Struct MeteredEmbedder 

pub struct MeteredEmbedder<E>
where
    E: Embedder,
{ /* private fields */ }

Embedder decorator that emits OTel-compatible telemetry per call (and optional cost via EmbeddingCostCalculator).

Wraps any inner E: Embedder and itself implements Embedder, so the wrapper drops in transparently anywhere the bare type was used.

Implementations

impl<E> MeteredEmbedder<E>

where E: Embedder,

pub fn new(inner: E, model: impl Into<Arc<str>>) -> Self

Wrap inner with a metered surface. model is the wire-name the operator wants surfaced in telemetry (gen_ai.embedding.model) and used as the lookup key in the cost calculator’s pricing table.

pub fn from_arc(inner: Arc<E>, model: impl Into<Arc<str>>) -> Self

Variant for callers that already hold an Arc<E> (typical when the embedder is shared across multiple memory backends).

pub fn with_cost_calculator( self, calculator: Arc<dyn EmbeddingCostCalculator>, ) -> Self

Attach an EmbeddingCostCalculator. When set, the wrapper emits gen_ai.embedding.cost on the success branch of every embed / embed_batch call whose (tenant, model) resolves to a pricing row.

pub fn model(&self) -> &str

Borrow the operator-supplied model name surfaced in telemetry — useful for tests and for dashboards that label rows by model.

Trait Implementations

impl<E> Embedder for MeteredEmbedder<E>

where E: Embedder,

fn dimension(&self) -> usize

Output vector dimension. Used by VectorStore impls to validate inserts against the configured index dimension.

fn embed<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, text: &'life1 str, ctx: &'life2 ExecutionContext, ) -> Pin<Box<dyn Future<Output = Result<Embedding>> + Send + 'async_trait>>

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

Embed one input string. Returns the vector plus optional usage metadata so downstream cost meters can charge the call without a second round-trip.

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

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

Batch embed. Default impl runs sequentially via embed, polling ExecutionContext::is_cancelled between iterations so a long batch bails out within one embed round-trip of cancellation rather than draining the full pool. Implementations that support a true batch endpoint should override — sequential calls amplify network latency by N.