What is Digit

DigIt is an MCP-enabled agent designed for efficient context management within complex data processing pipelines. It integrates the Model Context Protocol with the Dora dataflow framework and the MOFA Python framework.

Use cases

Use cases for DigIt include managing context in machine learning workflows, facilitating data flow in data engineering projects, and enhancing the efficiency of software applications that rely on complex data interactions.

How to use

To use DigIt, set up a Conda environment, install Rust, the Dora CLI, and the MOFA framework. Then clone the DigIt repository and follow the installation instructions provided in the README.

Key features

Key features of DigIt include seamless integration with the Dora dataflow framework, efficient context management capabilities, and the use of the MOFA framework for enhanced data processing.

Where to use

DigIt can be used in various fields that require complex data processing, such as data science, machine learning, and software development.

DigIt: An MCP-Enabled Agent with Dora Dataflow and MOFA Framework

Introduction

DigIt is an agent that implements the Model Context Protocol MCP to facilitate interaction and context management within dataflows. This project integrates an MCP functional agent within the Dora dataflow framework and leverages the MOFA Python framework for its development, providing a structured and efficient way to manage context in complex processing pipelines.

Installation

Follow these steps to set up and run DigIt:

1. Install Conda Environment

1. Download and install Conda from the official website: [Need official download link here]
2. Create a new conda environment:

   conda create -n mofa python=3.10
   

3. Activate the newly created environment:

   conda activate mofa
   

   All subsequent commands should be executed within this activated conda environment.

2. Install Rust

1. Install Rust using the following command:

   curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
   

   Follow the on-screen instructions to complete the installation.

3. Install Dora

1. Install the Dora command-line interface (CLI) using Cargo:

   cargo install dora-cli
   

   Ensure that Rust’s Cargo package manager is installed and configured correctly.

4. Install MOFA Environment

1. Clone the MOFA repository:

   git clone https://github.com/moxin-org/mofa.git
   

2. Navigate to the Python directory within the MOFA repository:

   cd mofa/python
   

3. Install MOFA in editable mode:

   pip install -e .
   

4. Go back to the root of the MOFA repository:

   cd ..
   

5. Clone This Project

1. Clone the DigIt repository:

   git clone https://github.com/KeriaDaring/DigIt.git
   

2. Navigate to the DigIt project directory:

   cd DigIt
   

3. Install the project dependencies:

   pip install -r requirement.txt
   

6. Configure API Keys

1. Edit the API key in the following configuration files:
   • ./agent/mcp-llm/mcp_llm/configs/chat_session.yml
   • ./configs/beaufy-context.yml
     Replace the placeholder API key with your actual API key in both files.

7. Configure MCP Server

1. Edit the MCP server configuration file: ./agent/mcp-llm/mcp_llm/configs/servers_config.json

   MCP Server Configuration: Modify the servers_config.json file to match your local setup:

   {
     "mcpServers": {
       "markdown_processor": {
         "command": "/path/to/your/uv",
         "args": [
           "--directory",
           "/path/to/your/project/mcp_servers",
           "run",
           "markdown_processor.py"
         ]
       }
     }
   }
   

   • Replace /path/to/your/uv with the actual path to your uv executable. You can find this path using the command which uv.
   • Replace /path/to/your/project/mcp_servers with the absolute path to the mcp_servers directory within your project.

8. Start the Project

1. Destroy any existing Dora sessions:

   dora destroy
   

2. Build and start the DigIt dataflow:

   dora up && dora build digit_dataflow.yml && dora start digit_dataflow.yml
   

3. Open another terminal window.

4. Activate the MOFA conda environment:

   conda activate mofa
   

5. In this new terminal, you can now input your prompt to start using DigIt.

   # Example:
   Send You Task : 告诉我关于mofa框架的相关细节