The pandas wrap-around is quite clever. Can you share more information on this or point me to a repo please?