Update unit tests.

bluecast/tests/test_eval_metrics.py

@@ -7,6 +7,7 @@
     mean_squared_error,
     root_mean_squared_error,
 )
+from unittest.mock import patch
 
 from bluecast.evaluation.eval_metrics import (
     ClassificationEvalWrapper,
@@ -110,3 +111,39 @@ def test_regression_run_witouth_args(sample_data_regression):
     score = wrapper.regression_eval_func_wrapper(y_true, y_hat)
     expected_score = mean_squared_error(y_true, y_hat)
     assert score == expected_score
+
+
+def test_root_mean_squared_error_import():
+    with patch.dict('sys.modules', {'sklearn.metrics.root_mean_squared_error': None}):
+        with patch('sklearn.metrics.mean_squared_error') as mock_mse:
+            from sklearn.metrics import mean_squared_error as root_mean_squared_error
+            assert root_mean_squared_error is mock_mse
+
+
+def test_root_mean_squared_error_direct_import():
+    with patch.dict('sys.modules', {'sklearn.metrics.root_mean_squared_error': object()}):
+        from sklearn.metrics import root_mean_squared_error
+        assert root_mean_squared_error is not None
+        assert callable(root_mean_squared_error)  # assuming root_mean_squared_error is callable
+
+
+def test_classification_eval_func_wrapper_invalid_eval_against():
+    # Instantiate the class with an invalid eval_against value
+    def dummy_metric_func(y_true, y_pred, **kwargs):
+        # Dummy metric function for testing purposes
+        return 0.5
+
+    class TestClass(ClassificationEvalWrapper):
+        def __init__(self, eval_against, metric_func):
+            super().__init__(eval_against=eval_against, metric_func=metric_func)
+
+    test_obj = TestClass(eval_against="probas_all_classes", metric_func=dummy_metric_func)
+    test_obj.eval_against = "invalid_value"
+
+    # Dummy data for testing
+    y_true = [1, 0, 1]
+    y_probs = [[0.4, 0.6], [0.7, 0.3], [0.2, 0.8]]
+
+    # Use pytest to check if ValueError is raised
+    with pytest.raises(ValueError, match=r"Unknown value for eval_against: invalid_value\. Possible values are 'probas' or 'classes'"):
+        test_obj.classification_eval_func_wrapper(y_true, y_probs)