Skip to content

Interface: OsfqlDiagnoseRequest

@kortexya/reasoninglayer


@kortexya/reasoninglayer / OsfqlDiagnoseRequest

Defined in: src/types/osfql.ts:428

Request to diagnose an OSFQL program for contradictions / inconsistencies.

The program is executed ephemerally (against a cloned hierarchy and a scratch store) and never mutates the tenant knowledge base. The check always runs under a deadline and a row cap; a breach of either answers a limit diagnostic, not an error.

optional maxRows?: number

Defined in: src/types/osfql.ts:461

The maximum number of rows any ONE candidate-set materialisation of the check may produce — the same per-materialisation cap as OsfqlRequest.maxRows.

Unlike on execute, the cap is ON by default: a join of a few scans over one sort materialises the product of their sizes. A supplied value is clamped to 1000000. A breach answers a limit diagnostic positioned at the statement that outgrew the cap.

100000


query: string

Defined in: src/types/osfql.ts:430

OSFQL program text to check.


optional timeoutMs?: number

Defined in: src/types/osfql.ts:447

A deadline for the whole check, in milliseconds.

Shared by every execution the check performs — the whole-program run and the re-runs that localise a failure to its statement. A supplied value is clamped to the engine’s program backstop (OSFKB_OSFQL_TIMEOUT_SECS).

Unlike OsfqlRequest.timeoutMs, 0 is NOT an opt-out: the engine refuses it with HTTP 400, because a check that executes client text on the tenant’s data has no unbounded mode. A breach answers a limit diagnostic with no range.

2000 — a lint budget, because an editor calls this route as its author types.