Some LinkML schemas such as Biolink contain class slots whose ranges are the same class, see biolink/biolink-model#1378 this causes a cyclic dependency when translating these to foreign key relationships. A few of these might be mistakes, but others may be intentional for the purpose of creating "denormalized" slots.
How should these be handled in cases where they're legitimate? Can denormalized slots be skipped or handled differently?
Some LinkML schemas such as Biolink contain class slots whose ranges are the same class, see biolink/biolink-model#1378 this causes a cyclic dependency when translating these to foreign key relationships. A few of these might be mistakes, but others may be intentional for the purpose of creating "denormalized" slots.
How should these be handled in cases where they're legitimate? Can denormalized slots be skipped or handled differently?