Josh Dula - CS 4800 Capstone Project

An intelligent study tool that automatically generates summaries, flashcards, and AI-powered quiz questions from your documents.

This repository is ready to run immediately. Follow these steps:

git clone https://github.com/jdula02/ai-study-tool.git
cd ai-study-tool

# Create virtual environment
python3 -m venv venv

# Activate it (Linux/WSL/Mac)
source venv/bin/activate

# Or on Windows
# venv\Scripts\activate

pip install -r requirements.txt

Note: If you get an "externally-managed-environment" error, make sure you're using the virtual environment above.

python -c "import nltk; nltk.download('punkt'); nltk.download('stopwords'); nltk.download('averaged_perceptron_tagger')"

Standard:

python simple_app.py

For WSL Users:

python run_wsl_server.py

Open your browser to: http://localhost:8000

WSL users may need to use: http://172.21.19.74:8000

• ** Key**: Claude AI key is already configured in .env
• ** Sample Data**: Database includes uploaded documents for immediate testing
• ** Sample Documents**: PDF and text files ready for upload testing
• ** All Dependencies**: Complete requirements.txt with exact versions

• Upload (/upload) - Drag-and-drop file upload interface
• Documents (/documents) - View all uploaded files with action buttons
• Quiz (/quiz) - NEW: AI-powered quiz generation
• Study (/study) - Spaced repetition flashcard system
• AI Tutor (/ai-tutor) - Interactive chat with Claude AI

1. Go to /quiz
2. Select any document
3. Enable "Use AI for smarter questions" (checkbox)
4. Choose number of questions (5-20)
5. Take the quiz with proper answer feedback
6. View detailed results with score percentage and question review
7. Retake the same quiz or start a new one

ai-study-tool/
├── simple_app.py              # Main FastAPI application
├── simple_config.py           # Configuration
├── .env                       # API keys (included for evaluation)
├── study_tool.db             # SQLite database with sample data
├── backend/
│   ├── api/                   # REST API endpoints
│   │   ├── simple_questions.py    # Question generation API
│   │   ├── ai_tutoring.py         # Claude AI integration
│   │   └── ...
│   └── services/              # Core business logic
│       ├── claude_ai_service.py   # AI service integration
│       ├── simple_question_generator.py # Basic questions
│       └── ...
├── frontend/
│   ├── templates/             # HTML pages
│   └── static/               # CSS/JS assets
├── uploads/                  # Sample uploaded files (included)
└── requirements.txt          # All dependencies

• Backend: FastAPI (Python 3.12)
• Database: SQLite with SQLAlchemy ORM
• AI Integration: Claude API (Anthropic)
• NLP: NLTK, Sumy, PyPDF2
• Frontend: HTML5, CSS3, Vanilla JavaScript
• Features: PDF processing, text summarization, spaced repetition

1. Document Processing: Upload PDFs/text files with automatic text extraction
2. Multiple Summarization: Extractive, abstractive, and hybrid approaches
3. Question Generation:
   • Basic: Rule-based fill-in-blank, multiple choice, true/false
   • AI-Powered: Intelligent questions that test comprehension vs. memorization
4. Flashcard System: Spaced repetition algorithm (Anki-inspired)
5. AI Tutoring: Interactive Q&A with document context

The repository includes sample documents and a populated database so you can immediately test all features without uploading new content.

Port 8000 in use?

# Change port in simple_app.py or kill existing process
pkill -f simple_app.py

Missing NLTK data?

python -c "import nltk; nltk.download('all')"

AI features not working?

• Claude AI API key is included in .env file
• Check console output for API errors
• If you see "Client.init() got an unexpected keyword argument 'proxies'", upgrade Anthropic:

pip install --upgrade anthropic

Virtual Environment Issues?

• Make sure you activated the virtual environment: source venv/bin/activate
• If you see "externally-managed-environment", you must use a virtual environment

WSL Network Issues?

• Try the WSL-specific server: python run_wsl_server.py
• Access via the IP shown in console output

1. Upload Feature: Try uploading a new PDF via /upload
2. Basic Questions: Generate questions from any document (uncheck AI option)
3. AI Quiz: Enable AI toggle and take a complete quiz with scoring
4. Quiz Completion: Complete a full quiz to see the results screen and retake functionality
5. Flashcards: Create and review flashcards with spaced repetition
6. AI Chat: Ask questions about uploaded documents via /ai-tutor
7. AI Summaries: Generate different summary types (extractive, abstractive, hybrid)

This project demonstrates full-stack web development, AI integration, natural language processing, and educational technology concepts.

Contact: Josh Dula - dulajb@appstate.edu