Exam Review Bot

Exam Review Bot (ERB) is a custom intelligent retrieval and hybrid RAG system for final exam preparation. Specifically, it

• Extracts specific lecture PDFs from your course schedule.

• Matches them to user questions (e.g., "What about transformers was discussed?").

• Supplements with internet knowledge (external search).

• Summarizes and explain the content intelligently.

Higher-level Architecture

[Your Documents: PDFs, Ed Posts, Images] ↓ [Loader modules] ↓ [Text Chunker (split large docs)] ↓ [Embedder (create vector embeddings)] ↓ [Vector Store (Chroma, FAISS, or similar)] ↓ [Retriever (search the database)] ↓ [LLM (OpenAI, Claude, or Local Model)] ↓ [Answer to your Question]

Project Setup

Prerequisites

• Python 3.8 or higher
• Node.js 16 or higher
• npm 7 or higher

Backend Setup

1. Create a virtual environment:

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

2. Install dependencies:

   pip install -r requirements.txt

3. Create a .env file in the backend directory:

   OPENAI_API_KEY=your_openai_api_key
   

Frontend Setup

1. Navigate to the frontend directory:

   cd frontend

2. Install dependencies:

   npm install

Running the Demo

Starting the Application

1. Start both frontend and backend concurrently:

   cd frontend
   npm run dev:all

   This will start:
   • Frontend at http://localhost:5173
   • Backend at http://localhost:8000

Using the Demo

1. Uploading Documents

1. Navigate to the Demo page using the navigation bar
2. In the "Upload Document" section:
   • Click "Choose File" to select a PDF document
   • Click "Upload" to process the document
   • Wait for the upload confirmation
   • The document will appear in the "Uploaded Documents" list

2. Chatting with AI

1. In the "Chat with AI" section:
   • Type your question in the text field
   • Click "Send" or press Enter
   • Wait for the AI's response
   • The conversation history will be displayed below

Features

• Document Management:

  • Upload PDF documents
  • View list of uploaded documents
  • Automatic document processing and indexing

• AI Chat Interface:

  • Real-time responses
  • Context-aware answers
  • Source references
  • Conversation history

• User Experience:

  • Loading indicators
  • Error handling
  • Responsive design
  • Real-time updates

Troubleshooting

Common Issues

1. Upload Fails:

   • Check file size (max 10MB)
   • Ensure file is PDF format
   • Check backend logs for errors

2. Chat Not Working:

   • Verify OpenAI API key in .env
   • Check internet connection
   • Ensure backend is running

3. Frontend Not Loading:

   • Clear browser cache
   • Check console for errors
   • Verify all dependencies are installed

Getting Help

• Check the console for error messages
• Review backend logs
• Ensure all services are running
• Verify environment variables

Development

Backend Development

• The backend is built with FastAPI
• API documentation is available at /docs when running the server
• Use uvicorn main:app --reload for development

Frontend Development

• The frontend uses Vite as the build tool
• Development server runs on http://localhost:5173 by default
• Hot Module Replacement (HMR) is enabled for faster development

To-Be-Implemented

• Custom Indexing + Custom Vectorstore database
• Hybrid RAG system
• File upload interface
• Chat interface
• Progress tracking
• User authentication

To-Dos

• Setting up frontend
• Backend frontend integration API service layer
• Get frontend + backend demo working PERFECTLY
• Figure out how to insert interative demo in personal website