# isy-rag-agent

**Repository Path**: deege/isy-rag-agent

## Basic Information

- **Project Name**: isy-rag-agent
- **Description**: 智能开源的企业级rag系统
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 11
- **Created**: 2026-06-14
- **Last Updated**: 2026-08-27

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# ISY-RAG-Agent

A Spring Boot-based RAG (Retrieval-Augmented Generation) intelligent chatbot management platform.

## Project Overview

ISY-RAG-Agent is a feature-rich intelligent chatbot platform integrating core RAG capabilities including knowledge base management, FAQ retrieval, tool execution, and web search. The system adopts a front-end/back-end separated architecture, with the back-end providing a complete RESTful API and supporting SSE streaming output, enabling seamless integration with various large language models.

## Core Features

### 🤖 Bot Management
- Create, configure, and manage multiple AI chatbots
- Customize bot avatars, welcome messages, and system prompts
- Configure response strategies, confidence thresholds, and context chunk counts
- Associate knowledge bases, FAQs, tools, and authorized users

### 📚 Knowledge Base Management
- Create and manage knowledge bases
- Support document upload (automatic chunking and vectorization)
- Integrated with Milvus vector database
- Chunk-level document management

### ❓ FAQ Management
- Categorized FAQ management
- Vector similarity matching
- Support for rich text responses
- Similar question configuration
- Click-through rate and satisfaction statistics

### 🔧 Tool Management
- Three tool types: MCP tools, API tools, and built-in functions
- MCP server integration (supports SSE and Streamable HTTP)
- HTTP API tool configuration
- Built-in functions: get time, mathematical calculations, string processing, JSON formatting, etc.

### 📊 Tracking & Monitoring
- Complete conversation tracking
- Input/output logging for each node
- Performance statistics (latency, tokens)
- Visualized tracking details

### 👥 User Management
- User registration and authentication
- JWT-based authentication
- Department management
- Daily call limits

## Technology Architecture

### Backend Stack
- **Spring Boot** - Application framework
- **MyBatis Plus** - ORM persistence layer
- **MySQL** - Relational database
- **Milvus** - Vector database
- **LangChain4j** - LLM integration
- **SSE** - Server-Sent Events streaming output

### Frontend Technology
- Native HTML/CSS/JavaScript
- No build tools required; runs directly

## Quick Start

### System Requirements
- JDK 17+
- MySQL 5.7+
- Milvus 2.0+ (optional, for vector retrieval functionality)

### Configuration Steps

1. **Create Database**
```sql
CREATE DATABASE isy_rag CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci;
```

2. **Modify Configuration File**
Edit `src/main/resources/application.yml` to configure database and Milvus connections:

```yaml
spring:
  datasource:
    url: jdbc:mysql://localhost:3306/isy_rag
    username: your_username
    password: your_password

milvus:
  uri: http://localhost:19530
  enabled: true
```

3. **Build and Run**
```bash
mvn clean package
java -jar target/isy-rag-agent.jar
```

4. **Access the System**
- Admin Panel: http://localhost:8080/admin/login
- User Portal: http://localhost:8080/portal/login

Default admin account: `admin` / `admin123`

## API Endpoints

### Authentication Endpoints
- `POST /api/auth/login` - User login
- `POST /api/auth/logout` - User logout
- `GET /api/auth/current-user` - Get current user

### Admin Panel Endpoints
- `/api/admin/bot` - Bot management
- `/api/admin/kb` - Knowledge base management
- `/api/admin/faq` - FAQ management
- `/api/admin/tool` - Tool management
- `/api/admin/model` - Model management
- `/api/admin/user` - User management
- `/api/admin/trace` - Tracking records

### Portal Endpoints
- `/api/portal/bots` - Get available bots for user
- `/api/portal/conversations` - Conversation list
- `/api/chat/stream` - Streaming chat (SSE)
- `/api/chat/send` - Regular chat

## Project Structure

```
src/main/java/com/isy/rag/
├── RagApplication.java          # Main application class
├── bootstrap/               # Bootstrap modules
│   ├── PageController.java   # Page routing
│   ├── admin/              # Admin panel
│   │   ├── controller/    # Controllers
│   │   ├── dao/          # Data access
│   │   ├── dto/          # Data transfer objects
│   │   ├── service/      # Business logic
│   │   └── mapper/       # MyBatis Mappers
│   ├── portal/            # User portal
│   ├── rag/             # RAG core
│   └── auth/            # Authentication
└── framework/            # Framework layer
    ├── config/          # Configuration
    ├── context/        # Context
    ├── convention/      # Conventions
    ├── database/       # Database
    ├── distributedid/  # Distributed ID
    ├── errorcode/       # Error codes
    ├── exception/      # Exceptions
    ├── idempotent/     # Idempotency
    ├── page/           # Pagination
    ├── trace/          # Tracing
    └── web/            # Web
```

## Configuration Properties

| Property | Description | Default |
|----------|-------------|---------|
| `server.port` | Server port | 8080 |
| `spring.datasource.*` | Database configuration | - |
| `milvus.uri` | Milvus connection address | http://localhost:19530 |
| `milvus.enabled` | Enable Milvus | true |
| `jwt.secret` | JWT secret key | - |
| `jwt.expiration` | JWT expiration time | 86400000 (24 hours) |
| `file.storage.local.base-path` | File upload directory | ./uploads |

## License

This project is intended solely for learning and research purposes.