yml2213
344eea2194
做了什么:
- 包改为分层: tcaptcha/{protocol,telemetry,solvers,ocr,adapters,models,tests}
* protocol/ client.py transport.py proxy.py
* telemetry/ crypto.py values.py key.py opcodes.py
disasm.py program.py layout.py modules.py
* solvers/ text_click.py icon_click.py icon_click_rotate.py slider.py
* ocr/ engines.py slicer.py
* adapters/ one_ms.py
- tools/ 里原本是**包运行时依赖**的 tdc_{disasm,program,layout,modules}.py 并入 telemetry/
（放在 tools/ 属于分层倒置），tools/ 只留 CLI: tdc_key_pure.py / tdc_profile.py / cdp_*.js
- crypto.py 529 -> 340 行: 切除「成功样本模板回填」与「Node 插桩兜底」两条老路；
generate_telemetry 只剩 python_values 一种模式，传别的会直接报错并说明原因
- 求解器改名分清: solver.py -> solvers/icon_click.py（client 实际在用的 YOLO+ResNet18 ONNX），
icon_click_solver.py -> solvers/icon_click_rotate.py（仅测试在用，留作对照）
- pyproject: 去掉 ultralytics/onnx（训练栈已删），补上真正在用的 onnxruntime
删除了什么:
- 雨云整条线 + dist/rainyun_qinglong（104M 旧快照）
- 模板回填: models/success_collect_template.b64、tdc_worker.js、analyze_collect_layout.py
- 「补环境」时代浏览器 JS: chrome_fetch/manual_captcha/probe_captcha/solve_in_browser/tdc_browser
- 旧 VM 探索工具 10 个(tdc_collectgen/tdc_slotmap/tdc_lift/tdc_vm_run/tdc_query/tdc_probe 等)
+ 8 个基于它们的旧 verify 脚本
- 训练/标注/评估: train_classifier/train_detector/prepare_yolo/web_labeler/collector/
siamese/evaluate/benchmark_verify
- tcaptcha/core_js/（60 个分析产物）、TOOLCHAIN_REPORT.md（整篇在描述已删工具）
- 3 个已无任何引用的 PyTorch 权重：icon_classifier.pt 43M / yolo_detector.pt 5.9M /
siamese_matcher.pt 1.6M —— 线上只跑 onnx，tcaptcha/ 117M -> 66M
为什么:
- 项目后面要专门承载「滑块 / 文字点选 / 图标点选」三种求解器；扁平包 + 运行时依赖混在
tools/ + 两条已废弃的 collect 生成路线，会让后续扩展持续踩坑
怎么验证:
- .venv/bin/python -m tcaptcha.selftest 全绿
- scripts/verify/verify_offline.py 逐槽全等 + 分段几何一致
- scripts/verify/verify_values.py 4 份真机明文 152/152 逐槽全等
- scripts/verify/send_test.py --mode python_values 1ms 签到页文字点选 errorCode=0
- 回滚点: tag backup/pre-refactor-20261012
94 lines
3.3 KiB
Python
94 lines
3.3 KiB
Python
# -*- coding: utf-8 -*-
|
||
"""把「同会话 collect 明文」与「同会话 tdc.js 的语义名」按位置精确对齐。
|
||
|
||
这是"不手工配平"的正解，依据都是已确证的链路：
|
||
* 槽位序 == 清单模块的 require 序（账本 #34/#35，运行时反证过）
|
||
* 语义名 == tdc_modules.analyze(同一份 tdc.js)（账本 #44，37/37 唯一）
|
||
* collect 4 段加密（账本 #41），解密后按空格 pad 切回 cd
|
||
|
||
输入（全部来自**同一次会话**，无需任何人工假设）:
|
||
/tmp/1ms_bridge/session_build.js 运行时 tdc.js（Debugger 取的）
|
||
/tmp/1ms_bridge/session_collect.json 同会话 TDC.getData(true) + getInfo
|
||
|
||
用法:
|
||
.venv/bin/python scripts/verify/align_session.py
|
||
"""
|
||
import base64
|
||
import json
|
||
import re
|
||
import sys
|
||
from pathlib import Path
|
||
from urllib.parse import unquote
|
||
|
||
ROOT = Path(__file__).resolve().parents[2]
|
||
sys.path.insert(0, str(ROOT))
|
||
sys.path.insert(0, str(ROOT / "tools"))
|
||
from tcaptcha.telemetry import modules as tm # noqa: E402
|
||
from tcaptcha.telemetry.crypto import split_collect_segments, xtea_decrypt_bytes # noqa: E402
|
||
from tcaptcha.telemetry.key import extract_xtea_key_pure # noqa: E402
|
||
|
||
BUILD = "/tmp/1ms_bridge/session_build.js"
|
||
SESS = "/tmp/1ms_bridge/session_collect.json"
|
||
|
||
js = open(BUILD, encoding="utf-8", errors="replace").read()
|
||
print("运行时 tdc.js: %d 字节" % len(js))
|
||
key = extract_xtea_key_pure(js)["key"]
|
||
print("本会话密钥 :", key)
|
||
print()
|
||
|
||
sess = json.load(open(SESS, encoding="utf-8"))
|
||
if "collect" not in sess and isinstance(sess.get("data"), dict):
|
||
sess = sess["data"] # cdp_collect.js 落盘的形状是 {target, ok, data:{...}}
|
||
collect = unquote(sess["collect"]) # 浏览器侧返回的是 URL-encoded
|
||
print("collect: %d 字符" % len(collect))
|
||
|
||
# 解密（4 段拼接：整体解一次即可，因为各段独立 XTEA-ECB）
|
||
raw = base64.b64decode(collect)
|
||
dec = xtea_decrypt_bytes(raw, key).decode("latin1")
|
||
plain = dec.rstrip("\x00")
|
||
print("解密 %d 字节" % len(dec))
|
||
runs = [m.span() for m in re.finditer(r" {2,}", plain)]
|
||
print("空格 pad 串:", runs)
|
||
print()
|
||
if len(runs) == 3:
|
||
c1, traj, c2, sd = split_collect_segments(plain)
|
||
print("段: chunk1=%d traj=%d chunk2=%d sd=%d" % (len(c1), len(traj), len(c2), len(sd)))
|
||
else:
|
||
c1 = traj = c2 = sd = None
|
||
print()
|
||
|
||
logical = re.sub(r" {2,}", "", plain)
|
||
obj = json.loads(logical)
|
||
cd = obj["cd"]
|
||
print("cd 槽数:", len(cd))
|
||
print()
|
||
|
||
rows = tm.analyze(BUILD)
|
||
names = [r["name"] for r in rows]
|
||
mods = [r["module"] for r in rows]
|
||
print("清单项 %d, 语义名 %d" % (len(mods), len(names)))
|
||
print()
|
||
|
||
# itoken 占 2 槽（^\\d+:\\d+$），其余 1 槽
|
||
multi = [i for i, n in enumerate(names) if n.startswith("itoken")]
|
||
slots = 0
|
||
print("%-2s %-6s %-34s %s" % ("#", "模块", "语义名", "值"))
|
||
print("-" * 110)
|
||
out = {}
|
||
k = 0
|
||
for i, (m, nm) in enumerate(zip(mods, names)):
|
||
n = 2 if i in multi else 1
|
||
if k + n > len(cd):
|
||
break
|
||
vals = cd[k:k + n]
|
||
k += n
|
||
val = vals[0] if n == 1 else vals
|
||
out[nm] = val
|
||
print("%-2d %-6d %-34s %s" % (i, m, nm, json.dumps(val, ensure_ascii=False)[:60]))
|
||
print()
|
||
print("覆盖槽数 %d / cd %d" % (k, len(cd)))
|
||
print()
|
||
Path("/tmp/1ms_bridge/name_values.json").write_text(
|
||
json.dumps(out, ensure_ascii=False, indent=1), encoding="utf-8")
|
||
print(">>> 语义名 -> 真值 已存 /tmp/1ms_bridge/name_values.json")
|