SmartWeld AI-基于MCP的物联网数据智能平台
🤖 一种先进的人工智能驱动系统，使用模型上下文协议（MCP）、LangChain和会话式人工智能将物联网焊接数据转化为可操作的见解。
🌟 概述
SmartWeld AI是一个完整的物联网数据智能平台，可以与工业焊接数据进行自然语言交互。建立在 模型上下文协议（MCP），它提供：
- 🗣️ 对话式AI界面 -用自然语言与您的数据聊天
- 🔧 MCP电动工具 -可扩展的数据访问工具生态系统
- 📊 动态可视化 -自动生成的图表和见解
- 🏭 物联网集成 -实时MQTT数据处理
- 🧠 智能分析 -LLM动力数据解释
🏗️ 建筑
┌─────────────────────────────────────────────────────────────────┐
│ SmartWeld AI Platform │
├─────────────────┬─────────────────┬─────────────────────────────┤
│ Frontend │ Agentic AI │ MCP Server │
│ (React) │ Server │ (Tools) │
│ │ (LangChain) │ │
│ ┌─────────────┐│ ┌─────────────┐ │ ┌─────────────────────────┐ │
│ │ Chat UI ││ │ Conversational│ │ │ Database Tools │ │
│ │ Components ││ │ Agent │ │ │ - LangChain SQL Agent │ │
│ │ Visualizations││ │ GPT-4 LLM │ │ │ - Schema Discovery │ │
│ │ Dashboard ││ │ Tool Calling │ │ │ - Query Execution │ │
│ └─────────────┘│ └─────────────┘ │ └─────────────────────────┘ │
│ │ │ ┌─────────────────────────┐ │
│ │ │ │ Weather Tools │ │
│ │ │ │ - Forecast API │ │
│ │ │ │ - Alerts Integration │ │
│ │ │ └─────────────────────────┘ │
└─────────────────┴─────────────────┴─────────────────────────────┘
│ │ │
▼ ▼ ▼
┌─────────────────┐ ┌─────────────────┐ ┌─────────────────┐
│ User Input │ │ LLM Processing │ │ Data Sources │
│ Natural Lang │ │ Tool Selection │ │ PostgreSQL │
│ Voice/Text │ │ Response Gen │ │ Weather APIs │
└─────────────────┘ └─────────────────┘ └─────────────────┘🚀 快速开始
先决条件
- Node.js v18+和npm
- PostgreSQL v13+数据库
- OpenAI API密钥 （用于GPT-4接入）
- Git 用于版本控制
1.克隆存储库
git clone
cd SmartWeld-MCP-AI2.环境设置
为每个组件创建环境文件：
MCP服务器 (.env 在……里面 mcp-server-stdio/)
# Database Configuration
DB_HOST=localhost
DB_PORT=5432
DB_NAME=smartweld
DB_USER=your_db_user
DB_PASSWORD=your_db_password
DB_SCHEMA=public
# AI Configuration
OPENAI_API_KEY=sk-your-openai-api-key
# Server Configuration
NODE_ENV=development代理服务器 (.env 在……里面 agentic-server/)
# Server Configuration
NODE_ENV=development
PORT=5001
CORS_ORIGIN=http://localhost:3000
# Database Configuration
DB_HOST=localhost
DB_PORT=5432
DB_NAME=smartweld
DB_USER=your_db_user
DB_PASSWORD=your_db_password
DB_SCHEMA=public
# AI Configuration
OPENAI_API_KEY=sk-your-openai-api-key
LLM_MODEL=gpt-4o
# Optional: Additional LLM Providers
GEMINI_API_KEY=your-gemini-api-key
ANTHROPIC_API_KEY=your-anthropic-api-key
# Security
JWT_SECRET=your-super-secret-jwt-key
ALLOWED_ORIGINS=http://localhost:3000,https://yourdomain.com
# Monitoring & Logging
LOG_LEVEL=info
MONITORING_ENABLED=true反应前端 (.env 在……里面 ai-chatbot-smartweld/)
# API Configuration
REACT_APP_API_URL=http://localhost:5001
NODE_ENV=development
# Optional: Authentication
REACT_APP_AUTH_ENABLED=false3.数据库设置
-- Create database
CREATE DATABASE smartweld;
-- Create user (optional)
CREATE USER smartweld_user WITH PASSWORD 'your_secure_password';
GRANT ALL PRIVILEGES ON DATABASE smartweld TO smartweld_user;
-- Connect to database and create sample tables
\c smartweld;
CREATE TABLE device_telemetry (
id SERIAL PRIMARY KEY,
device_id VARCHAR(50),
temperature DECIMAL(5,2),
voltage DECIMAL(5,2),
current DECIMAL(5,2),
gas_level DECIMAL(5,2),
timestamp TIMESTAMP DEFAULT CURRENT_TIMESTAMP
);
-- Insert sample data
INSERT INTO device_telemetry (device_id, temperature, voltage, current, gas_level) VALUES
('WELD001', 750.5, 220.0, 15.2, 85.5),
('WELD002', 680.3, 210.5, 12.8, 78.2),
('WELD003', 820.1, 230.2, 18.5, 92.1);4.安装与启动
终端1:MCP服务器
cd mcp-server-stdio
npm install
npm run build
npm start终端2：代理AI服务器
cd agentic-server
npm install
npm run dev终端3：反应前端
cd ai-chatbot-smartweld
npm install
npm start5.验证安装
- 检查MCP服务器：应显示“MCP正在stdio上运行”
- 检查AI服务器：参观
http://localhost:5001/health - 检查前端：参观
http://localhost:3000
📁 项目结构
SmartWeld-MCP-AI/
├── mcp-server-stdio/ # MCP Protocol Server
│ ├── src/
│ │ ├── tools/ # MCP Tools Implementation
│ │ │ ├── databaseTools/ # Database query tools
│ │ │ └── weatherForecastTool/ # Weather API tools
│ │ ├── config/ # Database & connection config
│ │ ├── interfaces/ # TypeScript interfaces
│ │ └── index.ts # MCP server entry point
│ ├── build/ # Compiled JavaScript
│ └── package.json
│
├── agentic-server/ # AI Agent Server
│ ├── src/
│ │ ├── agents/ # AI agent implementations
│ │ │ ├── conversational-agent.ts # Main chat agent
│ │ │ ├── sql-toolkit.ts # Database analysis tools
│ │ │ └── rag-retriever.ts # RAG for documents
│ │ ├── mcp/ # MCP client integration
│ │ │ ├── mcpClient.ts # MCP connection client
│ │ │ └── mcpToolWrapper.ts # Tool wrapper utilities
│ │ ├── config/ # Database & app configuration
│ │ ├── models/ # Data models
│ │ └── index.ts # Express server entry point
│ └── package.json
│
├── ai-chatbot-smartweld/ # React Frontend
│ ├── src/
│ │ ├── components/ # React components
│ │ │ ├── Chat.jsx # Main chat interface
│ │ │ └── chatbotComponents/ # Visualization components
│ │ ├── chatbotAPICalls/ # API integration
│ │ ├── contexts/ # React contexts
│ │ └── pages/ # Application pages
│ ├── public/
│ └── package.json
│
├── README.md # This file
└── blog-post-mcp-smartweld.md # Technical blog post🔧 配置指南
数据库配置
系统支持多种数据库配置：
// Production configuration
export const dbConfig = {
host: process.env.DB_HOST,
port: parseInt(process.env.DB_PORT || '5432'),
database: process.env.DB_NAME,
username: process.env.DB_USER,
password: process.env.DB_PASSWORD,
schema: process.env.DB_SCHEMA || 'public',
ssl: process.env.NODE_ENV === 'production'
};MCP工具配置
通过实现MCP工具接口添加新工具：
export const customTool: Tool = {
name: "custom-tool",
description: "Description of what the tool does",
inputSchema: {
type: "object",
properties: {
query: { type: "string" }
}
},
async execute(input: Record) {
// Tool implementation
return { content: "Tool result" };
}
};AI模型配置
支持多个LLM提供商：
// OpenAI (recommended)
const openaiModel = new ChatOpenAI({
model: "gpt-4o",
apiKey: process.env.OPENAI_API_KEY,
temperature: 0,
});
// Gemini (alternative)
const geminiModel = new ChatGoogleGenerativeAI({
model: "gemini-1.5-flash",
apiKey: process.env.GEMINI_API_KEY,
});🛠️ 可用命令
MCP服务器命令
npm run build # Compile TypeScript to JavaScript
npm run build:unix # Build with Unix permissions
npm run dev # Development mode with auto-reload
npm start # Start compiled server代理服务器命令
npm run start # Production mode
npm run dev # Development mode with nodemon
npm test # Run tests (if configured)前端命令
npm start # Development server
npm run build # Production build
npm test # Run test suite
npm run eject # Eject from Create React App🔐 安全和凭据管理
环境变量安全
- 永不承诺
.env文件 到版本控制 - 使用不同的凭据进行开发和生产
- 定期旋转API键
- 使用强大、唯一的密码访问数据库
API密钥管理
# Secure storage of API keys
export OPENAI_API_KEY="sk-your-key-here"
export GEMINI_API_KEY="your-gemini-key"
# Or use a secure key management service
# AWS Secrets Manager, Azure Key Vault, etc.数据库安全性
-- Create dedicated user with limited permissions
CREATE USER smartweld_app WITH PASSWORD 'secure_random_password';
GRANT SELECT, INSERT, UPDATE ON ALL TABLES IN SCHEMA public TO smartweld_app;
GRANT USAGE ON ALL SEQUENCES IN SCHEMA public TO smartweld_app;🚀 部署
生产部署检查表
- 环境设置
- \[\]设置 NODE_ENV=production - \[\]配置生产数据库 - \[\]设置安全的CORS源 - \[\]启用HTTPS
- 数据库迁移
# Run migrations
npm run migrate:up
# Seed initial data
npm run seed:production- 服务器部署
# Build all components
cd mcp-server-stdio && npm run build
cd ../agentic-server && npm run build
cd ../ai-chatbot-smartweld && npm run build
# Start with process manager
pm2 start ecosystem.config.js- Docker部署 （可选）
# Build Docker image
docker build -t smartweld-ai .
# Run with environment variables
docker run -d \
--name smartweld-ai \
-p 5001:5001 \
-e OPENAI_API_KEY=your-key \
-e DB_HOST=your-db-host \
smartweld-ai生产环境变量
# Production specific
NODE_ENV=production
PORT=5001
CORS_ORIGIN=https://your-domain.com
# Database (use connection pooling)
DB_MAX_CONNECTIONS=20
DB_IDLE_TIMEOUT=30000
# Security
JWT_SECRET=super-secure-64-character-random-string
RATE_LIMIT_WINDOW_MS=900000
RATE_LIMIT_MAX_REQUESTS=100
# Monitoring
LOG_LEVEL=warn
MONITORING_ENABLED=true
METRICS_PORT=9090🐛 故障排除
常见问题
1.MCP连接失败
# Check if MCP server is running
ps aux | grep "mcp-server"
# Check logs
tail -f mcp-server-stdio/logs/server.log
# Test connection manually
node -e "
const { connectToMcpServer } = require('./agentic-server/src/mcp/mcpClient');
connectToMcpServer('./mcp-server-stdio/build/index.js').then(console.log);
"2.数据库连接问题
# Test database connectivity
psql -h localhost -U smartweld_user -d smartweld -c "SELECT 1"
# Check connection string
echo $DATABASE_URL
# Verify credentials
npm run db:test3.OpenAI API错误
# Verify API key format
echo $OPENAI_API_KEY | grep "^sk-"
# Test API connectivity
curl -H "Authorization: Bearer $OPENAI_API_KEY" \
"https://api.openai.com/v1/models"
# Check rate limits
npm run ai:test4.前端API连接
// Test API endpoint
fetch('http://localhost:5001/health')
.then(r => r.json())
.then(console.log);
// Check CORS configuration
console.log('API URL:', process.env.REACT_APP_API_URL);调试模式
启用调试日志记录：
# In .env files
DEBUG=true
LOG_LEVEL=debug
VERBOSE_LOGGING=true查看调试信息：
# MCP Server debug
DEBUG=mcp* npm start
# AI Server debug
DEBUG=agent* npm run dev
# Frontend debug
REACT_APP_DEBUG=true npm start📊 监测和可观察性
健康检查
# Check all services
curl http://localhost:5001/health
curl http://localhost:3000/health度量与记录
// Built-in metrics endpoint
GET /metrics // Prometheus format
GET /status // System status性能监控
# Database query performance
npm run db:analyze
# LLM response times
npm run ai:benchmark
# Frontend bundle analysis
npm run analyze🤝 贡献
开发工作流程
- 分叉与克隆
git clone
cd SmartWeld-MCP-AI- 创建特征分支
git checkout -b feature/your-feature-name- 开发设置
# Install dependencies for all components
npm run install:all
# Start development environment
npm run dev:all- 代码规范
# Lint code
npm run lint
# Format code
npm run format
# Run tests
npm run test:all- 提交更改
git add .
git commit -m "feat: add your feature description"
git push origin feature/your-feature-name代码风格指南
- TypeScript：严格模式已启用
- 掉毛：带Prettier的ESLint
- 命名：camelCase用于变量，PasclCase用于组件
- 承诺：常规提交格式
📚 额外资源
文档
社区
相关项目
📄 许可证
此项目根据MIT许可证获得许可-请参阅 许可证 文件以获取详细信息。
👥 作者与致谢
- 开发团队：SmartWeld人工智能团队
- 特别感谢：Anthropic（MCP）、LangChain、OpenAI社区
- 灵感：工业物联网和人工智能民主化
______________________________________________________________________
🆘 支持
有问题吗？我们是来帮忙的！
- 📧 电子邮件: \[your-email@domain.com\]
- 💬 Discord 的中文翻译是“不和谐”或“纷争”。：\[您的Discord服务器\]
- 🐛 问题:
- 📖 维基: 维基工程
______________________________________________________________________
*内置于❤️ 面向智能制造的未来*