Fix test_sklearn.py

tests/test_sklearn.py

@@ -128,45 +128,30 @@ def test_fit_with_params(
         objective=objective,
     )
 
+    # See https://github.com/microsoft/LightGBM/issues/2328
     if boosting_type == "rf" and callable(objective):
-        # See https://github.com/microsoft/LightGBM/issues/2328
         with pytest.raises(lgb.basic.LightGBMError):
             clf.fit(X, y)
     else:
         clf.fit(X, y)
 
 
-@pytest.mark.parametrize("callbacks", [None, [callback]])
-@pytest.mark.parametrize("eval_metric", [None, "auc", zero_one_loss])
-def test_fit_with_fit_params(
-    callbacks: Optional[List[Callable]], eval_metric: Union[Callable, str]
-) -> None:
-    X, y = load_breast_cancer(return_X_y=True)
-
-    clf = OGBMClassifier(n_estimators=n_estimators, n_trials=n_trials)
-
-    clf.fit(X, y, callbacks=callbacks, eval_metric=eval_metric)
-
-
-def test_fit_with_unused_fit_params() -> None:
-    X, y = load_breast_cancer(return_X_y=True)
-
-    clf = OGBMClassifier(n_estimators=n_estimators, n_trials=n_trials)
-
-    clf.fit(X, y, eval_set=None)
-
-
-def test_fit_with_group_k_fold() -> None:
+def test_fit_with_empty_param_distributions() -> None:
     X, y = load_breast_cancer(return_X_y=True)
 
     clf = OGBMClassifier(
-        cv=GroupKFold(5), n_estimators=n_estimators, n_trials=n_trials
+        colsample_bytree=0.1,
+        n_estimators=n_estimators,
+        n_trials=n_trials,
+        param_distributions={},
     )
 
-    n_samples, _ = X.shape
-    groups = np.random.choice(10, size=n_samples)
+    clf.fit(X, y)
 
-    clf.fit(X, y, groups=groups)
+    df = clf.study_.trials_dataframe()
+    values = df["value"]
+
+    assert values.nunique() == 1
 
 
 def test_fit_with_pruning() -> None:
@@ -186,22 +171,37 @@ def test_fit_with_pruning() -> None:
     assert len(pruned_trials) > 0
 
 
-def test_fit_with_empty_param_distributions() -> None:
+@pytest.mark.parametrize("callbacks", [None, [callback]])
+@pytest.mark.parametrize("eval_metric", [None, "auc", zero_one_loss])
+def test_fit_with_fit_params(
+    callbacks: Optional[List[Callable]], eval_metric: Union[Callable, str]
+) -> None:
+    X, y = load_breast_cancer(return_X_y=True)
+
+    clf = OGBMClassifier(n_estimators=n_estimators, n_trials=n_trials)
+
+    clf.fit(X, y, callbacks=callbacks, eval_metric=eval_metric)
+
+
+def test_fit_with_unused_fit_params() -> None:
+    X, y = load_breast_cancer(return_X_y=True)
+
+    clf = OGBMClassifier(n_estimators=n_estimators, n_trials=n_trials)
+
+    clf.fit(X, y, eval_set=None)
+
+
+def test_fit_with_group_k_fold() -> None:
     X, y = load_breast_cancer(return_X_y=True)
 
     clf = OGBMClassifier(
-        colsample_bytree=0.1,
-        n_estimators=n_estimators,
-        n_trials=n_trials,
-        param_distributions={},
+        cv=GroupKFold(5), n_estimators=n_estimators, n_trials=n_trials
     )
 
-    clf.fit(X, y)
-
-    df = clf.study_.trials_dataframe()
-    values = df["value"]
+    n_samples, _ = X.shape
+    groups = np.random.choice(10, size=n_samples)
 
-    assert values.nunique() == 1
+    clf.fit(X, y, groups=groups)
 
 
 @pytest.mark.parametrize("n_jobs", [-1, 1])