TaxBot: A RAG Application for Indian Taxation

TaxBot is a Retrieval-Augmented Generation (RAG) application designed to assist users with queries related to Indian taxation. It combines state-of-the-art machine learning models and retrieval techniques to deliver accurate and context-aware responses.

Features

• RAG Architecture: Combines retrieval and generation to enhance response accuracy.
• Domain-Specific Expertise: Focused on Indian taxation laws and policies.
• Technologies Used: Built using LangChain, ChromaDB, and Zephyr-7b-beta model from Hugging Face.
• Efficient Knowledge Retrieval: Integrates a custom-built knowledge base for precise information extraction.

Tech Stack

• LangChain: Framework for building applications powered by language models.
• ChromaDB: Vector database for managing and retrieving embeddings.
• Hugging Face Zephyr-7b-beta: A fine-tuned large language model optimized for Indian taxation.

Installation

Prerequisites

• Python 3.8+
• Pip package manager
• Virtual environment (optional but recommended)

Steps

1. Clone the repository:

   git clone https://github.com/pratikroy311/TAXBot.git
   cd taxbot

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate # For Linux/macOS
   venv\Scripts\activate   # For Windows

3. Install dependencies:

   pip install -r requirements.txt

4. Set up your environment variables:
   • Create a .env file in the root directory.
   • Add the required API keys and configuration settings. Example:

     OPENAI_API_KEY=your_openai_api_key
     CHROMADB_PATH=./data/chromadb

5. Run the application:

   python app.py

Usage

• Open your browser and navigate to the local server URL (e.g., http://127.0.0.1:5000).
• Input your taxation-related query, and TaxBot will provide detailed responses.

Directory Structure

TaxBot/
├── app.py               # Main application entry point
├── requirements.txt     # Python dependencies
├── data/                # ChromaDB data storage
├── models/              # Model-related files
├── utils/               # Helper functions and utilities
└── README.md            # Project documentation

Future Enhancements

• Expand the knowledge base to include more domains beyond taxation.
• Enhance multi-language support for Indian regional languages.
• Integrate with external APIs for live updates on taxation rules.

Contributing

Contributions are welcome! Please fork the repository and create a pull request with your changes.

License

This project is licensed under the MIT License.

Acknowledgements

• Hugging Face for the Zephyr-7b-beta model.
• LangChain for the framework support.
• ChromaDB for efficient vector database management.

NOTE: You can run this notebook as well as chatbot in google colab. As long as colab is running the model will be hosted and link can be shared with others