Python 爬虫字体映射实战：HTML 里一个字都没有（CH-007）

1
"""CH-007：答案 = 5 段各自的暗号字按页码拼起来（共 5 个字）"""
2
import base64
3
import hashlib
4
import hmac
5
import json
6
import re
7
import time
8
import urllib.parse
9

10
import requests
11

12
BASE = "https://spider.jsnote.top"
13
ROUTE = "/verse"
14
INIT_URL = f"{BASE}/api/challenges/CH-007/init"
15
DATA_URL = f"{BASE}/api/challenges/CH-007/data"
16
SHEET_CSS_PATH = "/playground/verse/glyph.css"     # 静态字体库样式（不需登录）
17
TOTAL_PAGES = 5
18

19

20
def sign(hmac_key, page, timestamp):
21
    """base64(HMAC-SHA256("/verse|page|timestamp", hmacKey))"""
22
    raw = f"{ROUTE}|{page}|{timestamp}"
23
    digest = hmac.new(hmac_key.encode(), raw.encode(), hashlib.sha256).digest()
24
    return base64.b64encode(digest).decode()
25

26

27
def encode_params(page, timestamp):
28
    return urllib.parse.quote(json.dumps({"page": page, "timestamp": timestamp},
29
                                        separators=(",", ":")), safe="")
30

31

32
def decrypt(url_encoded, xor_key):
33
    """URL 解码 → Base64 解码 → 逐字节 XOR → utf-8 → JSON"""
34
    masked = base64.b64decode(urllib.parse.unquote(url_encoded))
35
    plain = bytes(byte ^ xor_key for byte in masked)
36
    return json.loads(plain.decode("utf-8"))
37

38

39
def parse_font_map(session):
40
    """从静态字体库扒出格子宽高和一行一行的字。"""
41
    css = session.get(BASE + SHEET_CSS_PATH, timeout=10).text
42
    head = re.search(r"\.next-glyph\{([^}]*)\}", css).group(1)
43
    cell_w = int(re.search(r"width:(\d+)px", head).group(1))
44
    cell_h = int(re.search(r"height:(\d+)px", head).group(1))
45
    svg_path = re.search(r"url\('([^']+)'\)", head).group(1)
46

47
    svg = session.get(BASE + svg_path, timeout=10).text
48
    rows = re.findall(r"<text[^>]*>(.*?)</text>", svg, re.S)
49
    return cell_w, cell_h, rows
50

51

52
def decode(glyphs, cell_w, cell_h, rows):
53
    """坐标 → 行/列 → 字。"""
54
    chars = []
55
    for glyph in glyphs:
56
        row = glyph["y"] // cell_h
57
        col = glyph["x"] // cell_w
58
        chars.append(rows[row][col])
59
    return "".join(chars)
60

61

62
def main():
63
    session = requests.Session()
64
    session.headers.update({
65
        "Content-Type": "application/json",
66
        "Referer": f"{BASE}/verse?challenge=CH-007",
67
        "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) Chrome/147.0.0.0 Safari/537.36",
68
    })
69
    session.cookies.set("session", "<把浏览器里的 session cookie 粘到这里>",
70
                        domain="spider.jsnote.top")
71

72
    config = session.get(INIT_URL, timeout=10).json()
73
    cell_w, cell_h, rows = parse_font_map(session)
74
    print(f"字体库：格子 {cell_w}x{cell_h}，{len(rows)} 行\n")
75

76
    keys = []
77
    for page in range(1, TOTAL_PAGES + 1):
78
        timestamp = str(int(time.time() * 1000))
79
        response = session.post(DATA_URL, headers={
80
            "X-URL-Sign": sign(config["hmacKey"], page, timestamp),
81
        }, data=json.dumps({"params": encode_params(page, timestamp)}), timeout=10)
82

83
        if response.status_code == 403:
84
            raise SystemExit("403：签名不对。自查 —— raw 是 '/verse|page|timestamp'，"
85
                             "结果是 HMAC-SHA256 的 Base64，路径带前导斜杠")
86
        response.raise_for_status()
87

88
        body = response.json()
89
        payload = decrypt(body["data"], int(body["xorKey"]))
90
        glyphs = payload["glyphs"]
91

92
        text = decode(glyphs, cell_w, cell_h, rows)
93
        key = decode([glyphs[payload["key_index"]]], cell_w, cell_h, rows)
94
        keys.append(key)
95
        print(f"第 {page} 段：{len(text)} 字 → 暗号字「{key}」")
96
        print(f"          开头：{text[:24]}…")
97
        time.sleep(0.5)
98

99
    print(f"\n答案：{''.join(keys)}")
100

101

102
if __name__ == "__main__":
103
    main()