Trait EmbeddingProvider 

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>, RagError>> + Send + 'async_trait>>
       where 'life0: 'async_trait,
             'life1: 'async_trait,
             Self: 'async_trait;
    fn dimensions(&self) -> usize;

    // 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>>, RagError>> + Send + 'async_trait>>
       where 'life0: 'async_trait,
             'life1: 'async_trait,
             'life2: 'async_trait,
             Self: 'async_trait { ... }
}

Available on crate feature rag only.

A provider that generates vector embeddings from text input.

Implementations wrap specific embedding backends (Gemini, OpenAI, etc.) behind a unified async interface. The default embed_batch implementation calls embed sequentially; backends that support native batching should override it.

Example

use adk_rag::EmbeddingProvider;

let provider = MyEmbeddingProvider::new();
let embedding = provider.embed("hello world").await?;
assert_eq!(embedding.len(), provider.dimensions());

Required Methods

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

where 'life0: 'async_trait, 'life1: 'async_trait, Self: 'async_trait,

Generate an embedding vector for a single text input.

fn dimensions(&self) -> usize

Return the dimensionality of embeddings produced by this provider.

Provided Methods

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

where 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, Self: 'async_trait,

Generate embedding vectors for a batch of text inputs.

The default implementation calls embed sequentially for each input. Override this method if the backend supports native batch embedding for better throughput.

Implementors