Struct SimdVectorSearch 

pub struct SimdVectorSearch { /* private fields */ }

SIMD-accelerated flat (brute-force) vector search.

Stores all vectors in memory and computes cosine distances to a query using scirs2_core’s SIMD dot-product primitive. Parallel execution is used for batches that exceed the configured threshold.

Implementations

impl SimdVectorSearch

pub fn new(parallel_threshold: usize) -> Self

Create a new empty flat index with the given parallel threshold.

pub fn default_threshold() -> Self

Create a flat index with the default threshold (256 vectors).

pub fn insert(&mut self, id: String, vector: Vec<f32>) -> Result<()>

Insert a vector under id. If id already exists, the vector is replaced.

pub fn len(&self) -> usize

Return the number of vectors in the index.

pub fn is_empty(&self) -> bool

Check whether the index is empty.

pub fn search(&self, query: &[f32], k: usize) -> Result<Vec<(String, f32)>>

Search for the k nearest neighbours of query by cosine distance.

Returns a Vec of (id, distance) pairs sorted by ascending distance, length ≤ k.

pub fn all_distances(&self, query: &[f32]) -> Result<Vec<(String, f32)>>

Compute raw cosine distances from query to all indexed vectors and return them in insertion order (not sorted).

Trait Implementations

impl Debug for SimdVectorSearch

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

Formats the value using the given formatter.

impl Default for SimdVectorSearch

fn default() -> Self

Returns the “default value” for a type.