Struct OllamaEmbeddingEngine 

pub struct OllamaEmbeddingEngine { /* private fields */ }

Embedding engine that calls the Ollama /api/embed HTTP endpoint.

Sends a batched array input per request, sub-batched by batch_size, and falls back to one concurrent request per text (via futures::future::join_all) for servers that do not accept array input. Transient HTTP errors (network failures, 429, 5xx) are retried with exponential back-off starting at 8 s (doubling to 128 s) for up to 128 s total.

Response shapes

Ollama can return embeddings in three shapes depending on the version and endpoint:

• {"embeddings": [[...]]} — standard /api/embed response
• {"embedding": [...]} — legacy single-embedding response
• {"data": [{"embedding": [...]}]} — OpenAI-compatible shape

All three shapes are handled transparently.

Implementations

impl OllamaEmbeddingEngine

pub fn new(config: &EmbeddingConfig) -> EmbeddingResult<Self>

Construct a new engine from the given EmbeddingConfig.

Returns EmbeddingError::ConfigError if the reqwest client cannot be built (e.g. invalid TLS or API key header value).

Trait Implementations

impl EmbeddingEngine for OllamaEmbeddingEngine

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

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

Embed a batch of text strings into normalized vectors

fn dimension(&self) -> usize

Get the dimensionality of embeddings produced by this engine

fn batch_size(&self) -> usize

Get the optimal batch size for this engine

fn max_sequence_length(&self) -> usize

Get the maximum sequence length (in tokens) supported