JayBot

A cloud-based, AI-driven tutoring platform that leverages Google Cloud, Pinecone, and GPT-4 to help students learn course content efficiently.

Table of Contents

• Project Overview
• Features
• Installation
• Usage
• Database Structure
• Flask API Endpoints
• Future Development
• License

Project Overview

The AI Tutor Platform provides a customized learning experience using AI. It allows proctors to upload course documents, processes them to understand content, and enables students to ask questions on the material.

Features

• Cloud-based AI tutoring with OpenAI GPT-4
• Google Cloud Integration for file storage and database management
• Pinecone Integration for vector comparisons and quick course context
• Secure access and role-based management with MySQL

Installation

1. Install XAMPP
2. Clone the repository in XAMPP's htdocs folder:

   cd C:\xampp\htdocs
   git clone https://github.com/yourusername/yourrepository.git
   cd yourrepository

3. Install the dependencies:

   cd setup
   python setup.py

4. Add gcloud_keys folder to backend
5. Set up environment variables in a .env file for Python backend:

   OPENAI_API_KEY=your_openai_api_key
   PINECONE_API_KEY=your_pinecone_api_key
   GOOGLE_APPLICATION_CREDENTIALS="path/to/your/credentials.json"
   DB_HOST=your_host
   DB_NAME=your_db
   DB_USER=your_user
   DB_PASS=your_password
   DB_PORT=your_port
   

6. Add db_config.php to /backend/includes/ and add environment variables for PHP frontend:

    <?php
    $host = your_host;
    $dbUsername = your_username;
    $dbPassword = your_password;
    $database = your_database;
    $port = your_port;
    ?>

7. Add db_connect.php to /backend/includes/ and add database connection instructions:

   <?php
   require_once "db_config.php"; 
   
   //Create database connection
   $connection = new mysqli($host, $dbUsername, $dbPassword, $database, $port);
   
   //Check if the connection was successful
   if ($connection->connect_error) {
       die("Connection failed: ".$connection->connect_error);
   }
   ?>

8. Turn on XAMPP's MySQL and initialize database:

   cd backend/database/
   python initializeTables.py

9. (Bonus Step - Recommended) Create a proctor account:
   Sign up with proctor username and password on the site and then run the following:

   cd backend/database/
   python createProctor.py <your_username_here>

Usage

1. Run the Flask Application:

  cd backend
  python app.py

2. Start Apache and MySQL on XAMPP Control Panel
   
3. Navigate here, accessible in your browser.

Database Structure

The platform includes the following database tables:

• Users - Stores id, username, password, and role
• Courses - Stores id, name, createdBy, filepath, courseCode, and description
• User_Courses - Stores userCoursesId, userId, courseId, interest, responseLength, and archived
• Messages - Stores messageId, userCoursesId, question, answer, timestamp, sourceName, feedbackRating, feedbackExplanation, feedbackTimestamp

Flask API Endpoints

• /upload - Uploads a single file to the databases
• /load-docs - Loads all files for a specified course
• /download - Downloads a specified file from GCS
• /delete - Deletes a specified file from GCS and Pinecone
• /delete-course - Deletes all files associated with a specified course from GCS and Pinecone
• /ask-question - Generates AI response based on course content, chat history, and user question
• /generate-report - Generates a report of student questions, AI answers, and student feedback

Future Development

• JayWing Academy Implementation
• More LLM Experimentation
• Advanced RAG Techniques

License

This project is licensed under the MIT License.