Struct OnnxEmbedding 

pub struct OnnxEmbedding { /* private fields */ }

ONNX-based local embedding provider.

Wraps an ONNX sentence-transformer model for on-device vector generation. When the onnx feature is not enabled, embed and embed_batch return an Error::Embedding explaining how to enable full inference.

Implementations

impl OnnxEmbedding

pub fn new(model_path: &str, dimensions: usize) -> Result<Self>

Create a new ONNX embedding provider from a model path.

The model should be an ONNX sentence-transformer model (e.g., all-MiniLM-L6-v2 exported to ONNX format).

Errors

Returns Error::Validation if the file at model_path does not exist on disk.

pub fn model_path(&self) -> &str

Get the model path.

Trait Implementations

impl Debug for OnnxEmbedding

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

Formats the value using the given formatter.

impl EmbeddingProvider for OnnxEmbedding

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 vector for a single text input.

Errors

Currently returns Error::Embedding because full ONNX Runtime inference requires the onnx feature (with ort, tokenizers, and ndarray crates).

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 embedding vectors for a batch of text inputs.

Errors

Currently returns Error::Embedding because full ONNX Runtime inference requires the onnx feature (with ort, tokenizers, and ndarray crates).

fn dimensions(&self) -> usize