154 ruff apply unsafe fixes to unit test (#155)

Author: c-w-feldmann

test_extras/test_chemprop/test_models.py

@@ -75,7 +75,7 @@ def test_get_params(self) -> None:
         chemprop_model.set_params(**new_params)
         model_params = chemprop_model.get_params(deep=True)
         for param_name, param in new_params.items():
-            if param_name in ["model__agg"]:
+            if param_name in {"model__agg"}:
                 self.assertIsInstance(model_params[param_name], type(param))
                 continue
             self.assertEqual(param, model_params[param_name])

test_extras/test_notebooks/test_notebooks.py

@@ -52,7 +52,6 @@ def get_notebook_paths_from_dir(notebook_dir: Path) -> list[Path]:
     # Find all Jupyter notebook files in the directory
     notebooks_paths = []
     for notebook_path in notebook_dir.rglob("*.ipynb"):
-
         if ".ipynb_checkpoints" in str(notebook_path.resolve()):
             # skip jetbrains checkpoints
             continue
@@ -115,7 +114,6 @@ def run_notebooks(
     nof_errors = 0
     # Loop through each notebook
     for notebooks_path in notebooks_paths:
-
         # Execute the notebook and capture the error code
         cmd = [
             "jupyter",

tests/test_elements/test_mol2any/test_mol2concatenated.py

@@ -52,7 +52,6 @@ def test_generation(self) -> None:
         ]
 
         for fp_output_type in fingerprint_morgan_output_types:
-
             concat_vector_element = MolToConcatenatedVector(
                 [
                     (

tests/test_estimators/test_leader_picker_clustering.py

@@ -87,7 +87,6 @@ def test_leader_picker_pipeline(self) -> None:
         for dist, exp_labels, exp_centroids in zip(
             distances, expected_labels, expected_centroids
         ):
-
             leader_picker = LeaderPickerClustering(distance_threshold=dist)
             pipeline = Pipeline(
                 [

tests/test_estimators/test_murcko_scaffold_clustering.py

@@ -14,7 +14,7 @@
     "c1ccccc1",
 ]
 
-SCAFFOLD_SMILES_TEST_GENERIC: list[str] = SCAFFOLD_SMILES + ["c1ncccc1"]
+SCAFFOLD_SMILES_TEST_GENERIC: list[str] = [*SCAFFOLD_SMILES, "c1ncccc1"]
 
 LINEAR_SMILES: list[str] = ["CC", "CCC", "CCCN"]
 

tests/test_estimators/test_nearest_neighbors.py

@@ -154,7 +154,7 @@ def test_fit_predict_precomputed(self) -> None:
 
     def test_fit_predict_invalid(self) -> None:
         """Test the fit_predict method with invalid smiles."""
-        with_invald_smiles = ["CC1CC"] + TEST_SMILES
+        with_invald_smiles = ["CC1CC", *TEST_SMILES]
 
         error_filter = ErrorFilter(filter_everything=True)
         model = Pipeline(
@@ -174,14 +174,14 @@ def test_fit_predict_invalid(self) -> None:
             ]
         )
         result = model.fit_predict(with_invald_smiles, with_invald_smiles).tolist()
-        self.assertListEqual(result, [["invalid", "invalid"]] + TWO_NN)
+        self.assertListEqual(result, [["invalid", "invalid"], *TWO_NN])
 
         result_only_valid = model.predict(TEST_SMILES).tolist()
         self.assertListEqual(result_only_valid, TWO_NN)
 
     def test_fit_and_predict_invalid_with_distance(self) -> None:
         """Test the fit_predict method with invalid smiles and distance."""
-        with_invald_smiles = ["CC1CC"] + TEST_SMILES
+        with_invald_smiles = ["CC1CC", *TEST_SMILES]
 
         error_filter = ErrorFilter(filter_everything=True)
         model = Pipeline(
@@ -204,7 +204,7 @@ def test_fit_and_predict_invalid_with_distance(self) -> None:
         result = model.predict(with_invald_smiles, **{"return_distance": True})
         neighbors = result[:, :, 0]
         distances = result[:, :, 1]
-        self.assertListEqual(neighbors.tolist(), [["invalid", "invalid"]] + TWO_NN)
+        self.assertListEqual(neighbors.tolist(), [["invalid", "invalid"], *TWO_NN])
         self.assertTrue(
             1 - np.allclose(distances[1:, :].astype(np.float64), TWO_NN_SIMILARITIES)
         )

tests/test_estimators/test_similarity_transformation.py

@@ -212,8 +212,8 @@ def test_error_handling(self) -> None:
                 ("error_replacer", PostPredictionWrapper(error_replacer)),
             ]
         )
-        full_pipeline.fit(COMPOUND_LIST + ["C#C#C"], IS_AROMATIC + [False])
-        prediction = full_pipeline.predict(COMPOUND_LIST + ["C#C#C"]).tolist()
+        full_pipeline.fit([*COMPOUND_LIST, "C#C#C"], [*IS_AROMATIC, False])
+        prediction = full_pipeline.predict([*COMPOUND_LIST, "C#C#C"]).tolist()
         self.assertListEqual(prediction[:-1], IS_AROMATIC)
         self.assertTrue(np.isnan(prediction[-1]))
 

tests/test_experimental/test_explainability/test_shap_explainers.py

@@ -87,7 +87,8 @@ def _test_valid_explanation(
 
         self.assertIsInstance(explanation.feature_vector, np.ndarray)
         self.assertEqual(
-            (nof_features,), explanation.feature_vector.shape  # type: ignore[union-attr]
+            (nof_features,),
+            explanation.feature_vector.shape,  # type: ignore[union-attr]
         )
 
         # feature names should be a list of not empty strings
@@ -98,7 +99,8 @@ def _test_valid_explanation(
             )
         )
         self.assertEqual(
-            len(explanation.feature_names), explanation.feature_vector.shape[0]  # type: ignore
+            len(explanation.feature_names),  # type: ignore
+            explanation.feature_vector.shape[0],  # type: ignore
         )
 
         self.assertIsInstance(explanation.molecule, RDKitMol)
@@ -112,7 +114,8 @@ def _test_valid_explanation(
         if is_regressor(estimator):
             self.assertTrue((1,), explanation.prediction.shape)  # type: ignore[union-attr]
             self.assertEqual(
-                (nof_features,), explanation.feature_weights.shape  # type: ignore[union-attr]
+                (nof_features,),
+                explanation.feature_weights.shape,  # type: ignore[union-attr]
             )
         elif is_classifier(estimator):
             self.assertTrue((2,), explanation.prediction.shape)  # type: ignore[union-attr]
@@ -123,21 +126,25 @@ def _test_valid_explanation(
                 # https://github.com/shap/shap/issues/3177 returning only one feature weight
                 # which is also based on log odds. This check is a workaround until the bug is fixed.
                 self.assertEqual(
-                    (nof_features,), explanation.feature_weights.shape  # type: ignore[union-attr]
+                    (nof_features,),
+                    explanation.feature_weights.shape,  # type: ignore[union-attr]
                 )
             elif isinstance(estimator, SVC):
                 # SVC seems to be handled differently by SHAP. It returns only a one dimensional
                 # feature array for binary classification.
                 self.assertTrue(
-                    (1,), explanation.prediction.shape  # type: ignore[union-attr]
+                    (1,),
+                    explanation.prediction.shape,  # type: ignore[union-attr]
                 )
                 self.assertEqual(
-                    (nof_features,), explanation.feature_weights.shape  # type: ignore[union-attr]
+                    (nof_features,),
+                    explanation.feature_weights.shape,  # type: ignore[union-attr]
                 )
             else:
                 # normal binary classification case
                 self.assertEqual(
-                    (nof_features, 2), explanation.feature_weights.shape  # type: ignore[union-attr]
+                    (nof_features, 2),
+                    explanation.feature_weights.shape,  # type: ignore[union-attr]
                 )
         else:
             raise ValueError("Error in unittest. Unsupported estimator.")
@@ -176,7 +183,6 @@ def test_explanations_fingerprint_pipeline(  # pylint: disable=too-many-locals
         for estimators, explainer_type in zip(
             explainer_estimators, explainer_types, strict=True
         ):
-
             # test explanations with different estimators
             for estimator in estimators:
                 pipeline = Pipeline(
@@ -241,7 +247,6 @@ def test_explanations_pipeline_with_invalid_inputs(self) -> None:
 
         for estimator in estimators:
             for fill_value in fill_values:
-
                 # pipeline with ErrorFilter
                 error_filter1 = ErrorFilter()
                 pipeline1 = Pipeline(
@@ -271,11 +276,10 @@ def test_explanations_pipeline_with_invalid_inputs(self) -> None:
                 )
 
                 for pipeline in [pipeline1, pipeline2]:
-
                     pipeline.fit(TEST_SMILES_WITH_BAD_SMILES, CONTAINS_OX_BAD_SMILES)
 
                     explainer = SHAPTreeExplainer(pipeline)
-                    log_block = rdBase.BlockLogs()  # pylint: disable=unused-variable
+                    log_block = rdBase.BlockLogs()
                     explanations = explainer.explain(TEST_SMILES_WITH_BAD_SMILES)
                     del log_block
                     self.assertEqual(
@@ -289,7 +293,7 @@ def test_explanations_pipeline_with_invalid_inputs(self) -> None:
                     self.assertIsNotNone(mol_reader_subpipeline)
 
                     for i, explanation in enumerate(explanations):
-                        if i in [3, 7]:
+                        if i in {3, 7}:
                             self.assertFalse(explanation.is_valid())
                             continue
 

tests/test_pipeline.py

@@ -233,7 +233,6 @@ def test_caching(self) -> None:
         for cache_activated in [False, True]:
             pipeline = get_exec_counted_rf_regressor(_RANDOM_STATE)
             with tempfile.TemporaryDirectory() as temp_dir:
-
                 if cache_activated:
                     cache_dir = Path(temp_dir) / ".cache"
                     mem = Memory(location=cache_dir, verbose=0)
@@ -300,7 +299,6 @@ def test_gridsearchcv(self) -> None:
         ]
 
         for test_data_dict in descriptor_elements_to_test:
-
             name = test_data_dict["name"]
             element = test_data_dict["element"]
             param_grid = test_data_dict["param_grid"]