It would be nice to have a thin Python layer on top of the Web API, which takes care of the most common Software Heritage data types (e.g., PIDs) and transparently handle functionalities shared by various endpoints (e.g., pagination).