show_ad=False).classification(image_bytes)
68
    normalised = (text.replace("×", "*").replace("✕", "*").replace("╳", "*")
69
                      .replace("x", "*").replace("X", "*")
70
                      .replace("÷", "/").replace("／", "/"))
71
    match = re.search(r"(\d+)\s*([+\-*/])\s*(\d+)", normalised)
72
    if not match:
73
        raise ValueError(f"验证码没认出来：{text!r}")
74
    left, operator, right = int(match.group(1)), match.group(2), int(match.group(3))
75
    return {"+": left + right, "-": left - right, "*": left * right}.get(operator, left // right)
76

77

78
def fetch_page(page_number: int) -> dict:
79
    """取一页：每页都要先拿一张新验证码；识别失败或答错都换一张重来。"""
80
    last_message = ""
81
    for _attempt in range(MAX_RETRY):
82
        captcha = call(CAPTCHA_URL, {"ajaxRequestTimestamp": int(time.time() * 1000)})
83
        image_base64 = str(captcha["data"]["captchaImage"]).split(",")[-1]
84

85
        try:
86
            answer = solve_captcha(base64.b64decode(image_base64))
87
        except (ValueError, ImportError) as error:
88
            last_message = str(error)
89
            time.sleep(0.5)
90
            continue
91

92
        data = call(LIST_URL, {
93
            "ajaxRequestTimestamp": int(time.time() * 1000),
94
            "pageSize": PAGE_SIZE,
95
            "pageNumber": page_number,
96
            "nameOrCode": "",
97
            "queryType": "first",
98
            "selectType": "pname",
99
            "captchaId": captcha["data"]["captchaId"],
100
            "yzm": str(answer),
101
        })
102
        if data.get("code") == "0":
103
            return data["data"]
104

105
        last_message = f"{data.get('code')} {data.get('message')}"
106
        time.sleep(0.5)
107

108
    raise SystemExit(f"第 {page_number} 页连续 {MAX_RETRY} 次都没过：{last_message}")
109

110

111
def main():
112
    total = 0
113
    rows_seen = 0
114
    total_page = 0
115

116
    for page_number in range(1, MAX_PAGES + 1):
117
        page = fetch_page(page_number)
118
        total_page = int(page.get("totalPage") or 0)
119

120
        for row in page.get("list", []):
121
            tail = str(row["projectCode"]).rsplit("-", 1)[-1]      # 2610-341504-04-01-429215 -> 429215
122
            if tail.isdigit():
123
                total += int(tail)
124
                rows_seen += 1
125

126
        print(f"第 {page_number}/{total_page or '?'} 页：{len(page.get('list', []))} 条，"
127
              f"累计 {rows_seen} 条，小计 {total}")
128

129
        if page.get("lastPage") or page_number >= (total_page or MAX_PAGES):
130
            break
131

132
    print(f"\n共 {rows_seen} 条，批复文号末段数字之和：{total}")
133

134

135
if __name__ == "__main__":
136
    main()