An AI-powered smart chat application developed as a graduation project. It combines a large language model with custom sentiment analysis and writing-style adaptation models to produce responses that match the tone and mood of the conversation.
- 💬 Conversational chat powered by an LLM (OpenAI API)
- 😊 Sentiment analysis — classical ML and a BERT-based model
- ✍️ Style adapter — adjusts the assistant's writing style to the user
- 🔐 User accounts with bcrypt password hashing
- 🗄️ Persistence with MySQL
- 🐳 Dockerized with docker-compose
- Python · Flask (REST backend) · Gunicorn
- OpenAI API
- scikit-learn / BERT for sentiment & style models
- MySQL
- Docker · docker-compose
| Path | Description |
|---|---|
backend/ |
Flask API (app.py) and services |
ai_module/ |
AI / model integration layer |
*.ipynb |
Model training notebooks (sentiment, BERT, style) |
*.pkl |
Trained model artifacts |
docs/, raporlar/ |
Project documentation & reports |
pip install -r requirements.txt
# Configure your OpenAI API key and MySQL connection, then:
python backend/app.pyOr run the full stack with Docker:
docker-compose up --buildOpen source, developed for academic purposes.