docs: 全面更新文档体系 — AI指南 + 人可读说明书

CLAUDE.md — AI 工作指南:
- 项目全景:目录结构/技术栈/端口
- 系统架构:前端路由/后端模块/认证流程
- 权限系统:三层角色/26项权限/守卫流水线/解析链路
- 考核系统:数据模型/出题算法/模板配置
- 测试脚本:7个Playwright测试说明
- 开发指南:启动/测试/重启/数据库管理
- 代码规范:TypeScript模式/权限装饰器/React约定
- Playwright测试技巧:React受控输入框/等待策略

README.md — 人可读英文说明书:
- 系统介绍 + 功能特性
- 完整使用指南(用户管理/权限管理/考核模板/组织考试)
- 核心流程说明(认证/出题/权限解析)
- 测试命令参考
- 项目结构 + 配置参考

README_ZH.md — 人可读中文说明书:
- 全面中文版本,包含所有新功能
- 步骤式操作指南,便于管理员使用

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
Developer
2026-06-09 13:19:45 +08:00
parent 1aee7e0baf
commit 65ede9fcff
3 changed files with 845 additions and 431 deletions
+364 -191
View File
@@ -1,225 +1,398 @@
# CLAUDE.md
# AuraK — AI 工作指南
This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
> 本文件指导 AI 助手(Claude Code 等)如何理解和使用此仓库。
> 项目:AuraK — 企业级多租户 AI 知识库与人才评价平台
## Project Overview
---
Simple Knowledge Base is a full-stack RAG (Retrieval-Augmented Generation) Q&A system built with React 19 + NestJS. It's a monorepo with Japanese/Chinese documentation but English code.
## 一、项目全景
**Key Features:**
- Multi-model support (OpenAI-compatible APIs + Google Gemini native SDK)
- Dual processing modes: Fast (Tika text-only) and High-precision (Vision pipeline)
- User isolation with JWT authentication and per-user knowledge bases
- Hybrid search (vector + keyword) with Elasticsearch
- Multi-language interface (Japanese, Chinese, English)
- Streaming responses via Server-Sent Events (SSE)
```
AuraK/
├── web/ # React 19 + Vite 前端 (端口 13001)
│ ├── components/ # UI 组件(含 views/ 目录下的主要页面)
│ ├── contexts/ # AuthContext / LanguageContext / ConfirmContext
│ ├── services/ # API 客户端(apiClient.ts + 各服务)
│ ├── hooks/ # usePermissions 等自定义 Hook
│ └── src/pages/ # 路由页面(薄壳,委托给 components/views
├── server/ # NestJS 后端 (端口 3001)
│ ├── src/auth/ # 认证(JWT/API Key+ 权限(RBAC
│ │ └── permission/ # Permission/Role 实体 + Service + Guard
│ ├── src/assessment/ # 考核评估子系统(最复杂)
│ ├── src/tenant/ # 多租户 + 租户成员管理
│ └── src/user/ # 用户 CRUD
├── docs/plans/ # 设计文档和计划
├── test-*.mjs # Playwright 测试脚本(项目根目录)
└── docker-compose.yml # 基础设施(Elasticsearch, Tika, LibreOffice
```
## Development Setup
### 核心技术栈
### Prerequisites
- Node.js 18+
- Yarn
- Docker & Docker Compose
| 层 | 技术 |
|---|---|
| 前端 | React 19, TypeScript, Vite 6, Tailwind CSS v4, Framer Motion, Lucide icons |
| 后端 | NestJS 11, TypeORM, SQLite (better-sqlite3), LangChain/LangGraph |
| 搜索 | Elasticsearch 9(向量+全文混合检索)|
| 认证 | JWT + API Key 双机制 |
| 部署 | Docker ComposeNginx 反向代理)|
---
## 二、系统架构
### 2.1 前端路由
| 路径 | 组件 | 说明 |
|---|---|---|
| `/login` | LoginPage | 公开 |
| `/` | OverviewPage | 工作台首页 |
| `/chat` | ChatPage | AI 对话 |
| `/assessment` | AssessmentPage | 考核评估 |
| `/assessment-stats` | AssessmentStatsView | 评估统计 |
| `/question-banks` | QuestionBankView | 题库管理 |
| `/settings` | SettingsPage | 系统设置 |
### 2.2 后端模块
app.module.ts 注册了 27 个模块,核心模块:
- **AuthModule** — 认证(JWT/API Key 双策略)
- **PermissionModule** — RBAC 权限(Role/RolePermission 实体 + PermissionService + PermissionsGuard
- **UserModule** — 用户 CRUD
- **TenantModule** — 多租户(全局模块)
- **AssessmentModule** — 考核评估(最复杂,含会话/出题/评分/证书)
- **AdminModule / SuperAdminModule** — 管理端 API
- **KnowledgeBaseModule** — 知识库索引和检索
- **ModelConfigModule** — AI 模型配置
### 2.3 认证流程
```
密码登录:
POST /api/auth/login (username, password)
→ LocalAuthGuard → JwtService.sign(payload) → { access_token, user }
获取 API Key:
GET /api/users/api-key (Authorization: Bearer <JWT>)
→ 返回 kb_xxxxxxxx...(存前端 localStorage
后续所有请求:
Headers: { x-api-key: <apiKey>, x-tenant-id: <tenantId> }
全局守卫: CombinedAuthGuardAPP_GUARD 注册)
→ 先尝试 API Key 认证,失败则回退 JWT
→ @Public() 装饰器跳过认证
```
---
## 三、权限系统(RBAC)核心
### 3.1 三层角色
| 角色 | 权限数 | 能做什么 |
|---|---|---|
| **SUPER_ADMIN** | 26 项 | 全部权限:用户/租户/知识库/考核/模型/插件/设置 |
| **TENANT_ADMIN** | 21 项 | 管理本租户用户和资源;不能跨租户/删用户/改系统设置 |
| **USER** | 5 项 | kb:view/create/edit, assess:view, plugin:view |
### 3.2 权限常量定义
位于 `server/src/auth/permission/permission.constants.ts`
```typescript
// 用户管理: user:view / user:create / user:edit / user:delete / user:role / user:password
// 租户管理: tenant:view / tenant:create / tenant:edit / tenant:delete / tenant:members
// 知识库: kb:view / kb:create / kb:edit / kb:delete / kb:publish
// 考核: assess:view / assess:manage / assess:template / assess:bank
// 模型: model:view / model:config
// 插件: plugin:view / plugin:manage
// 设置: settings:view / settings:system
```
### 3.3 权限守卫流水线
```
CombinedAuthGuard (认证) → 已有,全局注册
→ RolesGuard (角色门控) → @Roles(SUPER_ADMIN, TENANT_ADMIN)
→ PermissionsGuard (权限门控) → @Permission('user:view')
```
**系统角色保护**: 系统角色(isSystem=true)的权限不可通过 API 修改。
### 3.4 权限解析链路
```
PermissionService.getUserPermissions(userId, tenantId)
→ 1. 如果 user.isAdmin === true,返回全部 SUPER_ADMIN 权限
→ 2. 查找 TenantMember(userId, tenantId)
→ 3. TenantMember.role → Role.baseRole → RolePermission.permissionKey
```
---
## 四、考核评估系统
### 4.1 数据模型
```
AssessmentTemplate (模板)
└── question_count, dimensions[{name, label, weight}], passingScore
└── QuestionBank (题库,一对一关联模板)
└── QuestionBankItem (题目)
├── questionType: SHORT_ANSWER / MULTIPLE_CHOICE / TRUE_FALSE
└── dimension: PROMPT / LLM / IDE / DEV_PATTERN / WORK_CAPABILITY
AssessmentSession (会话)
├── questions_json (缓存的题目列表)
├── messages (对话历史)
├── scores / dimensionScores
└── finalReport (AI 生成的报告)
AssessmentCertificate (证书)
└── level: Novice / Proficient / Advanced / Expert
```
### 4.2 出题算法
`selectQuestions()``question-bank.service.ts`:
```typescript
// 按维度权重公平分配(floor + remainder,保证总和 = count
// 20题模板: PROMPT(30%)→6题, LLM(30%)→6题, IDE(20%)→4题, DEV_PATTERN(20%)→4题
```
### 4.3 模板配置(数据库 `assessment_templates` 表)
```typescript
// dimensions 字段是 JSON 数组,前端「测评模板」页面可编辑
dimensions = [
{ name: "PROMPT", label: "Prompt Engineering", weight: 30 },
{ name: "LLM", label: "LLM Knowledge", weight: 30 },
{ name: "IDE", label: "IDE Proficiency", weight: 20 },
{ name: "DEV_PATTERN", label: "Dev Patterns", weight: 20 },
]
```
---
## 五、测试脚本
所有 Playwright 测试位于项目根目录:
| 脚本 | 用途 | 运行命令 |
|---|---|---|
| `test-systematic.mjs` | **142 项系统性测试**(身份/CRUD/权限/边界/UI | `node test-systematic.mjs` |
| `test-e2e-full.mjs` | 94 项全角色 E2E 测试 | `node test-e2e-full.mjs` |
| `test-user-lifecycle.mjs` | 42 项用户生命周期测试 | `node test-user-lifecycle.mjs` |
| `test-multiround.mjs` | 考核多轮对话测试 | `node test-multiround.mjs` |
| `test-permission-flow.mjs` | 三层角色权限验证 | `node test-permission-flow.mjs` |
| `test-question-distribution.mjs` | 出题分布验证 | `node test-question-distribution.mjs` |
| `exam-organizer.mjs` | 考试组织者场景(创建考生+考核+查看结果) | `node exam-organizer.mjs` |
---
## 六、开发指南
### 6.1 启动项目
### Initial Setup
```bash
# Install dependencies
yarn install
# 后端
cd server && node dist/main.js # 生产模式(需要先 nest build)
cd server && yarn start:dev # 开发模式(热加载)
# Start infrastructure services
# 前端
cd web && npx vite --port 13001 # 开发模式
cd web && npx vite build # 编译
# 基础设施(需要 Docker
docker-compose up -d elasticsearch tika libreoffice
# Configure environment
cp server/.env.sample server/.env
# Edit server/.env with API keys and configuration
```
### Development Commands
### 6.2 运行测试
```bash
# Start both frontend and backend in development mode
yarn dev
# Frontend only (port 13001)
cd web && yarn dev
# Backend only (port 3001)
cd server && yarn start:dev
# Run tests
cd server && yarn test
cd server && yarn test:e2e
# Lint and format
cd server && yarn lint
cd server && yarn format
cd /d/AuraK && node test-systematic.mjs # 142 项全面测试
```
### Docker Services
- **Elasticsearch**: 9200 (vector storage)
- **Apache Tika**: 9998 (document text extraction)
- **LibreOffice Server**: 8100 (document conversion)
- **Backend API**: 3001
- **Frontend**: 13001 (dev), 80/443 (production via nginx)
### 6.3 重启服务
## Architecture
### Project Structure
```
simple-kb/
├── web/ # React frontend (Vite)
│ ├── components/ # UI components (ChatInterface, ConfigPanel, etc.)
│ ├── contexts/ # React Context providers
│ ├── services/ # API client services
│ └── utils/ # Utility functions
├── server/ # NestJS backend
│ ├── src/
│ │ ├── ai/ # AI services (embedding, etc.)
│ │ ├── api/ # API module
│ │ ├── auth/ # JWT authentication
│ │ ├── chat/ # Chat/RAG module
│ │ ├── elasticsearch/ # Elasticsearch integration
│ │ ├── import-task/ # Import task management
│ │ ├── knowledge-base/# Knowledge base management
│ │ ├── libreoffice/ # LibreOffice integration
│ │ ├── model-config/ # Model configuration management
│ │ ├── vision/ # Vision model integration
│ │ └── vision-pipeline/# Vision pipeline orchestration
│ ├── data/ # SQLite database storage
│ ├── uploads/ # Uploaded files storage
│ └── temp/ # Temporary files
├── docs/ # Comprehensive documentation (Japanese/Chinese)
├── nginx/ # Nginx configuration
├── libreoffice-server/ # LibreOffice conversion service (Python/FastAPI)
└── docker-compose.yml # Docker orchestration
```
### Key Architectural Concepts
**Dual Processing Modes:**
1. **Fast Mode**: Apache Tika for text-only extraction (quick, no API cost)
2. **High-Precision Mode**: Vision Pipeline (LibreOffice → PDF → Images → Vision Model) for mixed image/text documents (slower, incurs API costs)
**Multi-Model Support:**
- OpenAI-compatible APIs (OpenAI, DeepSeek, Claude, etc.)
- Google Gemini native SDK
- Configurable LLM, Embedding, and Rerank models
**RAG System:**
- Hybrid search (vector + keyword) with Elasticsearch
- Streaming responses via Server-Sent Events (SSE)
- Source citation and similarity scoring
- Chunk configuration (size, overlap)
## Code Standards
### Language Requirements
- **Code comments must be in English**
- **Log messages must be in English**
- **Error messages must support internationalization** to enable multi-language frontend interface
- **API response messages must support internationalization** to enable multi-language frontend interface
- Interface supports Japanese, Chinese, and English
### Testing
- Backend uses Jest for unit and e2e tests
- Frontend currently has no test framework configured
- Run tests: `cd server && yarn test` or `yarn test:e2e`
### Code Quality
- ESLint and Prettier configured for backend
- Format code: `cd server && yarn format`
- Lint code: `cd server && yarn lint`
## Common Development Tasks
### Adding a New API Endpoint
1. Create controller in appropriate module under `server/src/`
2. Add service methods with English comments
3. Update DTOs and validation
4. Add tests in `*.spec.ts` files
### Adding a New Frontend Component
1. Create component in `web/components/`
2. Add TypeScript interfaces in `web/types.ts`
3. Use Tailwind CSS for styling
4. Connect to backend services in `web/services/`
### Debugging
- Backend logs are in Chinese
- Check Elasticsearch: `curl http://localhost:9200/_cat/indices`
- Check Tika: `curl http://localhost:9998/tika`
- Check LibreOffice: `curl http://localhost:8100/health`
## Environment Configuration
Key environment variables (`server/.env`):
- `OPENAI_API_KEY`: OpenAI-compatible API key
- `GEMINI_API_KEY`: Google Gemini API key
- `ELASTICSEARCH_HOST`: Elasticsearch URL (default: http://localhost:9200)
- `TIKA_HOST`: Apache Tika URL (default: http://localhost:9998)
- `LIBREOFFICE_URL`: LibreOffice server URL (default: http://localhost:8100)
- `JWT_SECRET`: JWT signing secret
## Deployment
### Development
```bash
docker-compose up -d elasticsearch tika libreoffice
yarn dev
taskkill //F //IM node.exe 2>/dev/null # 杀所有 Node 进程
cd /d/AuraK/server && node dist/main.js & # 启动后端
cd /d/AuraK/web && npx vite --port 13001 & # 启动前端
```
### Production
### 6.4 数据库管理
```bash
docker-compose up -d # Builds and starts all services
# SQLite 文件路径
server/data/metadata.db
# 直接查询
cd /d/AuraK && node -e "
const sqlite3 = require('better-sqlite3');
const db = new sqlite3('server/data/metadata.db');
db.prepare('SELECT * FROM users').all();
db.close();
"
```
### Ports in Production
- Frontend: 80/443 (via nginx)
- Backend API: 3001 (proxied through nginx)
- Elasticsearch: 9200
- Tika: 9998
- LibreOffice: 8100
### 6.5 常见问题
## AI 工作流指令
- **端口占用**: `netstat -ano | grep 3001``taskkill //F //PID <pid>`
- **后端启动失败**: 确保从 `server/` 目录启动以加载 .env
- **Windows 杀进程**: 用 `taskkill //F //PID``kill` 不可靠
- **Dist 目录修改**: `server/dist/` 下的 .js 直接修改即可生效(无需 nest build)
本项目已安装 **gstack**54 个技能)和 **Superpowers**(14 个技能)。请按以下规则协调使用:
---
### 自动触发规则
当用户的意图匹配以下场景时,**自动调用对应的 gstack skill**(使用 Skill 工具),并在调用前向用户说明启动了哪个技能:
## 七、代码规范
- 用户讨论需求、想法、产品方向 → 调用 `office-hours` skill,说"正在启动 **office-hours**(产品策略顾问)..."
- 用户讨论功能范围、优先级 → 调用 `plan-ceo-review` skill,说"正在启动 **plan-ceo-review**(战略评审)..."
- 用户讨论技术方案、架构设计 → 调用 `plan-eng-review` skill,说"正在启动 **plan-eng-review**(架构评审)..."
- 用户要求审查代码 → 调用 `review` skill,说"正在启动 **review**(代码审查)..."
- 用户要求测试/QA → 调用 `qa` skill,说"正在启动 **qa**(自动化测试)..."
- 用户要求安全审查 → 调用 `cso` skill,说"正在启动 **cso**(安全审计)..."
- 用户要求发布/发版 → 调用 `ship` skill,说"正在启动 **ship**(发布流程)..."
- 用户报告 bug 需要调试 → 调用 `investigate` skill,说"正在启动 **investigate**(系统化调试)..."
### 7.1 TypeScript 模式
### Superpowers 保留自动触发
Superpowers 的技能(brainstorming、test-driven-development、systematic-debugging 等)继续保持原有自动触发机制,不做干预。
- **实体定义**: TypeORM `@Entity` + `@Column` + 枚举类型
- **控制器**: `@Controller` + `@UseGuards(CombinedAuthGuard)` + `@Permission()`
- **服务**: `@Injectable` 类,方法名用动词开头
- **前端组件**: React FC + TypeScript interface props
### 通知机制
每次自动调用 gstack skill 时,必须明确告知用户正在启动哪个技能及其作用。
### 7.2 权限装饰器使用
## Troubleshooting
```typescript
// 后端 API — 在控制器方法上标注所需权限
@Post()
@UseGuards(PermissionsGuard)
@Permission('user:create')
async createUser(@Request() req, @Body() body: CreateUserDto) { ... }
### Common Issues
1. **Elasticsearch not starting**: Check memory limits in docker-compose.yml
2. **File upload failures**: Ensure `uploads/` and `temp/` directories exist with proper permissions
3. **Vision pipeline errors**: Verify LibreOffice server is running and accessible
4. **API key errors**: Check environment variables in `server/.env`
// 前端 — 用 PermissionGate 组件或 usePermissions hook
<PermissionGate permission="user:create">
<Button></Button>
</PermissionGate>
### Database Management
- SQLite database: `server/data/metadata.db`
- Elasticsearch indices: Managed automatically by the application
- To reset: Delete `server/data/metadata.db` and Elasticsearch data volume
const { hasPermission } = usePermissions();
{hasPermission('user:view') && <UserList />}
```
## Documentation
### 7.3 通用 React 约定
- **README.md**: Project overview in Japanese
- **docs/**: Comprehensive documentation (mostly Japanese/Chinese)
- **DESIGN.md**: System architecture and design
- **API.md**: API reference
- **DEVELOPMENT_STANDARDS.md**: Mandates English comments/logs and internationalized messages
- 使用 `cn()` 工具函数(tailwind-merge)合并 className
- 图标用 `lucide-react`
- 动画用 `framer-motion`
- API 调用用 `apiClient` 或原生 `fetch`
- 模态框用 `createPortal` + `AnimatePresence`
When modifying code, always add English comments and logs as required by development standards. Error and UI messages must be properly internationalized. The project has extensive existing documentation in Japanese/Chinese - refer to `docs/` directory for detailed technical information.
---
## 八、Playwright 测试指南
### 8.1 测试模式
```javascript
import { chromium } from 'playwright';
const browser = await chromium.launch({ headless: true });
const page = await browser.newPage({ viewport: { width: 1440, height: 900 } });
// 模拟 API 调用
const r = await fetch(`${API}/api/auth/login`, {
method: 'POST',
headers: { 'Content-Type': 'application/json' },
body: JSON.stringify({ username, password }),
});
```
### 8.2 React 受控输入框操作(关键)⚠️
React 的受控 input/textarea 依赖 native setter 触发 `onChange`。**不要用 `type()``fill()`**,应使用:
```javascript
await page.evaluate((text) => {
const ta = document.querySelector('textarea');
if (!ta) return;
const setter = Object.getOwnPropertyDescriptor(HTMLTextAreaElement.prototype, 'value')?.set;
setter?.call(ta, text);
ta.dispatchEvent(new Event('input', { bubbles: true }));
}, '要输入的文字');
```
### 8.3 等待按钮可用
```javascript
await page.waitForFunction(() => {
const btn = document.querySelector('button:has(svg.lucide-send)');
return btn && !btn.disabled;
}, { timeout: 10000 });
```
### 8.4 等待 spinner 消失
```javascript
await page.waitForFunction(() => !document.querySelector('.animate-spin'), { timeout: 60000 });
```
### 8.5 检测考核题型
```javascript
const state = await page.evaluate(() => {
// 选择题选项按钮
const optionBtns = Array.from(document.querySelectorAll('button.w-full.text-left.px-5.py-4'))
.filter(b => !b.textContent?.includes('确认答案'));
// 简答题 textarea
const ta = document.querySelector('textarea');
return { choiceCount: optionBtns.length, hasTextarea: ta?.offsetParent !== null };
});
```
---
## 九、部署配置
### 端口
| 服务 | 端口 |
|---|---|
| 前端(开发) | 13001 |
| 前端(生产/Nginx | 80/443 |
| 后端 API | 3001 |
| Elasticsearch | 9200 |
| Apache Tika | 9998 |
| LibreOffice | 8100 |
### 环境变量
`server/.env` 关键配置:
```
PORT=3001
DATABASE_PATH=./data/metadata.db
ELASTICSEARCH_HOST=http://127.0.0.1:9200
JWT_SECRET=your-secret
```
---
## 十、快速参考命令
```bash
# 启动
cd /d/AuraK/server && node dist/main.js &
cd /d/AuraK/web && npx vite --port 13001 &
# 后端编译
cd /d/AuraK/server && npx nest build
# 运行测试
cd /d/AuraK && node test-systematic.mjs
# 数据库查询
cd /d/AuraK && node -e "
const s=require('better-sqlite3');
const d=new s('server/data/metadata.db');
const r=d.prepare('...').all(); console.log(r); d.close();"
# 清理端口
taskkill //F //IM node.exe 2>/dev/null
# 前端编译
cd /d/AuraK/web && npx vite build
```