Expose parametric in bootstrap_stats and improve relevant test

stepmix/stepmix.py

@@ -1150,7 +1150,13 @@ def bootstrap(
         )
 
     def bootstrap_stats(
-        self, X, Y=None, n_repetitions=1000, sample_weight=None, progress_bar=True
+        self,
+        X,
+        Y=None,
+        n_repetitions=1000,
+        sample_weight=None,
+        parametric=False,
+        progress_bar=True,
     ):
         """Non-parametric boostrap of StepMix estimator. Obtain boostrapped parameters and some statistics
         (mean and standard deviation).
@@ -1164,6 +1170,8 @@ def bootstrap_stats(
         sample_weight : array-like of shape(n_samples,), default=None
         n_repetitions: int
             Number of repetitions to fit.
+        parametric: bool, default=False
+            Use parametric bootstrap instead of non-parametric. Data will be generated by sampling the estimator.
         progress_bar : bool, default=True
             Display a tqdm progress bar for repetitions.
         Returns
@@ -1185,7 +1193,7 @@ def bootstrap_stats(
             Y=Y,
             n_repetitions=n_repetitions,
             sample_weight=sample_weight,
-            parametric=False,
+            parametric=parametric,
             progress_bar=progress_bar,
         )
 

test/test_bootstrap.py

@@ -83,7 +83,8 @@ def test_nested_permutation(data_nested, kwargs_nested):
     "ignore::sklearn.exceptions.ConvergenceWarning"
 )  # Ignore convergence warnings for same reason
 @pytest.mark.parametrize("model", EMISSION_DICT.keys())
-def test_bootstrap(data, kwargs, model):
+@pytest.mark.parametrize("parametric", [False, True])
+def test_bootstrap(data, kwargs, model, parametric):
     """Call the boostrap procedure on all models and make sure they don't raise errors.
 
     The data may not make sense for the model. We therefore do not test a particular output here."""
@@ -100,12 +101,12 @@ def test_bootstrap(data, kwargs, model):
     model_1 = StepMix(n_steps=1, **kwargs)
     model_1.fit(X, Y)
 
-    if model != "covariate":
-        model_1.bootstrap_stats(X, Y, n_repetitions=3)
-    else:
-        # Should raise error. Can't sample from a covariate model
+    if parametric and model == "covariate":
+        # Parametric bootstrap is not implemented for covariate model
         with pytest.raises(NotImplementedError) as e_info:
-            model_1.bootstrap_stats(X, Y, n_repetitions=3)
+            model_1.bootstrap_stats(X, Y, parametric=parametric, n_repetitions=3)
+    else:
+        model_1.bootstrap_stats(X, Y, parametric=parametric, n_repetitions=3)
 
 
 def test_nested_bootstrap(data_nested, kwargs_nested):