Regression Metrics: MAE, RMSE and R²

Implement regression_report(y_true, y_pred) returning a dict with:

• mae: mean absolute error
• rmse: root mean squared error
• r2: coefficient of determination, $1 - \frac{SS_{res}}{SS_{tot}}$

If y_true is constant, SS_tot is 0: return r2 = 1.0 for a perfect prediction and 0.0 otherwise.