Automated Testing for AI Systems

# tests/test_features.py
import pytest
from src.features import normalize_petal

def test_normalize_petal_range():
    assert normalize_petal(0.0) == 0.0
    assert normalize_petal(10.0) == 1.0

def test_normalize_petal_negative_raises():
    with pytest.raises(ValueError):
        normalize_petal(-1.0)

import pandera as pa
from pandera import Column, Check

iris_schema = pa.DataFrameSchema({
    "sepal_length": Column(float, Check.in_range(4.0, 8.0)),
    "sepal_width":  Column(float, Check.in_range(2.0, 4.5)),
    "petal_length": Column(float, Check.in_range(1.0, 7.0)),
    "petal_width":  Column(float, Check.in_range(0.1, 2.5)),
    "species":      Column(str,   Check.isin(["setosa", "versicolor", "virginica"])),
})

def test_training_data_schema(train_df):
    iris_schema.validate(train_df)  # fail fast on bad data

import joblib, numpy as np

model = joblib.load("models/iris.pkl")

def test_baseline_accuracy(test_df):
    acc = (model.predict(test_df.X) == test_df.y).mean()
    assert acc >= 0.90, f"Accuracy dropped: {acc:.3f}"

def test_invariance_to_irrelevant_noise():
    # small perturbation should not change prediction
    x = np.array([[5.1, 3.5, 1.4, 0.2]])
    base = model.predict(x)[0]
    noisy = model.predict(x + 1e-3)[0]
    assert base == noisy

def test_directional_petal_length():
    # larger petal → less likely setosa
    small = model.predict_proba([[5.0, 3.5, 1.0, 0.2]])[0][0]
    large = model.predict_proba([[5.0, 3.5, 5.0, 0.2]])[0][0]
    assert large < small

from fastapi.testclient import TestClient
from src.main import app

client = TestClient(app)

def test_predict_endpoint_contract():
    r = client.post("/predict", json={
        "sepal_length": 5.1, "sepal_width": 3.5,
        "petal_length": 1.4, "petal_width": 0.2,
    })
    assert r.status_code == 200
    body = r.json()
    assert "prediction" in body
    assert "model_version" in body
    assert body["prediction"] in {"setosa", "versicolor", "virginica"}

def test_predict_rejects_invalid_input():
    r = client.post("/predict", json={"sepal_length": "oops"})
    assert r.status_code == 422