RAG Chatbot

Overview

This project implements a standalone Question Answering (QA) chatbot using Retrieval-Augmented Generation (RAG). The chatbot uses a combination of local documents as its knowledge base to provide dynamic and contextually relevant responses.

Features

• RAG-based response generation for dynamic answers
• Ability to process and use information from various file types (txt, pdf, json, csv, html, md)
• Uses HuggingFace models for language understanding and generation
• Easy-to-use command-line interface for testing and interaction

Prerequisites

• Python 3.7+
• pip (Python package installer)

Setup

1. Clone the repository:

   git clone https://github.com/brandongraves08/ai_project.git
   cd ai_project
   

2. Set up a virtual environment (optional but recommended):

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`
   

3. Install dependencies:

   pip install -r requirements.txt
   

4. Prepare your data:

   • Create a data directory in the project root if it doesn't exist
   • Place your document files (pdf, txt, json, csv, html, md) in the data directory

Usage

1. Run the chatbot:

   python rag_chatbot.py
   

2. Once initialized, you can start asking questions. Type your questions and press Enter.

3. To exit the chatbot, type 'quit' and press Enter.

Project Structure

• rag_chatbot.py: Main script implementing the RAG-based chatbot
• data/: Directory for storing document files (not included in the repository)
• requirements.txt: List of Python dependencies
• README.md: This file, containing project documentation

Customization

• To customize the knowledge base, add or remove documents from the data directory
• To use a different language model, modify the model_name parameter in the RAGChatbot initialization

How It Works

1. The chatbot loads and processes documents from the data directory
2. It uses a text splitter to break documents into smaller chunks
3. These chunks are embedded and stored in a vector database (Chroma)
4. When a question is asked, the most relevant chunks are retrieved
5. The language model generates a response based on the retrieved information

Troubleshooting

• If you see a message about "Number of requested results is greater than number of elements in index", ensure that your data directory contains multiple documents with sufficient content.
• You may need to adjust the chunk_size in the RecursiveCharacterTextSplitter if you're working with very small documents.

Limitations

• The quality of responses depends on the documents provided in the data directory
• The chatbot may occasionally generate incorrect or inconsistent responses
• It doesn't maintain conversation context between questions

Future Improvements

• Implement caching to improve response time
• Add support for multi-turn conversations
• Implement user feedback mechanisms for continuous improvement
• Add support for online resources as part of the knowledge base
• Integrate Jira wiki support: Future versions will include the ability to connect to and retrieve information from Jira wikis, expanding the knowledge base to include up-to-date company documentation.
• Enhance RAG capabilities: We plan to improve the Retrieval-Augmented Generation system to provide more accurate and context-aware responses, possibly by incorporating more advanced language models and retrieval techniques.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request to the GitHub repository.

License

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