Skip to main content

EmbeddingProvider

fabryk_vector::embedding

Trait EmbeddingProvider 

Source 
pub trait EmbeddingProvider: Send + Sync {
    // Required methods
    fn embed<'life0, 'life1, 'async_trait>(
        &'life0 self,
        text: &'life1 str,
    ) -> Pin<Box<dyn Future<Output = Result<Vec<f32>>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait;
    fn dimension(&self) -> usize;
    fn name(&self) -> &str;

    // Provided method
    fn embed_batch<'life0, 'life1, 'life2, 'async_trait>(
        &'life0 self,
        texts: &'life1 [&'life2 str],
    ) -> Pin<Box<dyn Future<Output = Result<Vec<Vec<f32>>>> + Send + 'async_trait>>
       where Self: 'async_trait,
             'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait { ... }
}
Expand description

Trait for generating text embeddings.

Implementations wrap specific embedding libraries (fastembed, OpenAI, etc.) and provide a uniform async interface. The trait requires Send + Sync to allow safe sharing across async tasks.

§Thread Safety

Implementations should handle internal synchronization (e.g., Arc<Mutex<>>) for thread-unsafe underlying libraries.

Required Methods§

Source

fn embed<'life0, 'life1, 'async_trait>( &'life0 self, text: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<Vec<f32>>> + Send + 'async_trait>>
where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Generate an embedding for a single text.

Source

fn dimension(&self) -> usize

The embedding dimension.

Source

fn name(&self) -> &str

The provider name for diagnostics.

Provided Methods§

Source

fn embed_batch<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, texts: &'life1 [&'life2 str], ) -> Pin<Box<dyn Future<Output = Result<Vec<Vec<f32>>>> + Send + 'async_trait>>
where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Generate embeddings for a batch of texts.

Default implementation calls embed for each text sequentially. Backends that support native batching should override this.

Implementors§