📄 RAG PDF QA System with PostgreSQL & Ollama

🚀 Overview

This project implements a Retrieval-Augmented Generation (RAG) system that allows users to:

• Upload a PDF document
• Ask questions based on the document
• Get AI-generated answers using a local LLM
• Store chat history persistently using PostgreSQL

Unlike basic chatbot systems, this project combines:

• Vector search (FAISS)
• Embeddings (Sentence Transformers)
• Local LLM (Ollama)
• Persistent memory (PostgreSQL)

---

🧠 Architecture

PDF → Text Extraction → Chunking → Embeddings → FAISS Index
                                      ↓
                               Query Embedding
                                      ↓
                           Relevant Chunk Retrieval
                                      ↓
                    Ollama (LLM) → Answer Generation
                                      ↓
                        PostgreSQL → Store History

---

✨ Features

• 📄 PDF-based question answering
• 🧠 Retrieval-Augmented Generation (RAG)
• 🤖 Local LLM using Ollama (no API cost)
• 🗄️ Persistent chat history (PostgreSQL)
• 👤 Multi-user support with session tracking
• ⚡ Fast similarity search using FAISS

---

🛠️ Tech Stack

• Python 3.10
• PyMuPDF (PDF processing)
• FAISS (vector search)
• Sentence Transformers (embeddings)
• Ollama (local LLM)
• PostgreSQL (database)

---

📦 Installation

1. Clone Repository

git clone https://github.com/your-username/rag-pdf-system.git
cd rag-pdf-system

---

2. Create Virtual Environment

python -m venv venv
venv\Scripts\activate

---

3. Install Dependencies

pip install -r requirements.txt

---

4. Setup PostgreSQL

Create database:

CREATE DATABASE ragdb;

Update credentials in code:

DB_CONFIG = {
    "dbname": "ragdb",
    "user": "postgres",
    "password": "your_password",
    "host": "localhost",
    "port": "5432"
}

---

5. Install & Run Ollama

Download from: https://ollama.com/download

Run:

ollama pull llama3.2

Ollama runs automatically in the background.

---

▶️ Usage

Run the script:

python rag_postgres.py

Follow prompts:

Enter PDF path
Enter user ID (e.g., user1)
Ask questions

---

🧪 Example Queries

• What is Artificial Intelligence?
• Summarize the document
• Explain Machine Learning

---

🗄️ Database Schema

Table: userchat_history

Column	Type
id	SERIAL
user_id	TEXT
session_id	TEXT
question	TEXT
answer	TEXT
timestamp	TIMESTAMP

---

⚠️ Notes

• Ensure the PostgreSQL service is running
• Ollama must be installed and running
• First-time model download may take time

---

🚀 Future Improvements

• Multi-PDF support
• Web UI (Streamlit / React)
• pgvector integration
• Source highlighting
• API deployment (FastAPI)

---

📌 License

This project is for educational purposes.