Config reference: graph_indexing

• Enterprise tuning surface

  ---

  Defaults + constraints are rendered directly from Pydantic.

• Env keys when available

  ---

  Many fields have an env-style alias (from TriBridConfig.to_flat_dict()).

• Tooltip-level guidance

  ---

  If a matching glossary entry exists, you’ll see deeper tuning notes.

Config reference Config API & workflow Glossary

Total parameters: 26

Group index

• (root)

(root)

JSON key	Env key(s)	Type	Default	Constraints	Summary
graph_indexing.ast_calls_weight	—	float	1.0	≥ 0.0, ≤ 1.0	Edge weight for AST call relationships (function->callee).
graph_indexing.ast_contains_weight	—	float	1.0	≥ 0.0, ≤ 1.0	Edge weight for AST containment relationships (module->class/function, class->method).
graph_indexing.ast_imports_weight	—	float	1.0	≥ 0.0, ≤ 1.0	Edge weight for AST import relationships (module->imported_module).
graph_indexing.ast_inherits_weight	—	float	1.0	≥ 0.0, ≤ 1.0	Edge weight for AST inheritance relationships (class->base).
graph_indexing.build_lexical_graph	—	bool	true	—	Build lexical graph (Document/Chunk nodes + NEXT_CHUNK relationships)
graph_indexing.chunk_embedding_property	—	str	"embedding"	—	Chunk node property that stores the embedding vector
graph_indexing.chunk_vector_index_name	—	str	"tribrid_chunk_embeddings"	—	Neo4j vector index name for Chunk embeddings (mode='chunk')
graph_indexing.enabled	—	bool	true	—	Enable graph building during indexing (Neo4j)
graph_indexing.semantic_kg_allowed_entity_types	—	list[Literal["person", "org", "location", "event", "concept"]]	["concept"]	—	Allowed semantic KG entity types produced by extraction.
graph_indexing.semantic_kg_enabled	—	bool	false	—	Build semantic knowledge graph (concept entities + relations) linked to chunks during indexing
graph_indexing.semantic_kg_llm_model	—	str	""	—	LLM model name for semantic KG extraction when semantic_kg_mode='llm' (empty = use generation.enrich_model)
graph_indexing.semantic_kg_llm_timeout_s	—	int	30	≥ 5, ≤ 120	Timeout (seconds) for semantic KG LLM extraction per chunk
graph_indexing.semantic_kg_max_chunks	—	int	200	≥ 0, ≤ 100000	Maximum chunks to process for semantic KG extraction per indexing run (0 = disabled)
graph_indexing.semantic_kg_max_concepts_per_chunk	—	int	8	≥ 0, ≤ 50	Maximum semantic concepts to extract per chunk
graph_indexing.semantic_kg_max_relations_per_chunk	—	int	12	≥ 0, ≤ 200	Maximum semantic relations to create per chunk (heuristic mode)
graph_indexing.semantic_kg_min_concept_len	—	int	4	≥ 3, ≤ 20	Minimum length for semantic concept tokens
graph_indexing.semantic_kg_mode	—	Literal["heuristic", "llm"]	"heuristic"	allowed="heuristic", "llm"	Semantic KG extraction mode. 'heuristic' is deterministic and test-friendly; 'llm' uses an LLM to extract entities + relations.
graph_indexing.semantic_kg_reasoning_effort	—	Literal["minimal", "low", "medium", "high", "xhigh"]	"medium"	allowed="minimal", "low", "medium", "high", "xhigh"	Reasoning effort for semantic KG extraction when using OpenAI Responses-compatible models.
graph_indexing.semantic_kg_relation_weight_heuristic	—	float	0.5	≥ 0.0, ≤ 1.0	Edge weight for semantic concept relations in heuristic fallback mode.
graph_indexing.semantic_kg_relation_weight_llm	—	float	0.7	≥ 0.0, ≤ 1.0	Edge weight for semantic concept relations in LLM mode.
graph_indexing.semantic_kg_require_llm_success	—	bool	false	—	When true in LLM mode, fail semantic KG extraction for a chunk if LLM extraction fails instead of falling back.
graph_indexing.semantic_kg_typed_entities_enabled	—	bool	false	—	When true, semantic KG extraction preserves/uses typed entities (person, org, location, event, concept).
graph_indexing.store_chunk_embeddings	—	bool	true	—	Store chunk embeddings on Chunk nodes for Neo4j vector search (requires dense embeddings)
graph_indexing.vector_index_online_timeout_s	—	float	60.0	≥ 1.0, ≤ 600.0	Timeout waiting for Neo4j vector index ONLINE (seconds)
graph_indexing.vector_similarity_function	—	Literal["cosine", "euclidean"]	"cosine"	allowed="cosine", "euclidean"	Neo4j vector similarity function
graph_indexing.wait_vector_index_online	—	bool	true	—	Wait for the Neo4j vector index to come ONLINE after (re)creating it