Struct ToyEmbedder 

pub struct ToyEmbedder { /* private fields */ }

Deterministic hashed bag-of-tokens embedder.

embed(text) lowercases, ASCII-tokenises on non-alphanumeric boundaries, hashes each token to two bucket positions (FNV-1a-style) and adds 1.0 to each. The output vector is L2- normalised so dense cosine similarity ranks documents by (count-weighted) shared-token overlap.

Implementations

impl ToyEmbedder

pub fn new(dim: u32) -> Self

Construct a new embedder with the given dimension. dim must be > 0; values < 32 lead to heavy hash collisions on natural text.

pub fn model(&self) -> &str

Model identifier (passed to mnem’s vector lane so embeddings match query vectors at retrieve time).

pub const fn dim(&self) -> u32

Embedding dimension.

pub fn embed_text(&self, text: &str) -> Vec<f32>

Embed a string into a unit-norm vector.

Trait Implementations

impl Clone for ToyEmbedder

fn clone(&self) -> ToyEmbedder

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ToyEmbedder

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

Formats the value using the given formatter.