Struct EmbeddingService 

pub struct EmbeddingService { /* private fields */ }

Service for generating embeddings from spectrograms.

This is the main application service for the embedding bounded context. It coordinates between the model manager, ONNX inference, and domain entities.

Implementations

impl EmbeddingService

pub fn new(model_manager: Arc<ModelManager>, batch_size: usize) -> Self

Create a new embedding service.

Arguments

• model_manager - The model manager for ONNX model access
• batch_size - Maximum batch size for inference

pub fn with_config( model_manager: Arc<ModelManager>, config: EmbeddingServiceConfig, ) -> Self

Create with custom configuration

pub async fn embed_segment( &self, spectrogram: &Spectrogram, ) -> Result<EmbeddingOutput, EmbeddingError>

Generate an embedding from a single spectrogram.

Arguments

• spectrogram - The input spectrogram

Errors

Returns an error if inference fails or the embedding is invalid.

pub async fn embed_batch( &self, spectrograms: &[Spectrogram], ) -> Result<Vec<EmbeddingOutput>, EmbeddingError>

Generate embeddings for multiple spectrograms in batches.

This is more efficient than calling embed_segment multiple times as it uses batched inference.

Arguments

• spectrograms - Slice of input spectrograms

Errors

Returns an error if any inference fails. Partial results are not returned.

pub fn create_batch(&self, segment_ids: Vec<SegmentId>) -> EmbeddingBatch

Create a batch tracking object for monitoring progress.

pub fn model_version(&self) -> String

Get the current model version being used.

pub async fn is_ready(&self) -> bool

Check if the service is ready for inference.