Struct BatchCandidateGenerator 

pub struct BatchCandidateGenerator { /* private fields */ }

Batch candidate generator that combines multiple indexing strategies.

For best recall, combines:

1. ComponentIndex (ecosystem, prefix, trigram-based)
2. LSH index (for large SBOMs, catches approximate matches)
3. Cross-ecosystem mappings (optional)

The candidates from each source are deduplicated and merged.

Implementations

impl BatchCandidateGenerator

pub fn build(sbom: &NormalizedSbom, config: BatchCandidateConfig) -> Self

Create a new batch candidate generator from an SBOM.

pub fn find_candidates( &self, source_id: &CanonicalId, source_component: &Component, ) -> BatchCandidateResult

Generate candidates for a single component.

pub fn find_candidates_batch( &self, sources: &[(&CanonicalId, &Component)], ) -> Vec<BatchCandidateResult>

Generate candidates for multiple components in parallel.

pub fn all_candidates( &self, source_id: &CanonicalId, source_component: &Component, ) -> Vec<CanonicalId>

Get all unique candidates (deduplicated across all strategies).

pub const fn component_index(&self) -> &ComponentIndex

Get the underlying component index.

pub const fn has_lsh(&self) -> bool

Check if LSH is enabled.

pub const fn has_cross_ecosystem(&self) -> bool

Check if cross-ecosystem matching is enabled.

pub fn stats(&self) -> BatchCandidateStats

Get statistics about the generator.