---

PDF Question Answering System

Overview

The PDF Question Answering System is a web application that allows users to upload PDF documents and ask questions about their content. The system extracts text from the PDFs, processes it, and leverages state-of-the-art language models to provide accurate and context-aware answers to user queries. https://huggingface.co/spaces/Arbazkhan-cs/Retrieval-Augmented-Generation

Features

• PDF Text Extraction: Utilizes PyPDF2 to extract text from uploaded PDF documents.
• Text Chunking: Employs RecursiveCharacterTextSplitter to split text into manageable chunks based on newline characters.
• Vector Store Indexing: Uses FAISS for efficient text indexing and retrieval.
• Embeddings Generation: Leverages HuggingFace's sentence-transformers for creating embeddings.
• Natural Language Processing: Integrates the ChatGroq large language model (LLM) to handle and respond to natural language queries.
• User-Friendly Interface: Built with Streamlit for easy interaction and query submission.

How It Works

1. Upload a PDF: Users upload a PDF document through the Streamlit interface.
2. Text Processing: The application extracts and splits the text into chunks.
3. Indexing: The text chunks are indexed using FAISS.
4. Query Submission: Users input questions related to the PDF content.
5. Answer Retrieval: The system retrieves the most relevant text chunks and generates answers using the LLM.

Setup and Installation

Prerequisites

• Python 3.8 or higher
• Pip package manager

Installation

1. Clone the Repository:

   git clone https://github.com/Arbazkhan-cs/Retrieval-Augmented-Generation.git
   cd your-repo

2. Install Dependencies:

   pip install -r requirements.txt

Then, run the app using the following command:

streamlit run app.py

Technologies Used

• Languages and Frameworks: Python, Streamlit
• Libraries: PyPDF2, langchain, sentence-transformers, FAISS, dotenv
• Models and Tools: Hugging Face LLMs, ChatGroq, RecursiveCharacterTextSplitter

Contributing

Contributions are welcome! Please open an issue or submit a pull request for any improvements or bug fixes.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Contact

For any inquiries, please contact your email.

---

Feel free to customize the text to better fit your project's specifics and your preferences.