RAG Application using Gemini Flash

A Retrieval-Augmented Generation (RAG) application built with LangChain, Google's Gemini Flash, and Streamlit that allows users to ask questions about PDF documents and get AI-powered answers based on the document content.

Features

• 📄 PDF Document Processing: Automatically loads and processes PDF documents
• 🔍 Intelligent Chunking: Splits documents into manageable chunks for better retrieval
• 🧠 Vector Embeddings: Uses Google's embedding model for semantic search
• 🚀 Fast Retrieval: FAISS vector store for efficient similarity search
• 💬 Interactive Chat: Streamlit-powered chat interface
• ⚡ Powered by Gemini: Uses Google's Gemini 2.0 Flash model for responses

Prerequisites

• Python 3.8 or higher
• Google API Key (for Gemini Flash and embeddings)
• Ubuntu/Linux system (for installation instructions)

Installation

1. System Dependencies

sudo apt update && sudo apt install -y python3 python3-pip python3-venv python3-dev build-essential sqlite3 libsqlite3-dev

2. Virtual Environment Setup

# Create virtual environment
python3 -m venv ai_project_env

# Activate virtual environment
source ai_project_env/bin/activate

# Upgrade pip
pip install --upgrade pip

3. Install Python Dependencies

# Install all required packages
pip install pysqlite3-binary python-dotenv streamlit sentence-transformers unstructured langchain langchain-community langchain-google-genai langchain-experimental langchainhub faiss-cpu pypdf

Configuration

1. Environment Variables

Create a .env file in the project root:

GOOGLE_API_KEY=your_google_api_key_here

To get a Google API key:

1. Go to Google AI Studio
2. Create or select a project
3. Generate an API key
4. Copy the key to your .env file

2. PDF Document

Place your PDF document in the project root and name it source_file.pdf, or modify the filename in the code:

loader = PyPDFLoader("your_document.pdf")

Usage

1. Activate Virtual Environment

source ai_project_env/bin/activate

2. Run the Application

streamlit run app.py

3. Access the Application

Open your browser and navigate to http://localhost:8501

4. Ask Questions

Use the chat input at the bottom of the page to ask questions about your PDF document.

Project Structure

rag-gemini/
├── ai_project_env/          # Virtual environment
├── app.py                   # Main application file
├── .env                     # Environment variables (create this)
├── my_paper.pdf            # Your PDF document
├── requirements.txt        # Python dependencies (optional)
└── README.md              # This file

How It Works

1. Document Loading: PyPDFLoader extracts text from the PDF document
2. Text Splitting: RecursiveCharacterTextSplitter breaks the document into chunks
3. Embeddings: Google's embedding model converts text chunks into vectors
4. Vector Storage: FAISS stores and indexes the embeddings for fast retrieval
5. Query Processing: User questions are embedded and matched against document chunks
6. Answer Generation: Gemini Flash generates contextual answers based on retrieved chunks

Customization

Modify Chunk Size

text_splitter = RecursiveCharacterTextSplitter(chunk_size=1500)  # Adjust chunk size

Change Retrieval Parameters

retriever = vectorstore.as_retriever(
    search_type="similarity", 
    search_kwargs={"k": 5}  # Adjust number of retrieved chunks
)

Modify System Prompt

system_prompt = (
    "Your custom system prompt here..."
    "{context}"
)

Switch Gemini Model

llm = ChatGoogleGenerativeAI(
    model="gemini-2.0-flash",  # or "gemini-pro"
    temperature=0.3,     # Adjust creativity
    max_tokens=1000      # Adjust response length
)

Troubleshooting

Common Issues

1. SQLite3 Version Error The code includes a fix for SQLite3 version conflicts. If you still encounter issues:

import sys
try:
    import pysqlite3 as sqlite3
    sys.modules['sqlite3'] = sqlite3
except ImportError:
    import sqlite3

2. Google API Key Error

• Ensure your .env file is in the project root
• Verify your API key is valid and has the necessary permissions
• Check that you've enabled the Gemini API in Google Cloud Console

3. PDF Loading Error

• Ensure your PDF file exists in the specified path
• Try with a different PDF if the current one is corrupted
• Install additional PDF processing libraries if needed:

pip install pymupdf  # Alternative PDF reader

4. Memory Issues For large documents, consider:

• Reducing chunk size
• Limiting the number of retrieved chunks (k parameter)
• Using a more powerful machine

Dependencies

Core Libraries

• langchain: LLM application framework
• langchain-community: Community components
• langchain-google-genai: Google Gemini integration
• streamlit: Web application framework
• faiss-cpu: Vector similarity search
• python-dotenv: Environment variable management

Supporting Libraries

• pypdf: PDF processing
• sentence-transformers: Text embeddings
• pysqlite3-binary: SQLite3 compatibility

Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add some amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

License

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

Acknowledgments

• LangChain for the RAG framework
• Google AI for Gemini Pro and embedding models
• Streamlit for the web interface
• FAISS for vector search

Support

If you encounter any issues or have questions:

1. Check the troubleshooting section above
2. Open an issue on GitHub
3. Refer to the LangChain documentation
4. Check Google AI documentation

---