Entity disambiguation¶
When two extracted entities are merged into one, within a document and across documents.
Within a document¶
AGG_ENABLED¶
Run identity consolidation over the facts units: ontology anchoring, embedding clustering, natural-key pairs and the lexical/symbolic merge tiers. Off keeps every entity its own cluster -- naming, provenance and document metadata still run, and the aggregation encoder is never loaded. Requires AGG_UNIT_SCOPED_FACT_IRIS, or units fuse by local name anyway.
AGG_EMBEDDING_MODEL¶
Sentence-transformers model name used for entity embeddings. Spelled with the org prefix to match CHUNK_EMBEDDING_MODEL and EMBEDDING_MODEL_NAME: SharedEncoder keys its process-wide cache on the literal string, so the same checkpoint written two ways loads twice.
AGG_SIMILARITY_THRESHOLD¶
Cosine threshold of the cross-graph entity aligner when the caller names none: POST /match/entities, match-graphs and the ontocast_align_entities agent tool. The in-pipeline aggregator uses AGG_CANDIDATE_SIMILARITY_THRESHOLD; this setting does not affect it.
AGG_CANDIDATE_SIMILARITY_THRESHOLD¶
Cosine threshold of the in-pipeline aggregator: DBSCAN candidate clustering and the pairwise gate both use it. Deliberately permissive — candidates are validated symbolically afterwards.
AGG_LITERAL_CONFLICT_GUARD¶
Veto identity merges between entities asserting disjoint literal values on a shared predicate (numeric/temporal disjointness, or string sets with no compatible cross-pair). Turning it off isolates this guard's contribution to rejected merges.
AGG_INITIALS_DISTINCT_GUARD¶
Veto identity merges between entities whose labels are identical except for conflicting initials or single-letter identifiers ('company S.' vs 'company T.') — the shape authors write to distinguish entities.
AGG_LLM_COREFERENCE¶
After aggregation, put the identity pairs the symbolic tiers cannot settle to the model in one call per document of two or more units: merges made on a plausible lexical tier only, candidates refused for wording alone, and same-type entities of different units pointing to one node by one predicate. A 'same' verdict merges like a shared key and 'different' vetoes; every distinctness guard still applies.
AGG_LLM_COREFERENCE_MAX_PAIRS¶
Most pairs put to the model per document; the three kinds of candidate take turns up to it.
AGG_NATURAL_KEY_MERGE¶
Positive identity evidence from natural keys: instances sharing an identical short string value on a single-valued identifier-like predicate (schema max-1, or observed single-valued on every subject) become merge candidates even when their labels and embeddings disagree. All distinctness guards still apply.
AGG_TYPE_GUARD_UNTYPED¶
Type-compatibility guard behaviour for untyped entities. 'permissive' (default) lets a typed entity merge with an untyped one; 'strict' fails typed-vs-untyped pairs closed (two untyped entities stay comparable in both modes).
AGG_LEXICAL_LABEL_JACCARD¶
Minimum label token-set Jaccard for the fuzzy lexical-alias merge tier.
AGG_LEXICAL_SEQUENCE_RATIO¶
Minimum SequenceMatcher ratio on URI normal forms for the fuzzy lexical-alias merge tier.
AGG_LEXICAL_TOKEN_JACCARD¶
Minimum normal-form token Jaccard for the fuzzy lexical-alias merge tier (both sides >= 2 tokens).
AGG_FUNCTIONAL_MIN_EMPIRICAL_SUPPORT¶
Minimum distinct subjects a predicate must be observed on before it counts as empirically single-valued for the functional-object merge guard.
AGG_SIBLING_GUARD_SCOPE¶
Co-object sibling guard scope: 'subject' forbids merging any two objects of one subject; 'predicate' restricts the prohibition to objects sharing the same predicate.
AGG_UNIT_SCOPED_FACT_IRIS¶
Suffix every minted fact IRI with the index of the unit that minted it (
Across documents¶
LINK_ENABLED¶
Write each document's identity links graph (ontocast link build derives the same graphs later.