embedding

Documentation

Overview

Package embedding provides concrete Embedder implementations for generating text embedding vectors. Supports OpenAI, Google, and OpenAI-compatible endpoints (Ollama, LiteLLM, etc.).

Embedders satisfy the resume.Embedder interface and can be used for resume embedding, query embedding, and semantic similarity search.

Index

• type Config
• type Embedder
  • func New(cfg Config) (Embedder, error)

Types

type Config

type Config struct {
	// Provider name: "openai", "google", "openai-compat", "litellm", "none"
	Provider string `toml:"provider" json:"provider"`

	// Model name (e.g., "text-embedding-3-small", "text-embedding-004")
	Model string `toml:"model" json:"model"`

	// APIKey for the embedding provider
	APIKey string `toml:"api_key" json:"api_key"`

	// BaseURL for custom endpoints (OpenAI-compatible providers)
	BaseURL string `toml:"base_url" json:"base_url"`
}

Config holds configuration for creating an Embedder.

type Embedder

type Embedder interface {
	Embed(ctx context.Context, text string) ([]float64, error)
}

Embedder generates embedding vectors from text. Same interface as resume.Embedder — implementations satisfy both.

func New

func New(cfg Config) (Embedder, error)

New creates an Embedder from the given configuration. Returns nil if provider is "none" or empty.