ChatPDF

Interact with your PDF documents using Google Gemini!
This Streamlit app lets you upload PDF files, process their contents, and then ask questions that are answered contextually using Google's Gemini model.

---

Features

• Upload one or more PDF files.
• Extracts and stores their contents efficiently.
• Embeds PDF data using Google Generative AI embeddings.
• Stores embeddings locally with FAISS for fast retrieval.
• Ask natural language questions about your PDFs.
• Answers generated by Gemini via LangChain's conversational retrieval QA.

---

Installation

1. Clone the repository

   git clone https://github.com/avanshh99/Chat-with-PDF-using-Gemini.git
   cd Chat-with-PDF-using-Gemini

2. Install the required packages It's recommended to use a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows use venv\Scripts\activate
   pip install -r requirements.txt

   If requirements.txt is missing, install main packages:

   pip install streamlit PyPDF2 langchain langchain-google-genai google-generativeai langchain-community faiss-cpu python-dotenv

3. Set up your Google API Key

   • Create a .env file in the root directory:

     GOOGLE_API_KEY=your_google_api_key_here
     

   • You can get an API key from Google Generative AI documentation.

---

Usage

1. Start the Streamlit application

   streamlit run your_script_name.py

2. In the app:

   • Upload one or more PDF files via the sidebar.
   • Click "Submit & Process" to extract and embed the data.
   • Ask questions in the main input box — the app will answer using the content of your PDFs!

---

How it works

• PDF Extraction: Uses PyPDF2 to extract text from uploaded PDFs.
• Text Chunking: Splits the text using LangChain’s RecursiveCharacterTextSplitter for efficient processing.
• Embedding: Converts chunks into vector embeddings using Google Generative AI.
• Vector Store: Stores embeddings in a local FAISS index for similarity search.
• Conversational QA: Uses LangChain’s QA chain with Gemini to answer user questions based on retrieved document chunks.

---

File Structure

your_repo/
├── your_script_name.py
├── requirements.txt
├── .env.example
├── faiss_index/
└── README.md

Replace your_script_name.py with your actual script filename.

---

Requirements

• Python 3.8+
• Google Generative AI API Key

---

Notes

• For best results, upload clearly scanned PDFs with selectable text.
• The FAISS index is stored locally; delete the faiss_index folder to reset embeddings.
• Be mindful of Google API usage limits.

---

License

MIT License. See LICENSE for details.

---

Credits

• LangChain
• Streamlit
• FAISS

---

Troubleshooting

• Ensure your .env and API key are set up correctly.
• If you encounter package errors, check your Python version and dependency installations.
• For any issues, please open an issue on this repository.