add check for feat types

Author: ravinkohli

autoPyTorch/data/tabular_feature_validator.py

@@ -390,15 +390,37 @@ def get_columns_to_encode(
         """
         transformed_columns, feat_types = self._get_columns_to_encode(X)
         if self.feat_types is not None:
-            if len(self.feat_types) != len(X.columns):
-                raise ValueError(f"Expected number of `feat_types`: {len(self.feat_types)}"
-                                 f" to be the same as the number of features {len(X.columns)}")
+            self._validate_feat_types(X)
             transformed_columns = [X.columns[i] for i, col in enumerate(self.feat_types)
                                    if col.lower() == 'categorical']
             return transformed_columns, self.feat_types
         else:
             return transformed_columns, feat_types
 
+    def _validate_feat_types(self, X: pd.DataFrame):
+        """
+        Checks if the passed `feat_types` is compatible with what 
+        AutoPyTorch expects, i.e, it should only contain `numerical` 
+        or `categorical`. The case does not matter. 
+
+        Args:
+            X (pd.DataFrame):
+                input features set
+
+        Raises:
+            ValueError:
+                if the number of feat_types is not equal to the number of features
+                if the feature type are not one of "numerical", "categorical"
+
+        """
+        if len(self.feat_types) != len(X.columns):
+            raise ValueError(f"Expected number of `feat_types`: {len(self.feat_types)}"
+                             f" to be the same as the number of features {len(X.columns)}")
+        for feat_type in set(self.feat_types):
+            if feat_type.lower() not in ['numerical', 'categorical']:
+                raise ValueError(f"Expected type of features to be in `['numerical', "
+                                 f"'categorical']`, but got {feat_type}")
+
     def _get_columns_to_encode(
         self,
         X: pd.DataFrame,

test/test_data/test_feature_validator.py

@@ -439,6 +439,12 @@ def input_data_feature_feat_types(request):
             {'A': 3, 'B': '4'},
         ], dtype='category')
         return frame, ['categorical', 'categorical', 'numerical']
+    elif request.param == 'pandas_feat_type_error':
+        frame = pd.DataFrame([
+            {'A': 1, 'B': '2'},
+            {'A': 3, 'B': '4'},
+        ], dtype='category')
+        return frame, ['not_categorical', 'numerical']
     else:
         ValueError("Unsupported indirect fixture {}".format(request.param))
 
@@ -500,4 +506,23 @@ def test_feature_validator_get_columns_to_encode_error_length(input_data_feature
     X, feat_types = input_data_feature_feat_types
     validator = TabularFeatureValidator(feat_types=feat_types)
     with pytest.raises(ValueError, match=r"Expected number of `feat_types`: .*"):
-        validator.get_columns_to_encode(X)
+        validator._validate_feat_types(X)
+
+
+@pytest.mark.parametrize(
+    'input_data_feature_feat_types',
+    (
+        'pandas_feat_type_error',
+    ),
+    indirect=True
+)
+def test_feature_validator_get_columns_to_encode_error_feat_type(input_data_feature_feat_types):
+    """
+    Tests the correct error is raised when the length of feat types passed to
+    the validator is not the same as the number of features
+
+    """
+    X, feat_types = input_data_feature_feat_types
+    validator = TabularFeatureValidator(feat_types=feat_types)
+    with pytest.raises(ValueError, match=r"Expected type of features to be in .*"):
+        validator._validate_feat_types(X)