do more translation in core rather than API layer

it would be better to do data model transformation/translations in the core code so that the API code can be a simpler frontend for accessing the abstractions provided by the core code

Start by mapping where data-model transformations occur across the core and API layers; the issue does not name specific files, tests, or entry points. Define the desired core/API boundary before changing code, with completion meaning translations are handled by core abstractions and the API acts as a simpler frontend.

Written by the indexing model from the issue text.