LLMMemoryCandidatePipeline Python API

Map schema-valid model proposals back to exact bounded domain values.

Functions

constructor__init__(generator, *, evidence_projector=None) -> None

paramgeneratorStructuredGenerator[MemoryExtractionInput, MemoryExtractionOutput]

paramevidence_projectorMemoryEvidenceProjector | None

= None

Returns

None

funcextract(request) -> tuple[MemoryEntryInput, ...]

Generate and map candidates that still require MemoryService validation.

paramrequestMemoryCandidateRequest

Returns

tuple[powercontext.builtin.artifacts.memory.models.MemoryEntryInput, ...]