Struct ComponentIndex 

pub struct ComponentIndex { /* private fields */ }

Index for efficient component candidate lookup.

Reduces the O(n·m) comparison to O(n·k) where k << m by:

1. Grouping components by ecosystem
2. Bucketing by name prefix
3. Bucketing by trigrams (3-char substrings) for fuzzy matching
4. Pre-normalizing names for fast comparison

Uses Arc<CanonicalId> internally for efficient cloning during index building.

Implementations

impl ComponentIndex

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

Build an index from an SBOM.

Uses Arc<CanonicalId> internally to avoid expensive cloning of IDs across multiple index structures.

pub fn normalize_component(comp: &Component) -> NormalizedEntry

Normalize a component for indexing.

pub fn normalize_name(name: &str, ecosystem: Option<&str>) -> String

Normalize a component name for comparison.

Applies ecosystem-specific normalization rules:

• PyPI: underscores, hyphens, dots are all equivalent (converted to hyphen)
• Cargo: hyphens and underscores are equivalent (converted to underscore)
• npm: lowercase only, preserves scope
• Default: lowercase with underscore to hyphen conversion

This is also used by LSH for consistent shingle computation.

pub fn get_entry(&self, id: &CanonicalId) -> Option<&NormalizedEntry>

Get normalized entry for a component.

pub fn get_by_ecosystem(&self, ecosystem: &str) -> Option<Vec<CanonicalId>>

Get components by ecosystem.

Returns cloned CanonicalIds for API stability. The internal storage uses Arc to avoid expensive cloning during index building.

pub fn find_candidates( &self, source_id: &CanonicalId, source_entry: &NormalizedEntry, max_candidates: usize, max_length_diff: usize, ) -> Vec<CanonicalId>

Find candidate matches for a component.

Returns a list of component IDs that are likely matches, ordered by likelihood. Uses ecosystem and prefix-based filtering to reduce candidates.

Returns cloned CanonicalIds for API stability. The internal storage uses Arc to avoid expensive cloning during index building.

pub fn all_ids(&self) -> Vec<CanonicalId>

Get all component IDs (for fallback full scan).

Returns cloned CanonicalIds for API stability.

pub fn len(&self) -> usize

Get the number of indexed components.

pub fn is_empty(&self) -> bool

Check if the index is empty.

pub fn find_candidates_parallel<'a>( &self, sources: &[(&'a CanonicalId, &NormalizedEntry)], max_candidates: usize, max_length_diff: usize, ) -> Vec<(&'a CanonicalId, Vec<CanonicalId>)>

Find candidates for multiple source components in parallel.

This is significantly faster than calling find_candidates sequentially for large SBOMs (1000+ components). Uses rayon for parallel iteration.

Returns a vector of (source_id, candidates) pairs in the same order as input.

pub fn find_all_candidates_from( &self, other: &Self, max_candidates: usize, max_length_diff: usize, ) -> Vec<(CanonicalId, Vec<CanonicalId>)>

Find candidates for all components in another index in parallel.

Useful for diffing two SBOMs: build an index from the new SBOM, then find candidates for all components from the old SBOM.

pub fn stats(&self) -> IndexStats

Get statistics about the index.

pub fn trigram_similarity( entry_a: &NormalizedEntry, entry_b: &NormalizedEntry, ) -> f64

Compute trigram similarity between two entries (Jaccard coefficient).

Returns a value between 0.0 and 1.0 where 1.0 means identical trigram sets.