Justtryai Databricks MCP Server

🚀 Databricks MCP Server

A server implementing the Model Completion Protocol (MCP) for Databricks, providing access to Databricks functionality via the MCP protocol. This enables tools powered by large language models (LLMs) to interact with Databricks clusters, jobs, notebooks, etc.

🚀 Quick Start

The Databricks MCP Server allows LLMs to interact with Databricks through the MCP protocol, providing access to various Databricks features.

✨ Features

• MCP Protocol Support: Implements the MCP protocol, allowing LLMs to interact with Databricks.
• Databricks API Integration: Provides access to the functionality of the Databricks REST API.
• Tool Registration: Exposes Databricks functionality as MCP tools.
• Asynchronous Support: Uses asyncio for efficient operation.

🛠️ Available Tools

The Databricks MCP Server offers the following tools:

• list_clusters: Lists all clusters.
• get_cluster: Retrieves detailed information about a specified cluster.
• start_cluster: Starts a specified cluster.
• stop_cluster: Stops a specified cluster.
• delete_cluster: Deletes a specified cluster.
• create_job: Creates a new job.
• list_jobs: Lists all jobs.
• get_job: Retrieves detailed information about a specified job.
• run_job: Runs a specified job.
• cancel_run: Cancels a specified run.
• delete_run: Deletes a specified run.
• create_notebook: Creates a new notebook.
• list_notebooks: Lists all notebooks.
• get_notebook: Retrieves detailed information about a specified notebook.
• run_notebook: Runs a specified notebook.

📦 Installation

Prerequisites

1. Python 3.8 or higher.
2. The pip package manager.
3. Basic development tools (e.g., an IDE and a terminal).

Setup Steps

Step 1: Clone the Repository

Clone the Databricks MCP Server repository using the following command:

git clone https://github.com/your-organization/databricks-mcp-server.git

Step 2: Install Dependencies

After entering the project directory, install all dependencies:

pip install -r requirements.txt

Step 3: Start the Server

Run the following command to start the MCP server:

python src/server.py

💻 Running the Server

Start Script

To start the server, use the following script:

# Start the MCP server
.\scripts\start_server.ps1

# Start in detached mode
.\scripts\start_server.ps1 -Detached

Stop Script

To stop the running server, use the following script:

# Stop the MCP server
.\scripts\stop_server.ps1

💻 Usage Examples

Basic Usage

In the examples/ folder, you can find examples of direct usage of the server:

from databricks_mcp.client import DatabricksMCPClient

client = DatabricksMCPClient('http://localhost:8080', 'your_token')
cluster_info = client.get_cluster(cluster_id)
print(cluster_info)

Advanced Usage

In the examples/ folder, you can also find examples of using the MCP client:

from databricks_mcp.client import DatabricksMCPClient

client = DatabricksMCPClient('http://localhost:8080', 'your_token')
response = client.run_job(job_id)
print(response)

📁 Project Structure

The detailed structure of the project is as follows:

Databricks-MCP-Server/
├── src/
│   ├── server.py          # Main server file
│   └── client.py         # MCP client implementation
├── tests/
│   ├── test_clusters.py   # Test cluster functionality
│   └── test_jobs.py      # Test job functionality
├── scripts/
│   ├── start_server.ps1  # Script to start the server
│   └── stop_server.ps1   # Script to stop the server
├── examples/
│   ├── direct_usage.py   # Example of direct usage
│   └── mcp_client_usage.py # Example of using the MCP client
└── pyproject.toml        # Project configuration file

🔧 Development

Coding Standards

• Follow the PEP 8 style guide with a maximum line length of 100 characters.
• Use 4 spaces for indentation (no tabs).
• Use double quotes for strings.
• All classes, methods, and functions must include Google-style docstrings.
• All code, except test code, must have type hints.

Source Code Analysis Tools

The project uses the following source code analysis tools:

# Run all code check tools
uv run pylint src/ tests/
uv run flake8 src/ tests/
uv run mypy src/

🧪 Testing

The project uses pytest for testing. To run the tests, execute the following commands:

# Run all tests using a convenient script
.\scripts\run_tests.ps1

# Generate a code coverage report
.\scripts\run_tests.ps1 -Coverage

# Run specific tests and output detailed information
.\scripts\run_tests.ps1 -Verbose -Coverage tests/test_clusters.py

You can also use pytest directly to run the tests:

# Run all tests
pytest

# Generate a code coverage report
pytest --cov

📚 Documentation

The project documentation is located in the docs/ folder. You can refer to the API documentation and development guide.

🤝 Contribution

Contributions to this project are welcome! Please fork the repository and create a pull request before submitting your changes.

📄 License

This project is open-source under the LICENSE.