Struct EmbeddingEngine 

pub struct EmbeddingEngine { /* private fields */ }

Embedding engine for converting text to vectors.

Wraps different embedding providers with a unified interface.

Implementations

impl EmbeddingEngine

pub fn new(dimensions: usize) -> Self

Create a new embedding engine with hash-based embeddings.

Use this for development, testing, or offline scenarios. Hash embeddings are fast and free but lower quality.

pub fn from_env() -> Self

Create from environment variables.

Reads OPENAI_API_KEY. Falls back to hash embeddings if not set.

pub fn from_env_required() -> Result<Self>

Create from environment, returning error if not configured.

pub fn with_openai(api_key: impl Into<String>, model: Option<String>) -> Self

Create with OpenAI embeddings.

pub fn with_openai_config( api_key: impl Into<String>, config: OpenAIConfig, ) -> Self

Create with OpenAI using custom configuration.

pub fn with_provider(provider: Box<dyn EmbeddingProvider>) -> Self

Create with a custom provider.

pub fn dimensions(&self) -> usize

Get the embedding dimensions.

pub fn embed(&self, text: &str) -> Result<Vec<f32>>

Generate an embedding for the given text (sync wrapper).

pub async fn embed_async(&self, text: &str) -> Result<Vec<f32>>

Generate an embedding asynchronously.

pub async fn embed_batch(&self, texts: &[&str]) -> Result<Vec<Vec<f32>>>

Generate embeddings for multiple texts.

pub fn similarity(&self, a: &[f32], b: &[f32]) -> f32

Compute similarity between two embeddings.