Struct HashEmbedder 

pub struct HashEmbedder {
    pub dims: usize,
}

Deterministic, dependency-free embedder for tests, examples, and embeddings-less deployments.

Hashes the text into a fixed-dimension L2-normalized vector: no semantics, but consistent geometry — the same text always yields the same vector, so dedup and round-trip assertions work without a model. Not for production: similarity between different texts is noise. Model id is hash-embedder (embedding provenance still detects it).

Fields

dims: usize

Vector dimensionality (default DEFAULT_EMBEDDING_DIM).

Trait Implementations

impl Clone for HashEmbedder

fn clone(&self) -> HashEmbedder

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for HashEmbedder

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

Formats the value using the given formatter.

impl Default for HashEmbedder

fn default() -> Self

Returns the “default value” for a type.

impl Embedder for HashEmbedder

fn dim(&self) -> usize

Dimensionality of the vectors this embedder produces.

fn embed(&self, text: &str) -> Result<Embedding>

Embeds a single piece of text.

fn model_id(&self) -> &str

The model id that produced these embeddings (e.g. sentence-transformers/all-MiniLM-L6-v2@main). Used for embedding provenance (D10): memories stamp the model that embedded them, so a model swap is detectable and a re-embed pass can be triggered rather than silently producing incomparable vectors. Default "unknown" for backends that don’t track it.

fn embed_batch(&self, texts: &[&str]) -> Result<Vec<Embedding>>

Embeds a batch of texts. The default loops over [embed]; backends with batch acceleration override this.