🌍 Planet Classifier API

A FastAPI backend that serves a machine learning model to predict whether an object is likely to be a planet.
It takes 15 input parameters and returns a prediction.

📂 Project Structure

├── app.py # FastAPI app entrypoint
├── model_script.py # Model loading & prediction logic
├── artifacts/ # Trained model + scaler/encoder
├── requirements.txt # Dependencies
├── Procfile # Render start command
└── render.yaml # Optional Render config (IaC)

Local Dev

pyenv local 3.11.9

python -m venv venv source venv/bin/activate

pip install -r requirements.txt

Example

curl -X POST http://localhost:8000/predict \
-H "Content-Type: application/json" \
-d '{
  "param1": 0.5, "param2": 1.2, "param3": 0.3,
  "param4": 0.9, "param5": 1.7, "param6": 2.1,
  "param7": 0.4, "param8": 3.2, "param9": 0.8,
  "param10": 1.9, "param11": 2.7, "param12": 1.0,
  "param13": 4.1, "param14": 0.6, "param15": 2.3
}'

should return {"prediction": "Planet"} with confience margins