I would like to automatize pmdarima data preprocessing. I use the standard Pipeline, where the transform part is dynamically created based on data normality:
transform
