Revert changes to transfomr prim components for testing

alteryx · Feb 28, 2023 · 57dda43 · 57dda43
1 parent 1a364c8
commit 57dda43
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 10 deletions.
diff --git a/evalml/pipelines/components/transformers/preprocessing/transform_primitive_components.py b/evalml/pipelines/components/transformers/preprocessing/transform_primitive_components.py
@@ -73,6 +73,17 @@ def transform(self, X, y=None):
         es = self._make_entity_set(X_ww)
         features = ft.calculate_feature_matrix(features=self._features, entityset=es)
 
+        # Convert to object dtype so that pd.NA is converted to np.nan
+        # until sklearn imputer can handle pd.NA in release 1.1
+        # FT returns these as string types, currently there isn't much difference
+        # in terms of performance between object and string
+        # see https://pandas.pydata.org/docs/user_guide/text.html#text-data-types
+        # "Currently, the performance of object dtype arrays of strings
+        # "and arrays.StringArray are about the same."
+        features = features.astype(object, copy=False)
+        features.index = X_ww.index
+        features.ww.init(logical_types={col_: "categorical" for col_ in features})
+
         X_ww = X_ww.ww.drop(self._columns)
         X_ww = ww.concat_columns([X_ww, features])
 

diff --git a/evalml/tests/component_tests/test_ft_transform_primitive_components.py b/evalml/tests/component_tests/test_ft_transform_primitive_components.py
@@ -50,11 +50,9 @@ def make_answer_email_fit_transform(df_with_url_and_email):
         ["gmail.com", "yahoo.com", "abalone.com", "hotmail.com", "email.org"],
         dtype="category",
     )
-    expected.ww["IS_FREE_EMAIL_DOMAIN(email)"] = ww.init_series(
-        pd.Series(
-            [True, True, False, True, True],
-        ),
-        logical_type="BooleanNullable",
+    expected.ww["IS_FREE_EMAIL_DOMAIN(email)"] = pd.Series(
+        [True, True, False, True, True],
+        dtype="category",
     )
     expected.ww.drop(["email"], inplace=True)
     return expected
@@ -95,11 +93,11 @@ def make_answer_email_fit_transform_missing_values(df_with_url_and_email):
     )
     expected.ww["IS_FREE_EMAIL_DOMAIN(email)"] = pd.Series(
         [None, None, False, True, True],
-        dtype="boolean",
+        dtype="category",
     )
     expected.ww["IS_FREE_EMAIL_DOMAIN(email_2)"] = pd.Series(
         [None, None, False, True, None],
-        dtype="boolean",
+        dtype="category",
     )
     return expected
 
@@ -143,7 +141,7 @@ def make_expected_logical_types_email_fit_transform():
     return {
         "categorical": ww.logical_types.Categorical(),
         "numeric": ww.logical_types.Double(),
-        "IS_FREE_EMAIL_DOMAIN(email)": ww.logical_types.BooleanNullable(),
+        "IS_FREE_EMAIL_DOMAIN(email)": ww.logical_types.Categorical(),
         "EMAIL_ADDRESS_TO_DOMAIN(email)": ww.logical_types.Categorical(),
         "integer": ww.logical_types.Integer(),
         "boolean": ww.logical_types.Boolean(),
@@ -172,8 +170,8 @@ def make_expected_logical_types_email_fit_transform_missing_values():
         "numeric": ww.logical_types.Double(),
         "EMAIL_ADDRESS_TO_DOMAIN(email)": ww.logical_types.Categorical(),
         "EMAIL_ADDRESS_TO_DOMAIN(email_2)": ww.logical_types.Categorical(),
-        "IS_FREE_EMAIL_DOMAIN(email)": ww.logical_types.BooleanNullable(),
-        "IS_FREE_EMAIL_DOMAIN(email_2)": ww.logical_types.BooleanNullable(),
+        "IS_FREE_EMAIL_DOMAIN(email)": ww.logical_types.Categorical(),
+        "IS_FREE_EMAIL_DOMAIN(email_2)": ww.logical_types.Categorical(),
         "integer": ww.logical_types.Integer(),
         "boolean": ww.logical_types.Boolean(),
         "nat_lang": ww.logical_types.NaturalLanguage(),