CytoLens

CytoLens is a breast cancer predictor for cancer malignancy in breast tissue samples meant to be used in conjunction with measurements from a cytology lab. Uses a Logistic Regression model with an interactive radar chart visualization.

Built with FastAPI (backend) + Svelte (frontend), served from a single Docker container.

Thank you to Alejandro AO.

Development

Backend

cd backend
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
uvicorn app.main:app --reload --port 8000

Frontend

cd frontend
npm install
npm run dev

The frontend dev server runs on port 5173 and proxies /api requests to localhost:8000.

Retrain Model

cd backend
source .venv/bin/activate
python3 train.py

Docker

docker build -t cytolens .
docker run -p 8000:8000 cytolens

Open http://localhost:8000 to use the app.