Model Context Protocal (MCP) Implementation

This repository includes the Model Context Protocol (MCP) framework that ClimateGPT Team 1 is developing.

📂 Project Structure

/mcp-framework ├── modules/ # Core MCP components │ ├── context_manager.py # Stores execution context memory │ ├── data_loader.py # Handles dataset loading │ ├── query_manager.py # Routes queries dynamically │ ├── pipeline_manager.py # Executes MCP steps ├── models/ # Test EDA / initial models for MCP framework checking │ ├── scenario_projection.py # Temp trend analysis │ ├── temperature_trends.py # Climate scenario projections │ ├── Model3.py # Model 3 ├── config/ # Configuration settings │ ├── config.yaml # Defines dataset paths and pipeline steps ├── logs/ # Execution logs │ ├── mcp_execution.log ├── tests/ # Unit tests for MCP validation ├── main.py # Entry point for MCP execution ├── requirements.txt # Python dependencies ├── README.md # Project documentation

How to run MCP Framework

1. Clone the repository (if not already cloned):

   git clone https://github.com/ newsconsole/GMU_DAEN_2025_01_A.git
   

   Copy

2. Switch to the ClimateGPT Team 1 Branch:

   git checkout ClimateGPT_Team1
   

   Copy

3. Make sure to set up venv (Virtual Env)

   1. python -m venv venv
   2. venv\Scripts\Activate
   

   Copy

4. Install dependencies (requirements.txt):

   pip install -r requirements.txt
   

   Copy

5. Run the MCP Pipeline

   python main.py 
   

   Copy

Configuration & Execution

• The MCP pipeline is dynamically controlled by config/config.yaml which defines the datasets and pipeline steps
• Logs are stored in logs/mcp_execution.log for debugging and tracking execution results

Recent Updates

• Implemented initial MCP Framework with modular design
• Added dynamiic query routing & context memory