fix: Fixing the random state of ElasticNetClassifier by default, to ensure reproduciblity. Also included elasticnet in reproducibility tests (#1374)

Co-authored-by: Daniel Grindrod <Daniel.Grindrod@evotec.com>
Co-authored-by: Li Jiang <bnujli@gmail.com>

flaml/automl/model.py

@@ -2429,6 +2429,11 @@ class ElasticNetEstimator(SKLearnEstimator):
 
     def __init__(self, task="regression", **config):
         super().__init__(task, **config)
+        self.params.update(
+            {
+                "random_state": config.get("random_seed", 10242048),
+            }
+        )
         assert self._task.is_regression(), "ElasticNet for regression task only"
         self.estimator_class = ElasticNet
 

test/automl/test_regression.py

@@ -236,6 +236,7 @@ def test_multioutput():
     "estimator",
     [
         "catboost",
+        "enet",
         "extra_tree",
         "histgb",
         "kneighbor",
@@ -342,6 +343,7 @@ def test_reproducibility_of_catboost_regression_model():
     "estimator",
     [
         "catboost",
+        "enet",
         "extra_tree",
         "histgb",
         "kneighbor",
@@ -385,7 +387,6 @@ def test_reproducibility_of_underlying_regression_models(estimator: str):
             automl._state.X_train_all, automl._state.y_train_all, automl._state.kf, best_model.model, "regression"
         )
     )
-
     assert pytest.approx(val_loss_flaml) == reproduced_val_loss_underlying_model