Skip to content

Interface: DiscoverSchemaResponse

@kortexya/reasoninglayer


@kortexya/reasoninglayer / DiscoverSchemaResponse

Defined in: src/types/sources.ts:578

Response from schema discovery.

relations: DiscoveredSourceRelationDto[]

Defined in: src/types/sources.ts:584

Discovered relations between sorts.


sorts: DiscoveredSortDto[]

Defined in: src/types/sources.ts:582

Discovered sorts/types.


sourceId: string

Defined in: src/types/sources.ts:580

Source identifier.


totalColumns: number

Defined in: src/types/sources.ts:586

Total number of columns/features discovered.


optional totalDiscovered?: number

Defined in: src/types/sources.ts:595

Total number of sorts/tables discovered and returned.

Redundant with sorts.length on the happy path, but surfaced explicitly so a client can distinguish “the source has exactly N tables” from a silently truncated list. Discovery hard-refuses (400) rather than truncate, so a successful response always carries the complete set.


optional warnings?: string[]

Defined in: src/types/sources.ts:607

Warnings during discovery.

For a CSV type, discovery nominates at most one identifier column, and none when fewer than 10 rows were sampled or no column held distinct, key-shaped values. A type left without one gets a warning of the form "<type>: No identifier discovered: <reason>": bind such a type with explicit key columns, because a binding with no key is refused when its rows are projected.