#RAG

RAG is a production-grade Agentic Retrieval-Augmented Generation system. This project is built using Python, and uses PostgreSQL with pgvector for scalable, persistent vector storage.

Prerequisites

• Python 3.10+
• PostgreSQL Database running with the pgvector extension installed.

Setup Guide

1. Database Setup

Ensure PostgreSQL is running on your machine. You need to create an empty database (the default expected database name is ragdb, with user postgres).

CREATE DATABASE ragdb;

Note: The pgvector extension must be installed in your PostgreSQL environment.

2. Environment Configuration

Navigate to the rag_project directory and copy the environment template:

cd rag_project
cp .env.example .env

Fill out the required configuration variables in the .env file, especially your database credentials and API keys (e.g., OpenAI, Gemini, or Github).

3. Install Dependencies

It is highly recommended to use a virtual environment. From the project root folder, run:

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

pip install -r requirements.txt

4. Initialise Database Schema

As this project uses Alembic to manage database migrations, you must apply the initial schema before running the application:

cd rag_project
alembic upgrade head

5. Add Documents

Create a data/ directory at the root of the project and place any PDFs, PowerPoint files, or text data you want the RAG system to index inside it. The system's document loaders are configured to automatically scan the <repo_root>/data/ path by default.

Running the Application

To start the interactive DRAGON chat interface, run:

cd rag_project
python main.py

In-App Commands

• /refresh : Scans for new or changed files incrementally and updates the vector database.
• /reset : Completely wipes the database and forces a full re-index from scratch.
• exit : Quits the application.

Version Control

This repository has its database schema migrations consolidated to a single baseline. If you make modifications to the rag/db.py models, you can generate a new migration by running:

alembic revision --autogenerate -m "describe changes"
alembic upgrade head