Struct QueryEngine 

pub struct QueryEngine;

The query engine supports all query operations.

Implementations

impl QueryEngine

pub fn belief_revision( &self, graph: &MemoryGraph, params: BeliefRevisionParams, ) -> AmemResult<RevisionReport>

Analyze how a hypothesis would affect existing beliefs.

This is a read-only operation: it does not modify the graph. It finds contradictions via text overlap + vector similarity, checks for negation words, and propagates weakening through CausedBy / Supports edges.

pub fn gap_detection( &self, graph: &MemoryGraph, params: GapDetectionParams, ) -> AmemResult<GapReport>

Detect gaps in the knowledge graph: unjustified decisions, single-source inferences, low-confidence foundations, unstable knowledge, and stale evidence.

This is a read-only operation.

pub fn analogical( &self, graph: &MemoryGraph, params: AnalogicalParams, ) -> AmemResult<Vec<Analogy>>

Find subgraphs that are structurally and semantically analogous to the anchor.

Uses structural fingerprinting (node type counts, edge type counts, causal chain depth, branching factor) rather than full subgraph isomorphism.

combined_score = 0.6 * structural + 0.4 * content

This is a read-only operation.

pub fn drift_detection( &self, graph: &MemoryGraph, params: DriftParams, ) -> AmemResult<DriftReport>

Track how beliefs on a given topic have changed over time.

Finds relevant nodes, groups them into SUPERSEDES chains and clusters of similar content across sessions, and reports stability.

This is a read-only operation.

impl QueryEngine

pub fn centrality( &self, graph: &MemoryGraph, params: CentralityParams, ) -> AmemResult<CentralityResult>

Compute centrality scores for nodes in the graph.

pub fn shortest_path( &self, graph: &MemoryGraph, params: ShortestPathParams, ) -> AmemResult<PathResult>

Find the shortest path between two nodes.

impl QueryEngine

pub fn consolidate( &self, graph: &mut MemoryGraph, params: ConsolidationParams, ) -> AmemResult<ConsolidationReport>

Run a set of consolidation operations against graph.

If params.dry_run is true, the graph is not mutated; the returned report describes what would happen.

When dry_run is false:

• DeduplicateFacts adds Supersedes edges from the surviving node to each duplicate.
• LinkContradictions adds Contradicts edges.
• PromoteInferences changes event_type from Inference to Fact.
• PruneOrphans and CompressEpisodes are always dry-run-only in V1.

impl QueryEngine

pub fn new() -> Self

Create a new query engine.

pub fn traverse( &self, graph: &MemoryGraph, params: TraversalParams, ) -> AmemResult<TraversalResult>

Traverse from a starting node following specific edge types.

pub fn pattern<'a>( &self, graph: &'a MemoryGraph, params: PatternParams, ) -> AmemResult<Vec<&'a CognitiveEvent>>

Find nodes matching conditions.

pub fn temporal( &self, graph: &MemoryGraph, params: TemporalParams, ) -> AmemResult<TemporalResult>

Compare graph state across time ranges or sessions.

pub fn causal( &self, graph: &MemoryGraph, params: CausalParams, ) -> AmemResult<CausalResult>

Impact analysis: what depends on a given node?

pub fn similarity( &self, graph: &MemoryGraph, params: SimilarityParams, ) -> AmemResult<Vec<SimilarityMatchResult>>

Find similar nodes using feature vector cosine similarity.

pub fn memory_quality( &self, graph: &MemoryGraph, params: MemoryQualityParams, ) -> AmemResult<MemoryQualityReport>

Evaluate memory quality across confidence, freshness, and graph structure.

pub fn context( &self, graph: &MemoryGraph, node_id: u64, depth: u32, ) -> AmemResult<SubGraph>

Get the full context for a node: the node itself, all edges, and connected nodes.

pub fn resolve<'a>( &self, graph: &'a MemoryGraph, node_id: u64, ) -> AmemResult<&'a CognitiveEvent>

Get the latest version of a node, following SUPERSEDES chains.

impl QueryEngine

pub fn text_search( &self, graph: &MemoryGraph, term_index: Option<&TermIndex>, doc_lengths: Option<&DocLengths>, params: TextSearchParams, ) -> AmemResult<Vec<TextMatch>>

BM25 text search over node contents. Uses TermIndex if available, falls back to full scan.

pub fn hybrid_search( &self, graph: &MemoryGraph, term_index: Option<&TermIndex>, doc_lengths: Option<&DocLengths>, params: HybridSearchParams, ) -> AmemResult<Vec<HybridMatch>>

Hybrid BM25 + vector search with Reciprocal Rank Fusion.

Trait Implementations

impl Default for QueryEngine

fn default() -> Self

Returns the “default value” for a type.