feat(collect): 纯离线生成打通 1ms 签到文字点选 —— errorCode=0 收口
做了什么:
- 新增 tcaptcha/tdc_values.py: 37 个语义名的取值生成器（静态画像 / 会话量 /
每次采集量三层）+ build_plain 四段组装 + generate_collect 纯离线出包
- crypto: 新增 generate_telemetry(mode="python_values")；轨迹生成统一到
trajectory_points()（首点绝对坐标 + 增量 + 固定收尾 [1,1,12]）
- crypto.split_collect_segments 重写: pad 只有 1 个空格时旧 ` {2,}` 判据会漏段，
且 UA 串里的普通空格会假阳性 —— 改为「模 24 算术 + 不落在 JSON 字符串内部 + 结构校验」
- tools/tdc_modules: trajectory 指纹原先指向只做 JSON.stringify 的模块，改为
mouseMoveEvent/touchStartEvent（真轨迹模块），37/37 不变
- client 默认 telemetry_mode 改为 python_values
- 新增 scripts/verify/{align_session,audit_names,verify_values,verify_offline,send_test}.py
与 tools/{cdp_collect,cdp_dump_scripts,tdc_profile}.py
为什么:
- 账本 #40「谁在发 HTTP 是 12 的剩余自变量」被推翻。同 aid/同客户端/同求解器、
答案都对，只换 collect 构造方式: 本次会话构建生成 -> 0 (4/4)，
旧模板回填 -> 12 (2/2)。真因是 collect 必须与本次会话构建同源。
怎么验证:
- .venv/bin/python -m tcaptcha.selftest 全绿（含新增四段切分与纯离线闭环回归）
- scripts/verify/verify_values.py 4 份真机明文 152/152 逐槽全等
- scripts/verify/verify_offline.py 逐槽全等 + 分段几何一致
- scripts/verify/send_test.py --mode python_values 1ms 签到页文字点选 errorCode=0 x4
This commit is contained in:
yml2213
@@ -0,0 +1,202 @@
|
||||
# -*- coding: utf-8 -*-
|
||||
"""
|
||||
tdc_profile.py —— 设备画像（fingerprint profile）的**提取**与**定义**。
|
||||
|
||||
背景（账本 §五 第 0 项，`har/1ms_dianxuan/PROGRESS.md`）:
|
||||
「纯离线生成」的唯一阻塞点，是需要为每个**语义名**提供取值。
|
||||
多数语义名是静态环境量（UA / 屏幕 / 时区 / 语言 / 核数 / charset /
|
||||
WebGL / Canvas），可以从一份真机明文里**按值的内容反查**出来。
|
||||
|
||||
为什么不靠"明文属于哪份构建":
|
||||
真机明文的槽顺序 = **它自己那次会话构建**的 require 顺序，而那份 tdc.js
|
||||
通常没保存（本项目实测：自洽性打分全为负，即明文不属于手上任何构建）。
|
||||
但**值本身携带强语义**：`"MacIntel"` 只可能是 platform，`"UTF-8"` 只可能是
|
||||
charset，`"ANGLE (Apple, …)"` 只可能是 webgl_renderer。所以按内容反查
|
||||
比按位置配对更稳，也不需要那份丢失的构建。
|
||||
|
||||
镜像自洽（案例 §6.3）: 一套画像内的值必须互相一致 ——
|
||||
screen_fingerprint `"1440-2560-2530-24-…"` 里的 1440/2560/24 应当分别等于
|
||||
screen_width / screen_height / screen_color_depth（字段序 = 宽-高-可用高-色深）。
|
||||
本模块会做这个交叉校验。
|
||||
|
||||
标准画像请以 `tcaptcha/tdc_values.py::DEVICE_PROFILE` 为准（那里是 37 个语义名
|
||||
全部对齐后的产物）；本文件只是**从明文按值反查**的开发期工具。
|
||||
|
||||
用法:
|
||||
# 从一份已解密的真机明文提取画像基线
|
||||
.venv/bin/python tools/tdc_profile.py from-sample <明文文件> [--json 输出]
|
||||
|
||||
# 打印内置画像
|
||||
.venv/bin/python tools/tdc_profile.py show
|
||||
"""
|
||||
from __future__ import annotations
|
||||
|
||||
import argparse
|
||||
import json
|
||||
import re
|
||||
import sys
|
||||
from pathlib import Path
|
||||
from typing import Any, Dict, List, Optional, Tuple
|
||||
|
||||
ROOT = Path(__file__).resolve().parents[1]
|
||||
|
||||
|
||||
# --------------------------------------------------------------------------- #
|
||||
# 内容反查：值 -> 语义名（可能多个候选，按置信度排序） #
|
||||
# --------------------------------------------------------------------------- #
|
||||
def guess_names(v: Any) -> List[str]:
|
||||
"""按**值的内容**反推它可能出自哪个语义名（不依赖槽位置）。"""
|
||||
out: List[str] = []
|
||||
if isinstance(v, str):
|
||||
if "Mozilla" in v and "AppleWebKit" in v:
|
||||
out.append("user_agent")
|
||||
if v == "MacIntel":
|
||||
out.append("platform")
|
||||
if re.fullmatch(r"[+-]\d{2}(:\d{2})?", v):
|
||||
out.append("timezone_offset")
|
||||
if v.upper() in ("UTF-8", "UTF8"):
|
||||
out.append("charset")
|
||||
if "ANGLE" in v or "Metal Renderer" in v or "OpenGL" in v:
|
||||
out.append("webgl_renderer")
|
||||
if "Inc." in v or ("Google" in v and "Apple" in v):
|
||||
out.append("webgl_vendor")
|
||||
if v.startswith("GgoA") and len(v) > 40:
|
||||
out.append("canvas_fingerprint")
|
||||
if "captcha" in v and "rand=" in v:
|
||||
out.append("template_url")
|
||||
if v == "98k":
|
||||
out.append("constant_98k")
|
||||
if v == "iframe":
|
||||
out.append("iframe_detect")
|
||||
if "-" in v and "*" in v and re.fullmatch(r"[\d\-*|]+", v):
|
||||
out.append("screen_fingerprint")
|
||||
if v.startswith("http") and "1ms.run" in v:
|
||||
out.append("entry_url")
|
||||
elif v.startswith("http"):
|
||||
out.append("location_href")
|
||||
elif isinstance(v, list):
|
||||
if v and all(isinstance(x, list) and len(x) == 3 for x in v):
|
||||
out.append("trajectory")
|
||||
if v and all(isinstance(x, str) for x in v) and 1 <= len(v) <= 4:
|
||||
out.append("languages")
|
||||
elif isinstance(v, int) and not isinstance(v, bool):
|
||||
if 1_600_000_000 <= v <= 2_100_000_000:
|
||||
out.append("timestamp")
|
||||
return out
|
||||
|
||||
|
||||
# --------------------------------------------------------------------------- #
|
||||
# 内置画像（MacBook 画像基线；取自真机样本，可按需替换/随机化） #
|
||||
# --------------------------------------------------------------------------- #
|
||||
DEFAULT_PROFILE: Dict[str, Any] = {
|
||||
"name": "Mac / Chrome",
|
||||
# --- 明文里能直接看到的静态量 ---
|
||||
"user_agent": ("Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) "
|
||||
"AppleWebKit/537.36 (KHTML, like Gecko) Chrome/155.0.0.0 Safari/537.36"),
|
||||
"platform": "MacIntel",
|
||||
"languages": ["zh-CN", "zh"],
|
||||
"timezone_offset": "+08",
|
||||
"charset": "UTF-8",
|
||||
"hardware_concurrency": 10,
|
||||
# ⚠️ 方向别弄反：真机 screen.width == 1440、screen.height == 2560
|
||||
# （2026-10-12 浏览器实测；1440x2560 竖屏，availHeight 2530）
|
||||
"screen_width": 1440,
|
||||
"screen_height": 2560,
|
||||
"screen_color_depth": 24,
|
||||
"screen_fingerprint": "1440-2560-2530-24-*-*-|-*",
|
||||
"webgl_renderer": "ANGLE (Apple, ANGLE Metal Renderer: Apple M4, Unspecified Version)",
|
||||
"webgl_vendor": "Google Inc. (Apple)",
|
||||
"canvas_fingerprint": ("GgoAAAANSUhEUgAAASwAAACWCAYAAATf3v8bh+ryWcfSBSLkmN3N1"
|
||||
"PQinPL7Bi3TG0qE2CogAAAABJRU5ErkJggg=="),
|
||||
"constant_98k": "98k",
|
||||
"cookie_enabled": 1,
|
||||
"touch_detect": 2, # 真机实测（该机 maxTouchPoints 为 0，值仍是 2）
|
||||
"viewport_size": [360, 360],
|
||||
}
|
||||
|
||||
|
||||
def cross_check(profile: Dict[str, Any]) -> List[str]:
|
||||
"""做镜像自洽性交叉校验（案例 §6.3）。返回不一致项列表。"""
|
||||
problems: List[str] = []
|
||||
fp = profile.get("screen_fingerprint") or ""
|
||||
parts = fp.split("-")
|
||||
if len(parts) >= 4:
|
||||
try:
|
||||
# 字段序 = 宽-高-可用高-色深（2026-10-12 与浏览器真值 1440/2560/2530/24 对齐；
|
||||
# 旧代码把它当成 高-宽-?-色深，方向是反的）
|
||||
w, h, _avail, depth = int(parts[0]), int(parts[1]), int(parts[2]), int(parts[3])
|
||||
if h != profile.get("screen_height"):
|
||||
problems.append("screen_fingerprint 的高度 %d != screen_height %s"
|
||||
% (h, profile.get("screen_height")))
|
||||
if w != profile.get("screen_width"):
|
||||
problems.append("screen_fingerprint 的宽度 %d != screen_width %s"
|
||||
% (w, profile.get("screen_width")))
|
||||
if depth != profile.get("screen_color_depth"):
|
||||
problems.append("screen_fingerprint 的色深 %d != screen_color_depth %s"
|
||||
% (depth, profile.get("screen_color_depth")))
|
||||
except ValueError:
|
||||
problems.append("screen_fingerprint 格式无法解析: %r" % fp)
|
||||
ua = profile.get("user_agent") or ""
|
||||
if "Chrome/" in ua and "Macintosh" in ua and profile.get("platform") != "MacIntel":
|
||||
problems.append("UA 说是 Mac，但 platform=%r" % profile.get("platform"))
|
||||
return problems
|
||||
|
||||
|
||||
def from_sample(path: str) -> Tuple[Dict[str, Any], List[Dict]]:
|
||||
"""从一份真机明文，按内容反查提取画像基线。
|
||||
|
||||
返回 (识别到的状态量, 未识别的槽列表)。
|
||||
"""
|
||||
raw = open(path, encoding="latin1").read().rstrip("\x00")
|
||||
logical = re.sub(r" {2,}", "", raw)
|
||||
cd = json.loads(logical)["cd"]
|
||||
|
||||
found: Dict[str, Any] = {}
|
||||
unidentified: List[Dict] = []
|
||||
for i, v in enumerate(cd):
|
||||
cands = guess_names(v)
|
||||
if len(cands) == 1:
|
||||
found[cands[0]] = v
|
||||
elif cands:
|
||||
found.setdefault(cands[0], v)
|
||||
else:
|
||||
unidentified.append({"slot": i, "value": v})
|
||||
return found, unidentified
|
||||
|
||||
|
||||
def main() -> int:
|
||||
ap = argparse.ArgumentParser(description="TDC 设备画像提取/查看")
|
||||
sub = ap.add_subparsers(dest="cmd", required=True)
|
||||
p1 = sub.add_parser("from-sample", help="从真机明文提取画像基线")
|
||||
p1.add_argument("plain")
|
||||
p1.add_argument("--json", default=None)
|
||||
sub.add_parser("show", help="打印内置画像")
|
||||
|
||||
args = ap.parse_args()
|
||||
if args.cmd == "show":
|
||||
print(json.dumps(DEFAULT_PROFILE, ensure_ascii=False, indent=1))
|
||||
probs = cross_check(DEFAULT_PROFILE)
|
||||
print()
|
||||
print("交叉校验:", "全部自洽" if not probs else probs)
|
||||
return 0
|
||||
|
||||
found, unk = from_sample(args.plain)
|
||||
print("从 %s 反查到的状态量:" % Path(args.plain).name)
|
||||
for k, v in found.items():
|
||||
print(" %-22s %s" % (k, json.dumps(v, ensure_ascii=False)[:70]))
|
||||
print()
|
||||
print("未识别的槽: %d 个" % len(unk))
|
||||
for u in unk:
|
||||
print(" 槽%-3d %s" % (u["slot"], json.dumps(u["value"], ensure_ascii=False)[:60]))
|
||||
print()
|
||||
probs = cross_check(found)
|
||||
print("交叉校验:", "全部自洽 ✓" if not probs else probs)
|
||||
if args.json:
|
||||
Path(args.json).write_text(json.dumps(found, ensure_ascii=False, indent=1),
|
||||
encoding="utf-8")
|
||||
print("画像 -> %s" % args.json)
|
||||
return 0
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
raise SystemExit(main())
|
||||
Reference in New Issue
Block a user