Module embedding 

Embedding generation and async queue management (RML-873)

Supports multiple embedding backends:

• OpenAI API (text-embedding-3-small) - requires openai feature
• TF-IDF fallback (no external dependencies)

Features:

• LRU embedding cache with zero-copy Arc<f32> sharing
• Async queue processing for batch operations

Feature Flags

• openai: Enables OpenAI embedding backend (requires API key)

Structs

EmbeddingCache

Thread-safe LRU embedding cache with bytes-based capacity

EmbeddingCacheStats

Statistics for the embedding cache

EmbeddingProviderInfo

Static metadata describing an embedding provider.

EmbeddingQueue

Embedding queue for async processing

EmbeddingRegistry

A runtime registry of named EmbeddingProvider implementations.

EmbeddingWorker

Background worker for processing embeddings

OpenAIEmbedder

OpenAI embedding client

TfIdfEmbedder

TF-IDF based embedder using hashing trick

Traits

Embedder

Trait for embedding generators

EmbeddingProvider

An Embedder that also exposes self-describing metadata.

Functions

cosine_similarity

Cosine similarity between two vectors

create_embedder

Create an embedder from configuration

get_embedding

Get embedding for a memory

get_embedding_status

Get embedding status for a memory