the meshery pr (meshery/meshery#917) adds summary filter apis for components and relationships in the registry. reviewers (@lekaf974 , @aabidsofi19 ) agreed that the filter/response structs should live in meshery/schemas as the source of truth, not in the meshery repo directly.
looking for alignment on the schema design before updating the existing pr (#616).
proposed schemas:
ComponentSummaryFilter filter params (model, category, registrant, annotations)
ComponentSummary response with grouped entries
ComponentGroupEntry single group row (key + count)
RelationshipSummaryFilter filter params (model, kind, type, subtype)
RelationshipSummary response with grouped entries
RelationshipGroupEntry single group row (key + count)
open question from gemini review: the generated Summary structs use anonymous structs instead of reusing GroupEntry types should the schema use $ref to the entry type to avoid duplication in codegen output?
ref: meshery/meshery#917, #616, #615
the meshery pr (meshery/meshery#917) adds summary filter apis for components and relationships in the registry. reviewers (@lekaf974 , @aabidsofi19 ) agreed that the filter/response structs should live in meshery/schemas as the source of truth, not in the meshery repo directly.
looking for alignment on the schema design before updating the existing pr (#616).
proposed schemas:
ComponentSummaryFilterfilter params (model, category, registrant, annotations)ComponentSummaryresponse with grouped entriesComponentGroupEntrysingle group row (key + count)RelationshipSummaryFilterfilter params (model, kind, type, subtype)RelationshipSummaryresponse with grouped entriesRelationshipGroupEntrysingle group row (key + count)open question from gemini review: the generated
Summarystructs use anonymous structs instead of reusingGroupEntrytypes should the schema use$refto the entry type to avoid duplication in codegen output?ref: meshery/meshery#917, #616, #615