AI Chatbot for Student Data Management

Overview

This project is an AI-driven chatbot fine-tuned with the Gemini 1.5 Flash LLM model. It interacts with a dataset of fictional student information, answering questions about student profiles while handling both in-dataset and out-of-dataset queries. The chatbot is built with a FastAPI backend and a React frontend, ensuring seamless communication and efficient response handling. It also features chat history storage using MongoDB, with session persistence via local storage, and provides users with a dedicated history page for reviewing past interactions.

Features

• AI-Driven Responses: Fine-tuned with student data using Gemini 1.5 Flash LLM to provide precise answers to queries.
• FastAPI Backend: Efficient API handling and scalable architecture for seamless server-client communication.
• React Frontend: Interactive user interface with real-time communication and response rendering.
• MongoDB Integration: Stores chat history dynamically, allowing real-time retrieval and review of previous conversations.
• Local Storage: Persists chat history across sessions for better user experience and continuity.
• Dedicated History Page: Displays a complete log of past conversations for easy reference.
• In-Dataset and Out-of-Dataset Handling: Accurately manages both known and unknown queries with appropriate responses.

Tech Stack

• Backend: FastAPI, Python
• Frontend: React, JavaScript
• Database: MongoDB (for chat history)
• Model: Gemini 1.5 Flash LLM (fine-tuned with student data)
• Deployment: Google Cloud SDK, Google AI Studio
• Storage: Local Storage (for session persistence)

Setup and Installation

Prerequisites

Ensure you have the following installed:

• Python 3.7+
• Node.js and npm (for React frontend)
• MongoDB (for storing chat history)
• Google Cloud SDK (for model fine-tuning and deployment)
• FFmpeg (for Flask backend audio handling)
• Gemini API Key and Model ID

Backend Setup

1. Clone the repository:

   git clone https://github.com/GauravOOO2/Fine_Tuned_Student_Data_AI_ChatBot_Backend.git
   cd Fine_Tuned_Student_Data_AI_ChatBot_Backend

2. Install dependencies:

   pip install -r requirements.txt

3. Set up environment variables:

   Create a .env file in the backend folder and add your credentials:

   GEMINI_API_KEY=<your-gemini-api-key>
   MONGODB_URI=<your-mongodb-uri>

4. Run the FastAPI server:

   uvicorn main:app --reload

Frontend Setup

1. Navigate to the frontend folder:

   git clone https://github.com/GauravOOO2/Fine_Tuned_Student_Data_AI_ChatBot_FrontEnd.git
   

      cd Fine_Tuned_Student_Data_AI_ChatBot_FrontEnd

2. Install dependencies:

   npm install

3. Start the React development server:

   npm start

MongoDB Setup

1. Ensure MongoDB is running locally or provide a connection URI. The chat history is stored dynamically in MongoDB.
2. Set the MongoDB URI in the backend's .env file.

Model Fine-Tuning

1. Fine-tune the Gemini 1.5 Flash LLM using Google Cloud SDK and AI Studio. The dataset used contains fictional student information in JSON format.

2. Model ID and API Key should be included in the .env file.

Usage

Once both the backend and frontend servers are running, you can access the chatbot by navigating to:

http://localhost:3000

1. Chat with the Bot: Ask questions about students, and the chatbot will respond using the fine-tuned Gemini model.
2. Chat History: Check the "History" tab to view past conversations.
3. Session Persistence: The chatbot will remember past conversations using local storage.

Example Queries

• "What is the age of John Doe?"
• "Who is Jane Smith's oncologist?"
• "Which school does Alex attend?"
• "What is the percentage of Lily Thompson?"

For any out-of-dataset queries, the chatbot will gracefully handle them with appropriate fallback responses.

Project Structure

.
├── backend
│   ├── app
│   │   ├── main.py         # FastAPI backend implementation
│   │   ├── database.py     # MongoDB connection and schema
│   │   └── models.py       # Defines schema for chat history
├── frontend
│   ├── public              # Public files (index.html, etc.)
│   ├── src
│   │   ├── components      # React components
│   │   ├── pages           # Pages for chat and history
│   │   ├── App.js          # Main React component
│   │   └── index.js        # Entry point for React app
└── README.md

Future Enhancements

• Support for Multiple Datasets: Extend the chatbot’s ability to handle multiple datasets or domains.
• Enhanced Natural Language Understanding: Improve chatbot’s understanding and flexibility in answering more complex queries.
• User Authentication: Add user login and personalized chat history for individual users.

Contributing

Contributions are welcome! Please fork the repository and create a pull request with your changes. Ensure that your code follows the existing code style and includes relevant tests.

Contact

If you have any questions or suggestions, feel free to reach out:

• Gaurav Varma – LinkedIn
• GitHub – YourGitHubProfile