db2-mcp-server

Overview

The db2-mcp-server is a Python-based server utilizing the MCP framework to interact with IBM DB2 databases. It provides tools for listing tables and retrieving table metadata.

Features

• List Tables: Retrieve a list of tables from the connected DB2 database.
• Get Table Metadata: Fetch metadata for a specific table, including column details and data types.
• Dynamic Prompt Loading: Load custom prompts from JSON configuration files for flexible query assistance.
• Built-in Prompts: Pre-configured prompts for DB2 query help and schema analysis.

Requirements

• Python 3.12
• FastMCP (latest stable version)
• IBM DB2 Python driver (ibm_db)
• Pydantic

Setup Instructions

1. Clone the Repository

   git clone <repository-url>
   cd db2-mcp-server
   

2. Set Up Virtual Environment

   uv v0.6.x
   source uv/bin/activate
   

3. Install Dependencies

   pip install -r requirements.txt
   

4. Run the Server

   python src/db2_mcp_server/core.py
   

Dynamic Prompts

The server supports dynamic prompt loading from JSON configuration files. Set the PROMPTS_FILE environment variable to specify a custom prompts configuration file.

Example Usage

export PROMPTS_FILE=/path/to/your/prompts_config.json
python src/db2_mcp_server/core.py

Configuration Format

See examples/prompts_config.json for a complete example and docs/DYNAMIC_PROMPTS.md for detailed documentation.

Testing

• Use pytest (version ≥ 7.0.0) for running tests.
• Current test coverage: 92.98% (exceeding the 80% requirement).
• Comprehensive test suite includes:
  • Core functionality tests (test_core.py)
  • Database tools tests (test_list_tables.py, test_metadata_retrieval.py)
  • Caching mechanism tests (test_cache.py)
  • Logging configuration tests (test_logger.py)
  • Dynamic prompt loading tests (test_dynamic_loader.py, test_integration_dynamic_prompts.py)
• Run tests with:

  pytest --cov=src/db2_mcp_server --cov-report=html tests/
  

• For verbose output:

  pytest --cov=src/db2_mcp_server --cov-report=term-missing -v
  

Security and API Restrictions

• The server is read-only, prohibiting SQL INSERT, UPDATE, or DELETE operations.
• Uses a database user with only SELECT privileges.

Logging

• Errors are logged with structured logs in JSON format, excluding sensitive data.

Contribution

• Contributions are welcome. Please follow the project's coding standards and testing guidelines.