Struct EmbeddingIndex 

pub struct EmbeddingIndex { /* private fields */ }

Brute-force vector index for semantic search.

Stores (unit_id, feature_vec) pairs copied from the graph for cache locality during sequential scan.

Implementations

impl EmbeddingIndex

pub fn build(graph: &CodeGraph) -> Self

Build an EmbeddingIndex from all code units in the given graph.

Only includes units whose feature vectors have the expected dimension and are not all-zero (i.e., have a non-negligible norm).

pub fn search( &self, query: &[f32], top_k: usize, min_similarity: f32, ) -> Vec<EmbeddingMatch>

Search for the most similar vectors to query.

Returns at most top_k results with cosine similarity >= min_similarity, ordered by descending score.

If the query vector dimension does not match the index dimension, or the query vector has near-zero norm, returns an empty vector.

pub fn dimension(&self) -> usize

Returns the dimensionality of vectors in this index.

pub fn len(&self) -> usize

Returns the number of indexed entries (units with valid feature vectors).

pub fn is_empty(&self) -> bool

Returns true if the index contains no entries.

Trait Implementations

impl Clone for EmbeddingIndex

fn clone(&self) -> EmbeddingIndex

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for EmbeddingIndex

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for EmbeddingIndex

fn default() -> Self

Returns the “default value” for a type.