feat(v1): add secure configuration baseline

This commit is contained in:
ila
2026-07-21 09:45:59 +08:00
parent 7b05efe56c
commit 0b01e1bf24
13 changed files with 281 additions and 19 deletions
+17
View File
@@ -0,0 +1,17 @@
# Local configuration and credentials
v1/config.local.json
*.local.json
.env
.env.*
# Runtime artifacts and private evaluation materials
artifacts/
testdata/private/
testdata/videos/
*.onnx
*.pt
# Python tooling
.pytest_cache/
__pycache__/
*.py[cod]
+2 -1
View File
@@ -42,7 +42,8 @@
| 基线检查 | `./init.ps1` |
| 旧代码编译检查 | `python -m compileall -q demo` |
| 旧 GUI 启动 | `Set-Location demo; python main.py` |
| V1 测试 | T-101 创建 `v1/requirements.txt` 和测试后写入真实 `pytest` 命令 |
| V1 依赖安装 | `python -m pip install -r v1/requirements.txt`(显式执行;`init.ps1` 不自动安装) |
| V1 单元测试 | `python -m pytest v1/tests -v` |
| V2 测试 | T-302 创建 Go 模块后使用 `go test ./...` |
## 演进决策
+1 -1
View File
@@ -14,7 +14,7 @@
| ID | 任务 | 依赖 | 验收要点 | 状态 |
| --- | --- | --- | --- | --- |
| T-101 | 创建 V1 包、依赖清单、示例配置和忽略规则 | T-000 | `v1/` 可导入;真实 RTSP 凭证被拒绝提交;`pytest` 能运行。 | DOING |
| T-101 | 创建 V1 包、依赖清单、示例配置和忽略规则 | T-000 | `v1/` 可导入;真实 RTSP 凭证被拒绝提交;`pytest` 能运行。 | DONE |
| T-102 | 实现可重连的视频源与录像回放适配器 | T-101 | 有效本地录像可按时间戳产帧;无效源进入连接错误状态且不崩溃。 | TODO |
| T-103 | 实现 Pose 适配器与模型来源校验 | T-102 | 输出 person box、17 点和置信度;错误模型或哈希不符时给出明确错误。 | TODO |
| T-104 | 实现人员跟踪与姿态质量门控 | T-103 | 连续人员维持 ID;低质量、缺失膝踝或空帧不会产生倒地候选。 | TODO |
+6 -6
View File
@@ -36,7 +36,7 @@
- Modify: `.gitignore`
- Modify: `init.ps1`, `docs/03-tech-stack.md`, `docs/current-state.md`
- [ ] **Step 1: Write the failing configuration test**
- [x] **Step 1: Write the failing configuration test**
```python
from v1.config import load_config
@@ -55,12 +55,12 @@ def test_load_config_resolves_rtsp_environment_variable(tmp_path, monkeypatch):
assert load_config(config_file).source_url == "rtsp://example"
```
- [ ] **Step 2: Run the test to verify it fails**
- [x] **Step 2: Run the test to verify it fails**
Run: `python -m pytest v1/tests/test_config.py -v`
Expected: FAIL because `v1.config` does not exist.
- [ ] **Step 3: Implement the smallest configuration interface**
- [x] **Step 3: Implement the smallest configuration interface**
```python
import json
@@ -114,12 +114,12 @@ def load_config(path: Path) -> AppConfig:
`config.local.json`, `artifacts/`, `testdata/private/`, `*.onnx` and camera credentials must be ignored; `config.example.json` contains only `rtsp_url_env`.
- [ ] **Step 4: Run baseline tests and update the standard command**
- [x] **Step 4: Run baseline tests and update the standard command**
Run: `python -m pytest v1/tests -v`
Expected: PASS. Update `init.ps1` so it installs `v1/requirements.txt` and runs this command.
Expected: PASS. Update `init.ps1` so it checks the required runtime imports and runs this command; keep `v1/requirements.txt` as an explicit installation command rather than mutating the environment during verification.
- [ ] **Step 5: Commit**
- [x] **Step 5: Commit**
```powershell
git add .gitignore init.ps1 v1 docs/03-tech-stack.md docs/current-state.md docs/06-tasks.md progress.md
+2 -2
View File
@@ -15,7 +15,7 @@
},
"model": {
"path": "models/best.pt",
"sha256": "填写受控模型哈希",
"sha256": "0000000000000000000000000000000000000000000000000000000000000000",
"confidence_threshold": 0.25
},
"event": {
@@ -32,7 +32,7 @@
```
- `rtsp_url_env` 必填;应用从同名环境变量读取真实 URL。
- 数值是待现场录像校准的默认值;每个值必须真正进入事件逻辑。
- 数值是待现场录像校准的默认值;每个值必须真正进入事件逻辑。示例中的全零 SHA-256 只占位配置形状,T-103 必须以受控模型的真实哈希替换并验证后才能启动推理。
- 缺少环境变量、模型不存在或哈希不符时,应用显示配置错误,不启动监控。
## 核心数据
+8 -8
View File
@@ -4,18 +4,18 @@
## 当前快照
- 日期:2026-07-20
- 阶段:V1 工程化起步;T-101 进行中。
- 日期:2026-07-21
- 阶段:V1 工程化起步;T-101 已验收,等待 T-102。
- 已验证环境:Windows PowerShell;Python 3.8.10;Ultralytics 8.3.205;PyQt5 可导入。
- 旧生产基线:`demo/main.py`、`demo/fall_detection_gui.py`、`demo/detect_fall.py`、`demo/best.pt`。
- V1 代码:`v1/` 目录存在但尚无实现。
- V1 代码:已建立 `v1/__init__.py`、`v1/config.py`、安全的 `config.example.json`、固定依赖清单和配置单元测试;视频源、Pose、跟踪、证据、状态机和 GUI 尚未实现。
- V2 代码:`v2/` 目录存在但尚无实现。
- 非代码设计工件:docs/ui/silver-pose-ui-ux-spec.md、docs/ui/2026-07-20-html-prototype-plan.md、docs/ui/silver-pose-v1-prototype.html 与 docs/ui/silver-pose-v2-prototype.html 已建立。v2 HTML 是符合正式浅色 Windows 规范的当前视觉参考:浅灰蓝底、白色卡片,红色只表示确认摔倒、其弹窗和事件证据;文件名中的 v2 只表示原型设计修订,不能理解为 Go V2 实现已开始。v1 HTML 保留为历史深色对照。两者均使用顶部双 Tab、设置草稿与状态交互,且画面、事件和时间都是模拟数据,不连接真实摄像头、模型或网络,也不改变 Phase 1 任务顺序。
- 测试:`python -m compileall -q demo` 已通过;T-101 开始前 `v1/tests` 尚不存在,因此 `python -m pytest v1/tests -v` 以目录未找到退出,这是待建立的初始测试基线。
- 测试:`python -m compileall -q demo` 已通过;`python -m pytest v1/tests -v` 当前有 2 个配置测试并已通过。`init.ps1` 会检查运行时依赖、编译旧基线并运行 V1 测试,但不会安装软件包。
- 模型:`demo/best.pt` 可加载为 YOLO Pose,类别 `person`,`kpt_shape=[17, 3]`;与 `D:\PythonP\fall_detection\best.pt` 哈希一致。
- 当前标准启动:`./init.ps1`。
- 当前标准验证:`python -m compileall -q demo`。
- 当前 blocker:V1 依赖清单、示例配置、回归录像及事件标签尚未创建;真实海康 RTSP 流尚未接入。
- 当前 blocker:回归录像及事件标签尚未创建;真实海康 RTSP 流尚未接入。
全局环境的 `pip check` 存在其他项目的包冲突,因此它不是 Silver Pose 的验收命令。`init.ps1` 只检查本项目实际导入的 OpenCV、NumPy、Ultralytics 与 PyQt5,并在命令非零退出时失败。
@@ -32,9 +32,9 @@
## 任务状态
- 已完成:T-000(Harness 文档与旧基线快照)。
- 正在进行:T-101(创建 V1 包、依赖清单、示例配置和忽略规则)。
- 下一个可领取:完成 T-101 后为 T-102。
- 已完成:T-000(Harness 文档与旧基线快照)、T-101(V1 安全配置基线)。
- 正在进行:无。
- 下一个可领取:T-102。
## 当前可运行内容
+3 -1
View File
@@ -1,13 +1,14 @@
#!/usr/bin/env pwsh
# Silver Pose 的 Windows 统一启动与验证入口。
# 当前阶段验证 demo 基线;T-101 创建 v1 后,应同步更新本文件和 docs/03-tech-stack.md。
# V1 的 requirements.txt 是显式安装清单;本脚本只检查已安装依赖,避免修改当前环境。
$ErrorActionPreference = "Stop"
Set-Location -Path $PSScriptRoot
$InstallCmd = 'python -c "import cv2, numpy, ultralytics, PyQt5; print(''required Python packages available'')"'
$VerifyCmd = "python -m compileall -q demo"
$V1TestCmd = "python -m pytest v1/tests -v"
$StartCmd = "Set-Location demo; python main.py"
function Invoke-Checked {
@@ -22,6 +23,7 @@ function Invoke-Checked {
Write-Host "==> 当前目录: $($PWD.Path)"
Invoke-Checked -Label "检查 Silver Pose Python 依赖" -Command $InstallCmd
Invoke-Checked -Label "运行 demo 基线验证" -Command $VerifyCmd
Invoke-Checked -Label "运行 V1 单元测试" -Command $V1TestCmd
Write-Host "==> 图形界面启动命令"
Write-Host " $StartCmd"
+9
View File
@@ -80,3 +80,12 @@
- 阻塞:无。
- 决策:`init.ps1` 保持无副作用的依赖检查与基线验证,不自动安装包;`requirements.txt` 将作为显式、可复现的安装清单。
- 下一步:先写配置加载与敏感值拒绝的失败测试,再实现最小配置接口。
## 【2026-07-21】T-101 创建 V1 包、依赖清单、示例配置和忽略规则(完成)
- 状态:DONE
- 变更:新增 `v1` 包、经验证的 `config.py`、仅含环境变量名的公开配置示例、与当前环境一致的 `requirements.txt` 和两项 pytest 配置测试;新增 `.gitignore`,忽略本地配置、事件工件、私有录像和模型导出物。`init.ps1` 现在运行 V1 测试但不自动安装依赖。
- 验证:先运行 `python -m pytest v1/tests/test_config.py -v`,确认因缺少 `v1.config` 产生预期导入失败;实现后 `./init.ps1` 和 `python -m pytest v1/tests -v` 均通过(2 passed),`python -m compileall -q v1 demo` 通过。公开配置的地址/凭证扫描无匹配,`git check-ignore` 已确认本地配置、事件工件、`.pt` 与 `.onnx` 工件被忽略。
- 阻塞:无。
- 决策:配置只允许 `rtsp_url_env` 指向运行环境变量,拒绝内嵌来源地址;确认窗口限制为 1–3 秒。公开示例使用 64 位全零哈希占位,T-103 必须替换为受控模型真实哈希后才可加载模型。
- 下一步:T-102,先建立录像回放和错误来源的失败测试。
+5
View File
@@ -0,0 +1,5 @@
"""Silver Pose Python V1 package.
V1 keeps configuration, video acquisition, pose inference, tracking, evidence,
and temporal event decisions in separate modules.
"""
+21
View File
@@ -0,0 +1,21 @@
{
"source": {
"id": "lobby-camera-01",
"rtsp_url_env": "SILVER_POSE_RTSP_URL"
},
"model": {
"path": "models/best.pt",
"sha256": "0000000000000000000000000000000000000000000000000000000000000000",
"confidence_threshold": 0.25
},
"event": {
"keypoint_confidence_threshold": 0.4,
"suspect_window_seconds": 0.5,
"confirm_window_seconds": 1.8,
"recovery_window_seconds": 2.0,
"cooldown_seconds": 10.0
},
"artifacts": {
"event_dir": "../artifacts/events"
}
}
+146
View File
@@ -0,0 +1,146 @@
"""Validated, credential-safe configuration loading for Silver Pose V1."""
import json
import os
import re
from dataclasses import dataclass
from pathlib import Path
from typing import Any, Dict
class ConfigError(ValueError):
"""Raised when a configuration file cannot safely start V1."""
@dataclass(frozen=True)
class EventConfig:
keypoint_confidence_threshold: float
suspect_window_seconds: float
confirm_window_seconds: float
recovery_window_seconds: float
cooldown_seconds: float
@dataclass(frozen=True)
class AppConfig:
source_id: str
source_url: str
model_path: Path
model_sha256: str
confidence_threshold: float
event: EventConfig
event_dir: Path
_ENVIRONMENT_NAME = re.compile(r"^[A-Za-z_][A-Za-z0-9_]*$")
_SHA256 = re.compile(r"^[0-9a-fA-F]{64}$")
def _mapping(value: Any, field_name: str) -> Dict[str, Any]:
if not isinstance(value, dict):
raise ConfigError("{0} must be an object".format(field_name))
return value
def _text(value: Any, field_name: str) -> str:
if not isinstance(value, str) or not value.strip():
raise ConfigError("{0} must be a non-empty string".format(field_name))
return value.strip()
def _number(value: Any, field_name: str, minimum: float, maximum: float) -> float:
try:
result = float(value)
except (TypeError, ValueError):
raise ConfigError("{0} must be numeric".format(field_name))
if not minimum <= result <= maximum:
raise ConfigError(
"{0} must be between {1} and {2}".format(field_name, minimum, maximum)
)
return result
def _resolve_path(config_path: Path, value: Any, field_name: str) -> Path:
raw_path = Path(_text(value, field_name))
if raw_path.is_absolute():
return raw_path
return (config_path.parent / raw_path).resolve()
def load_config(path: Path) -> AppConfig:
"""Load one public/local configuration pair without persisting credentials.
The config file can name an environment variable but must not embed a source
address. The resolved address remains in memory only.
"""
config_path = Path(path).resolve()
try:
raw = json.loads(config_path.read_text(encoding="utf-8"))
except OSError as exc:
raise ConfigError("cannot read config: {0}".format(exc))
except json.JSONDecodeError as exc:
raise ConfigError("invalid JSON config: {0}".format(exc))
root = _mapping(raw, "config")
source = _mapping(root.get("source"), "source")
if "url" in source or "rtsp_url" in source:
raise ConfigError("source must define rtsp_url_env, not an embedded address")
environment_name = _text(source.get("rtsp_url_env"), "source.rtsp_url_env")
if not _ENVIRONMENT_NAME.match(environment_name):
raise ConfigError("source.rtsp_url_env must be an environment variable name")
source_url = os.environ.get(environment_name)
if not source_url:
raise ConfigError("missing RTSP environment variable: {0}".format(environment_name))
model = _mapping(root.get("model"), "model")
model_sha256 = _text(model.get("sha256"), "model.sha256").lower()
if not _SHA256.match(model_sha256):
raise ConfigError("model.sha256 must be a 64-character SHA-256 value")
event_raw = _mapping(root.get("event"), "event")
event = EventConfig(
keypoint_confidence_threshold=_number(
event_raw.get("keypoint_confidence_threshold"),
"event.keypoint_confidence_threshold",
0.0,
1.0,
),
suspect_window_seconds=_number(
event_raw.get("suspect_window_seconds"),
"event.suspect_window_seconds",
0.0,
30.0,
),
confirm_window_seconds=_number(
event_raw.get("confirm_window_seconds"),
"event.confirm_window_seconds",
1.0,
3.0,
),
recovery_window_seconds=_number(
event_raw.get("recovery_window_seconds"),
"event.recovery_window_seconds",
0.0,
300.0,
),
cooldown_seconds=_number(
event_raw.get("cooldown_seconds"),
"event.cooldown_seconds",
0.0,
3600.0,
),
)
artifacts = _mapping(root.get("artifacts"), "artifacts")
return AppConfig(
source_id=_text(source.get("id"), "source.id"),
source_url=source_url,
model_path=_resolve_path(config_path, model.get("path"), "model.path"),
model_sha256=model_sha256,
confidence_threshold=_number(
model.get("confidence_threshold"), "model.confidence_threshold", 0.0, 1.0
),
event=event,
event_dir=_resolve_path(config_path, artifacts.get("event_dir"), "artifacts.event_dir"),
)
+5
View File
@@ -0,0 +1,5 @@
numpy==1.24.4
opencv-python==4.6.0.66
ultralytics==8.3.205
PyQt5==5.15.9
pytest==8.3.5
+56
View File
@@ -0,0 +1,56 @@
import json
import pytest
from v1.config import ConfigError, load_config
def _write_config(path, source):
path.write_text(
json.dumps(
{
"source": source,
"model": {
"path": "models/best.pt",
"sha256": "a" * 64,
"confidence_threshold": 0.25,
},
"event": {
"keypoint_confidence_threshold": 0.4,
"suspect_window_seconds": 0.5,
"confirm_window_seconds": 1.8,
"recovery_window_seconds": 2.0,
"cooldown_seconds": 10.0,
},
"artifacts": {"event_dir": "artifacts/events"},
},
ensure_ascii=False,
),
encoding="utf-8",
)
def test_load_config_resolves_source_from_environment_variable(tmp_path, monkeypatch):
config_file = tmp_path / "config.json"
_write_config(
config_file,
{"id": "lobby-camera-01", "rtsp_url_env": "SILVER_POSE_RTSP_URL"},
)
monkeypatch.setenv("SILVER_POSE_RTSP_URL", "rtsp://demo.invalid/live")
config = load_config(config_file)
assert config.source_id == "lobby-camera-01"
assert config.source_url == "rtsp://demo.invalid/live"
assert config.event.confirm_window_seconds == 1.8
def test_load_config_rejects_embedded_source_address(tmp_path):
config_file = tmp_path / "config.json"
_write_config(
config_file,
{"id": "lobby-camera-01", "url": "rtsp://demo.invalid/live"},
)
with pytest.raises(ConfigError, match="rtsp_url_env"):
load_config(config_file)