Currently the Patientinformation transfered to the Core Bundles is the whole Relationship Graph.
This is later computed into extraction information in the BatchCopyRedacter.
For different use cases like serialization and simplicity this info should be precomputed in a separate service and shared with the core.