CI Remove pandas FutureWarning in scipy-dev build (scikit-learn#24206)

lesteve · web-flow · commit 5fd66bc55f03 · 2022-08-19T11:52:14.000-04:00
diff --git a/sklearn/datasets/_arff_parser.py b/sklearn/datasets/_arff_parser.py
@@ -392,18 +392,21 @@ def _pandas_arff_parser(
     # `"'some string value'"` with pandas.
     single_quote_pattern = re.compile(r"^'(?P<contents>.*)'$")
 
-    def strip_quotes_regex(s):
-        return s.group("contents")
+    def strip_single_quotes(input_string):
+        match = re.search(single_quote_pattern, input_string)
+        if match is None:
+            return input_string
+
+        return match.group("contents")
 
     categorical_columns = [
         name
         for name, dtype in frame.dtypes.items()
         if pd.api.types.is_categorical_dtype(dtype)
     ]
     for col in categorical_columns:
-        frame[col].cat.categories = frame[col].cat.categories.str.replace(
-            single_quote_pattern, strip_quotes_regex, regex=True
-        )
+        frame[col] = frame[col].cat.rename_categories(strip_single_quotes)
+
     X, y = _post_process_frame(frame, feature_names_to_select, target_names_to_select)
 
     if output_arrays_type == "pandas":
