Module embeddings 

Embedding provider abstraction for vector generation

This module provides a trait-based abstraction for embedding generation, supporting multiple providers (FastEmbed, OpenAI, Ollama) with a unified interface.

Architecture

┌──────────────────────────────────────────────────────────────┐
│                  EmbeddingProvider Trait                     │
│  embed_documents, embed_query, dimensions, model_name       │
└──────────────────────────────────────────────────────────────┘
                             │
         ┌───────────────────┼───────────────────┐
         ▼                   ▼                   ▼
  ┌─────────────┐    ┌─────────────┐    ┌─────────────┐
  │  FastEmbed  │    │   OpenAI    │    │   Ollama    │
  │  (local)    │    │   (API)     │    │  (local)    │
  └─────────────┘    └─────────────┘    └─────────────┘

Example

use skill_runtime::embeddings::{EmbeddingProvider, FastEmbedProvider, EmbeddingConfig};

// Create a provider
let provider = FastEmbedProvider::new(FastEmbedModel::AllMiniLM)?;

// Embed a query
let query_embedding = provider.embed_query("search for kubernetes tools").await?;

// Embed multiple documents
let texts = vec!["doc1".to_string(), "doc2".to_string()];
let embeddings = provider.embed_documents(texts).await?;

Structs

EmbeddingConfig

Configuration for embedding providers

EmbeddingProviderFactory

Factory for creating embedding providers from configuration

EmbeddingResult

Embedding result with metadata

FastEmbedProvider

FastEmbed embedding provider

OllamaProvider

Ollama embedding provider

OpenAIEmbedProvider

OpenAI embedding provider

Enums

EmbeddingProviderType

Supported embedding provider types

FastEmbedModel

FastEmbed model variants

OpenAIEmbeddingModel

OpenAI embedding model variants

Traits

EmbeddingProvider

Trait for embedding generation providers

Functions

create_provider

Convenience function to create a provider from configuration