Module embedding 

Embedding generation for semantic search.

Provides embedding generation using fastembed (when available) or a hash-based fallback for deterministic pseudo-embeddings.

Feature Flags

• fastembed-embeddings: Enables FastEmbed with BGE-M3 (1024 dimensions, 8192 token max)
• Without the feature: Uses hash-based fallback (deterministic but not semantic)

Structs

FallbackEmbedder

Hash-based fallback embedder.

FastEmbedEmbedder

FastEmbed embedder using BGE-M3.

Constants

DEFAULT_DIMENSIONS

Default embedding dimensions for the BGE-M3 model.

Traits

Embedder

Trait for embedding generators.

Functions

cosine_similarity

Computes cosine similarity between two embedding vectors.

create_embedder

Creates the default embedder based on available features.