Struct CandleEmbeddingProvider 

pub struct CandleEmbeddingProvider { /* private fields */ }

Local embedding provider powered by Candle.

Uses a small BERT-based model to generate embeddings entirely on CPU, with no external API calls. The model is downloaded from Hugging Face on first use and cached in the HF cache directory.

Implementations

impl CandleEmbeddingProvider

pub fn new() -> MenteResult<Self>

Create a new provider with the default model (all-MiniLM-L6-v2).

Downloads the model on first use. Subsequent calls load from cache.

pub fn with_model(model_id: &str) -> MenteResult<Self>

Create a new provider with a specific Hugging Face model ID.

pub fn with_cache_dir(cache_dir: PathBuf) -> MenteResult<Self>

Create a new provider with a custom cache directory.

Trait Implementations

impl EmbeddingProvider for CandleEmbeddingProvider

fn embed(&self, text: &str) -> MenteResult<Vec<f32>>

Generate an embedding vector for a single text.

fn embed_batch(&self, texts: &[&str]) -> MenteResult<Vec<Vec<f32>>>

Generate embedding vectors for a batch of texts.

fn dimensions(&self) -> usize

The dimensionality of the embedding vectors produced by this provider.

fn model_name(&self) -> &str

The name of the model used by this provider.