McpProject

[//]: # (# 模块化 MCP 客户端) [//]: # () [//]: # (这是一个模块化、可扩展的 MCP (Model Control Protocol) 客户端实现，用于连接大模型 API 和多个工具服务器。) ## Structure Description: 1. **Configuration Management** (`config.py`) - Responsible for loading and validating configurations from environment variables. 2. **MCP Server Type Connector** (`server_connector.py`) - Responsible for connecting to and managing multiple MCP servers. - Supports connecting to local scripts and NPX packages. 3. **Model Client** (`model_client.py`) - Responsible for interacting with large model APIs. - Handles streaming responses, tool calls, and result processing. 4. **Local MCP Service Storage (py)** (`mcpserver`) - Stores local Python MCP services, which can be extended and developed. 5. **MCP Configuration Information Loader** (`mcp_config_loader.py`) - Parses the configuration information from the `mcp_servers.json` file. 6. **MCP Configuration File** (`mcp_servers.json`) - Configuration file with the same functionality as Cursor's MCP file. 7. **Main Application** (`main.py`) - Integrates all the above modules. - Provides a command-line interface and interactive chat. [//]: # (## 特性) [//]: # () [//]: # (- 支持连接多个 MCP 服务器) [//]: # (- 模块化设计便于扩展) [//]: # (- 统一的工具调用接口) [//]: # (- 改进的错误处理和日志记录) [//]: # (- 支持流式输出和推理过程显示) ## Usage ## Environment Configuration: 1. After pulling the project, navigate to the root directory: ``` # Create a virtual environment uv venv # Activate the virtual environment .venv/bin/activate # Install dependencies uv pip install -r requirements.txt uv add dotenv ``` 2. In the root directory, copy `.env example` and rename it to `.env`, then set the necessary environment variables: ``` # apikey for 阿里百练 (Ali BaiLian) DASHSCOPE_API_KEY=your_dashscope_api_key # Address for 阿里百练 (Ali BaiLian) BASE_URL=https://dashscope.aliyuncs.com/compatible-mode/v1 # Select the model MODEL=xxxx ``` ## Adding MCP Unified management based on JSON files. Copy `mcp_servers.json example` and rename it to `mcp_servers.json`. 1. If the newly added MCP service is a self-written `.py` file, refer to `weather_server.py`. 2. If it is another remote MCP, copy the corresponding MCP's JSON and add it to `mcp_servers.json` (consistent with Cursor's configuration). ```json { "mcpServers": { "YourMcpName": { "disabled": false, "command": "xxx", "args": ["xxx"], "env": {} } } } ``` ### Running (Either 1 or 2 will work) 1. Command Line After activating the virtual environment: ```bash python src/main.py ``` 2. PyCharm Can also be run in PyCharm.