database2prompt

An open-source project designed to extract relevant data from databases and transform it into context for Retrieval-Augmented Generation (RAG) in generative AI applications.

How is it useful?

database2prompt makes it easy to generate prompts to LLMS by reading your database and generating a markdown containing its schema. This provides context for the AI to maximize the effectiveness of your prompts.

Databases Support (WIP)

Databases	Support
PostgreSQL	✅

We will add support for most databases including analytical databases

Output Formats

Output Format	Support
JSON	✅
Markdown	✅

Example Outputs

You can find example outputs generated by database2prompt in the following files:

• summary-database.md - Example of markdown output
• summary-database.json - Example of JSON output

Usage

Installation

pip install database2prompt

Quick Start

Here's a simple example of how to use database2prompt:

from database2prompt.database.core.database_config import DatabaseConfig
from database2prompt.database.core.database_params import DatabaseParams
from database2prompt.database.core.database_factory import DatabaseFactory
from database2prompt.database.processing.database_processor import DatabaseProcessor
from database2prompt.markdown.markdown_generator import MarkdownGenerator

# 1. Configure database connection
config = DatabaseConfig(
    host="localhost",
    port=5432,
    user="your_user",
    password="your_password",
    database="your_database",
    schema="your_schema"
)

# 2. Connect to database
strategy = DatabaseFactory.run("pgsql", config)
next(strategy.connection())

# 3. Configure which tables to document
params = DatabaseParams()

# Option A: Document specific tables
params.tables(["schema.table1", "schema.table2"])

# Option B: Ignore specific tables
params.ignore_tables(["schema.table_to_ignore"])

# 4. Process database information
database_processor = DatabaseProcessor(strategy, params)

# 5. Generate content to prompt (markdown or json)
content = database_processor.database_to_prompt(output_format="json")

Configuration

Configure the database connection:

   # .env file
   DB_HOST=localhost
   DB_PORT=5432
   DB_USER=postgres
   DB_PASSWORD=postgres
   DB_NAME=postgres
   DB_SCHEMA=public

   config = DatabaseConfig.from_env()

Contributing

Development Setup

1. Clone the repository:

   git clone https://github.com/orladigital/database2prompt.git
   cd database2prompt

2. Create a virtual environment:

   python -m venv .venv
   source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install development dependencies:

   pip install poetry
   poetry install

4. Start the development database (optional):

   docker compose up -d

5. Run the project:

   poetry run python database2prompt/main.py

How to Contribute

You can contribute to database2prompt in many different ways:

• Suggest a feature
• Code an approved feature idea (check our issues)
• Report a bug
• Fix something and open a pull request
• Help with documentation
• Spread the word!

License

Licensed under the MIT License, see LICENSE for more information.