Fix code quality tests

sktime/forecasting/tests/test_neuralforecast.py

@@ -318,7 +318,7 @@ def test_neural_forecast_with_auto_freq_on_missing_date_like(
 
     model = model_class(freq="auto", max_steps=1, trainer_kwargs={"logger": False})
 
-     with pytest.raises(
+    with pytest.raises(
         ValueError, match="(could not interpret freq).*(use a valid offset in index)"
     ):
         model.fit(y, fh=[1, 2, 3])