RAG-based Document QA Assistant

An end-to-end Retrieval-Augmented Generation (RAG) system built with Python, Google Gemini, and ChromaDB that allows you to ingest any PDF document and ask questions with precise, context-aware answers.

---

Features

• Universal Document Support: Works with resumes, research papers, textbooks, technical reports, and multi-page PDFs.
• Smart Sliding-Window Chunking: Splits text into coherent, overlapping chunks with paragraph and sentence preservation.
• Fast Batch Vector Embeddings: Uses Google's gemini-embedding-001 model with batching for high speed and rate-limit protection.
• Local Persistent Vector Storage: Stores and retrieves document embeddings locally with ChromaDB.
• Context-Aware QA: Answers questions using Google's gemini-3.6-flash model with retrieved source citations.
• Cross-Platform & Windows Safe: Configured with UTF-8 console output to prevent terminal encoding crashes.

---

Project Structure

├── pdf_utils.py        # PDF text extraction, normalization, and universal chunking
├── read_pdf.py         # Utility to inspect extracted text from a PDF
├── chunk_pdf.py        # Utility to inspect generated text chunks
├── embed_test.py       # Test script for generating vector embeddings
├── store_chunks.py     # Main ingestion script: embeds & stores chunks in ChromaDB
├── ask_question.py     # Q&A CLI: queries vector database and answers via Gemini
├── test.py             # Basic Gemini API connectivity test
├── requirements.txt    # Project dependencies
├── .env.example        # Environment variable template
└── .gitignore          # Git ignore rules

---

Prerequisites

• Python 3.9+
• A Google Gemini API Key (get one free at Google AI Studio)

---

Setup & Installation

1. Clone the repository:

   git clone https://github.com/hems705/RAG-based-QA-document.git
   cd RAG-based-QA-document

2. Create and activate a virtual environment:

   python -m venv .venv
   # Windows (PowerShell):
   .\.venv\Scripts\activate
   # Linux / macOS:
   source .venv/bin/activate

3. Install dependencies:

   pip install -r requirements.txt

4. Set up your environment variables: Create a .env file in the root directory:

   GEMINI_API_KEY=your_gemini_api_key_here

5. Verify API connectivity:

   python test.py

---

Usage

1. Ingest a PDF into ChromaDB

Place your PDF in the project folder or specify its path:

# Ingest by specifying a path:
python store_chunks.py path/to/document.pdf

# Or simply place your PDF in the project folder and run:
python store_chunks.py

2. Ask Questions

Ask specific questions or query the document:

# Ask a specific question:
python ask_question.py What are the key findings discussed in this document?

# Or run without arguments for an automatic summary & key takeaways:
python ask_question.py

3. Inspection Tools (Optional)

• Inspect raw extracted text:

  python read_pdf.py [document.pdf]

• Inspect text chunking:

  python chunk_pdf.py [document.pdf]

• Test embedding generation:

  python embed_test.py [document.pdf]

---

License

MIT License