Each endpoint gives its response in a different format, with different metadata; some have more metadata and linked data than others. How opinionated should the open-source app be at flattening and normalizing these various endpoints for easier placement into a single table or database?