Fix tests

bluecast/tests/test_cast.py

@@ -314,32 +314,29 @@ def transform(
     )  # due to custom model and fit method
 
 
-@pytest.fixture
-def bluecast_instance():
-    # Create a fixture to instantiate the BlueCast class with default values for testing
-    return BlueCast(class_problem="binary", target_column="target")
-
-
-def test_enable_feature_selection_warning(bluecast_instance, capsys):
+def test_enable_feature_selection_warning(capsys):
     # Test if a warning is raised when feature selection is disabled
     df = pd.DataFrame({"feature1": [1, 2, 3], "target": [0, 1, 0]})
+    bluecast_instance = BlueCast(class_problem="binary", target_column="target")
     bluecast_instance.initial_checks(df)
     captured = capsys.readouterr()
     assert "Feature selection is disabled." in captured.err
 
 
-def test_hypertuning_cv_folds_warning(bluecast_instance, capsys):
+def test_hypertuning_cv_folds_warning(capsys):
     # Test if a warning is raised when hypertuning_cv_folds is set to 1
     df = pd.DataFrame({"feature1": [1, 2, 3], "target": [0, 1, 0]})
+    bluecast_instance = BlueCast(class_problem="binary", target_column="target")
     bluecast_instance.conf_training.hypertuning_cv_folds = 1
     bluecast_instance.initial_checks(df)
     captured = capsys.readouterr()
     assert "Cross validation is disabled." in captured.err
 
 
-def test_missing_feature_selector_warning(bluecast_instance, capsys):
+def test_missing_feature_selector_warning(capsys):
     # Test if a warning is raised when feature selection is enabled but no feature selector is provided
     df = pd.DataFrame({"feature1": [1, 2, 3], "target": [0, 1, 0]})
+    bluecast_instance = BlueCast(class_problem="binary", target_column="target")
     bluecast_instance.conf_training.enable_feature_selection = True
     bluecast_instance.initial_checks(df)
     captured = capsys.readouterr()
@@ -349,18 +346,20 @@ def test_missing_feature_selector_warning(bluecast_instance, capsys):
     )
 
 
-def test_missing_xgboost_tune_params_config_warning(bluecast_instance, capsys):
+def test_missing_xgboost_tune_params_config_warning(capsys):
     # Test if a warning is raised when XgboostTuneParamsConfig is not provided
     df = pd.DataFrame({"feature1": [1, 2, 3], "target": [0, 1, 0]})
+    bluecast_instance = BlueCast(class_problem="binary", target_column="target")
     bluecast_instance.conf_xgboost = None
     bluecast_instance.initial_checks(df)
     captured = capsys.readouterr()
     assert "No XgboostTuneParamsConfig has been provided." in captured.err
 
 
-def test_min_features_to_select_warning(bluecast_instance, capsys):
+def test_min_features_to_select_warning(capsys):
     # Test if a warning is raised when min_features_to_select is greater than or equal to the number of features
     df = pd.DataFrame({"feature1": [1, 2, 3], "target": [0, 1, 0]})
+    bluecast_instance = BlueCast(class_problem="binary", target_column="target")
     bluecast_instance.conf_training.enable_feature_selection = True
     bluecast_instance.conf_training.min_features_to_select = 3
     bluecast_instance.initial_checks(df)
@@ -371,9 +370,10 @@ def test_min_features_to_select_warning(bluecast_instance, capsys):
     )
 
 
-def test_shap_values_and_ml_algorithm_warning(bluecast_instance, capsys):
+def test_shap_values_and_ml_algorithm_warning(capsys):
     # Test if a warning is raised when calculate_shap_values is True and cat_encoding_via_ml_algorithm is True
     df = pd.DataFrame({"feature1": [1, 2, 3], "target": [0, 1, 0]})
+    bluecast_instance = BlueCast(class_problem="binary", target_column="target")
     bluecast_instance.conf_training.calculate_shap_values = True
     bluecast_instance.conf_training.cat_encoding_via_ml_algorithm = True
     bluecast_instance.initial_checks(df)
@@ -384,9 +384,10 @@ def test_shap_values_and_ml_algorithm_warning(bluecast_instance, capsys):
     )
 
 
-def test_cat_encoding_via_ml_algorithm_and_ml_model_warning(bluecast_instance, capsys):
+def test_cat_encoding_via_ml_algorithm_and_ml_model_warning(capsys):
     # Test if a warning is raised when cat_encoding_via_ml_algorithm is True and ml_model is provided
     df = pd.DataFrame({"feature1": [1, 2, 3], "target": [0, 1, 0]})
+    bluecast_instance = BlueCast(class_problem="binary", target_column="target")
     bluecast_instance.conf_training.cat_encoding_via_ml_algorithm = True
     bluecast_instance.ml_model = (
         bluecast_instance.ml_model
@@ -399,9 +400,10 @@ def test_cat_encoding_via_ml_algorithm_and_ml_model_warning(bluecast_instance, c
     )
 
 
-def test_precise_cv_tuning_warnings(bluecast_instance, capsys):
+def test_precise_cv_tuning_warnings(capsys):
     # Test if warnings are raised for precise_cv_tuning conditions
     df = pd.DataFrame({"feature1": [1, 2, 3], "target": [0, 1, 0]})
+    bluecast_instance = BlueCast(class_problem="binary", target_column="target")
     bluecast_instance.conf_training.precise_cv_tuning = True
     bluecast_instance.initial_checks(df)
     captured = capsys.readouterr()
@@ -416,7 +418,7 @@ def test_precise_cv_tuning_warnings(bluecast_instance, capsys):
     )
 
 
-def test_class_problem_mismatch_warnings(bluecast_instance, capsys):
+def test_class_problem_mismatch_warnings(capsys):
     # Test if warnings are raised for class problem mismatch
     df_binary = pd.DataFrame({"feature1": [1, 2, 3], "target": [0, 1, 0]})
     df_multiclass = pd.DataFrame({"feature1": [1, 2, 3], "target": [0, 1, 2]})