fix failing to create error report when filter_features is empty list

microsoft · May 12, 2022 · fcc44f0 · fcc44f0
1 parent c53ba4d
commit fcc44f0
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 5 deletions.
diff --git a/erroranalysis/erroranalysis/analyzer/error_analyzer.py b/erroranalysis/erroranalysis/analyzer/error_analyzer.py
@@ -334,7 +334,7 @@ def create_error_report(self,
                                        num_leaves=num_leaves,
                                        min_child_samples=min_child_samples)
         matrix = None
-        if filter_features is not None:
+        if filter_features:
             matrix = self.compute_matrix(filter_features,
                                          None,
                                          None)

diff --git a/erroranalysis/tests/test_error_report.py b/erroranalysis/tests/test_error_report.py
@@ -51,7 +51,8 @@ def test_error_report_housing(self):
             run_error_analyzer(model, X_test, y_test, feature_names,
                                categorical_features)
 
-    def test_error_report_housing_pandas(self):
+    @pytest.mark.parametrize('filter_features', [None, [], ['MedInc', 'HouseAge']])
+    def test_error_report_housing_pandas(self, filter_features):
         X_train, X_test, y_train, y_test, feature_names = \
             create_housing_data()
         X_train = create_dataframe(X_train, feature_names)
@@ -61,7 +62,7 @@ def test_error_report_housing_pandas(self):
         for model in models:
             categorical_features = []
             run_error_analyzer(model, X_test, y_test, feature_names,
-                               categorical_features)
+                               categorical_features, filter_features=filter_features)
 
 
 def is_valid_uuid(id):
@@ -73,7 +74,8 @@ def is_valid_uuid(id):
 
 
 def run_error_analyzer(model, X_test, y_test, feature_names,
-                       categorical_features, expect_user_warnings=False):
+                       categorical_features, expect_user_warnings=False,
+                       filter_features=None):
     if expect_user_warnings and pd.__version__[0] == '0':
         with pytest.warns(UserWarning,
                           match='which has issues with pandas version'):
@@ -84,7 +86,7 @@ def run_error_analyzer(model, X_test, y_test, feature_names,
         model_analyzer = ModelAnalyzer(model, X_test, y_test,
                                        feature_names,
                                        categorical_features)
-    report1 = model_analyzer.create_error_report(filter_features=None,
+    report1 = model_analyzer.create_error_report(filter_features=filter_features,
                                                  max_depth=3,
                                                  num_leaves=None,
                                                  compute_importances=True)
@@ -109,6 +111,11 @@ def run_error_analyzer(model, X_test, y_test, feature_names,
     assert ea_deserialized.importances == report1.importances
     assert ea_deserialized.root_stats == report1.root_stats
 
+    if not filter_features:
+        assert ea_deserialized.matrix == None
+    else:
+        assert ea_deserialized.matrix != None
+
     # validate error report does not modify original dataset in ModelAnalyzer
     if isinstance(X_test, pd.DataFrame):
         assert X_test.equals(model_analyzer.dataset)