from fastapi import FastAPI                        # The web framework (builds your API)
from sklearn.datasets import load_iris             # The iris flower dataset
from sklearn.tree import DecisionTreeClassifier    # The ML model
import numpy as np                                 # For handling numbers

app = FastAPI()

# This runs ONCE when the server starts — trains the model immediately
iris = load_iris()
model = DecisionTreeClassifier(random_state=42)
model.fit(iris.data, iris.target)
class_names = ["setosa", "versicolor", "virginica"]

@app.get("/health")          # Visit /health → tells you the server is alive
async def health():
    
    return {"status": "ok"}

@app.get("/predict")         # Visit /predict?sl=4.8&sw=4&pl=3.6&pw=0.8 → get prediction
async def predict(sl: float, sw: float, pl: float, pw: float):
    features = np.array([[sl, sw, pl, pw]])
    pred = int(model.predict(features)[0])
    return {"prediction": pred, "class_name": class_names[pred]}