Struct NativeDiscoveryEngine 

pub struct NativeDiscoveryEngine { /* private fields */ }

The main RuVector-native discovery engine

This engine uses RuVector’s core algorithms:

• Vector similarity via HNSW index
• Graph coherence via Stoer-Wagner min-cut
• Temporal windowing for streaming analysis

Implementations

impl NativeDiscoveryEngine

pub fn new(config: NativeEngineConfig) -> Self

Create a new engine with the given configuration

pub fn add_vector(&mut self, vector: SemanticVector) -> u32

Add a vector to the engine In production, this would use ruvector-core’s vector storage

pub fn add_correlation_edge( &mut self, source: u32, target: u32, correlation: f64, )

Add a correlation-based edge

pub fn compute_coherence(&self) -> CoherenceSnapshot

Compute current coherence using Stoer-Wagner min-cut

The min-cut value represents the “weakest link” in the network. A drop in min-cut indicates the network is becoming fragmented.

pub fn detect_patterns(&mut self) -> Vec<DiscoveredPattern>

Detect patterns by comparing current state to history

pub fn domain_coherence(&self, domain: Domain) -> Option<f64>

Get domain-specific coherence

pub fn stats(&self) -> EngineStats

Get statistics about the current state

pub fn get_patterns(&self) -> Vec<DiscoveredPattern>

Get all detected patterns from the latest detection run

pub fn export_graph(&self) -> GraphExport

Export the current graph structure

pub fn get_coherence_history(&self) -> Vec<CoherenceHistoryEntry>

Get the coherence history