Expand description
Deterministic entity extraction for associative (spreading-activation) recall.
SAG-inspired: a chunk’s high-signal tokens (error codes, CLI flags, paths,
code symbols) become index entities in chunk_entities. Two chunks that share
an entity are associatively linked even when their prose is dissimilar, which
lets recall spread activation across that link — the ACT-R associative term
(S_ji) that base-level activation alone cannot supply.
Deterministic on purpose — no LLM, no regex dependency (manual char
scanning) — so it runs on the no-LLM capture hot path and in migration
backfill. Mirrors the project constraint that capture never depends on the LLM
(see ResilientDistiller).
Deliberately ignores plain lowercase words: those are already recoverable via the lexical/BM25 channel, and entities must stay discriminative so the ACT-R fan term keeps promiscuous tokens from dominating the spread.
Structs§
- Extracted
Entity - One extracted entity: the normalized join key plus a coarse type tag. The tag
is diagnostic only — recall keys on
entity, never onetype.
Functions§
- extract_
entities - Extract high-signal entities from a chunk’s content and optional trigger desc.