Struct Embeddings 

pub struct Embeddings {
    pub vectors: Vec<(String, Vec<f64>)>,
    /* private fields */
}

Lightweight semantic embeddings for repo memory. §3.8: “optional embeddings (per project)”

Fields

vectors: Vec<(String, Vec<f64>)>

Stored text + normalized hashing-vector embedding.

Implementations

impl Embeddings

pub const DEFAULT_DIMENSIONS: usize = 512

pub fn new() -> Self

pub fn with_dimensions(dimensions: usize) -> Self

pub fn embed(&self, text: &str) -> Vec<f64>

Build a deterministic hashing-vector embedding from text.

This is intentionally local-first: no model/API key required, fixed dimensions across documents, stable across sessions, and good enough for lexical semantic recall in memory. It uses signed feature hashing with unigram + adjacent bigram features, sublinear term frequency, and L2 normalization.

pub fn add(&mut self, text: &str)

pub fn add_many<I, S>(&mut self, texts: I)

where I: IntoIterator<Item = S>, S: AsRef<str>,

pub fn search(&self, query: &str, k: usize) -> Vec<String>

Find the most similar stored text to the query

pub fn search_scored(&self, query: &str, k: usize) -> Vec<(f64, String)>

pub fn save_to_path(&self, path: impl AsRef<Path>) -> Result<()>

pub fn load_from_path(path: impl AsRef<Path>) -> Result<Self>

Trait Implementations

impl Clone for Embeddings

fn clone(&self) -> Embeddings

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for Embeddings

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

Formats the value using the given formatter.

impl Default for Embeddings

fn default() -> Self

Returns the “default value” for a type.