Rationalise how we determine what to pass through/remove/transform from upstream APIs

[chris48s]

The way we pass through data from upstream APIs is currently a bit haphazard:

• Sometimes we just pass objects through, so if we change the upstream API, the change just automatically happens here too
• Sometimes we explicitly pick out specific fields, so an upstream change that we want to reflect here has to be explicitly accounted for
• Sometimes we apply a transformation to the data

This is a bit inconsistent. The idea is currently a bit half-baked, but it feels like there is scope to make this all much more obvious/predictable/consistent at all layers of the object hierarchy by applying a schema to the input and output data instead of adding ad-hoc rules, but I need to put more thought into it..