Architecture

DIVA uses a versioned pet_schema = 3 data model.

Each pet record is normalized into explicit sections:

• profile
• plan
• medical
• analytics
• external_links

Runtime state is persisted to JSON and grouped into typed sections:

• medical
• mobility
• behavior
• operations
• reports

The domain layer exposes explicit sub-engine boundaries inside PetEngine:

• routines
• nutrition
• medical
• mobility
• behavior
• reporting