You will build a feature selection + regression pipeline to predict disease progression using the Diabetes dataset. Your goal is to combine preprocessing, feature selection, and model training in one efficient workflow.

Follow these steps:

1. Load the dataset using load_diabetes().
2. Split it into train/test sets (test_size=0.3, random_state=42).
3. Build a pipeline with:
   • StandardScaler().
   • SelectFromModel(Lasso(alpha=0.01, random_state=42)) for automatic feature selection.
   • LinearRegression() as the final model.
4. Fit the pipeline and evaluate it using R² score on the test set.
5. Print:
   • The R² score (rounded to 3 decimals).
   • The number of features selected.