embedder

type CohereConfig struct {
	// APIKey for Cohere API (required).
	APIKey string

	// BaseURL for the API (default: https://api.cohere.com).
	BaseURL string

	// Model name (default: embed-english-v3.0).
	// Supported: embed-english-v3.0, embed-multilingual-v3.0, embed-v4.0, etc.
	Model string

	// Dimension of embeddings (auto-detected from model if 0).
	Dimension int

	// Timeout for API requests (default: 30s).
	Timeout time.Duration

	// BatchSize for batch embedding (default: 96, Cohere's max per request).
	BatchSize int

	// InputType specifies the type of input (required for v3+ models).
	// Values: "search_document", "search_query", "classification", "clustering"
	// Default: "search_document"
	InputType string

	// OutputDimension for v4+ models (optional).
	// Values: 256, 512, 1024, 1536
	// If set, overrides model's default dimension.
	OutputDimension *int

	// Truncate specifies how to handle inputs longer than max tokens.
	// Values: "NONE", "START", "END" (default: "END")
	Truncate string
}

type CohereEmbedder struct {
	// contains filtered or unexported fields
}

type Embedder interface {
	// Embed converts text to a vector embedding.
	Embed(ctx context.Context, text string) ([]float32, error)

	// EmbedBatch converts multiple texts to vector embeddings.
	// More efficient than calling Embed multiple times.
	EmbedBatch(ctx context.Context, texts []string) ([][]float32, error)

	// Dimension returns the embedding vector dimension.
	Dimension() int

	// Model returns the model name being used.
	Model() string

	// Close releases any resources held by the embedder.
	Close() error
}

type OllamaConfig struct {
	// BaseURL for Ollama API (default: http://localhost:11434).
	BaseURL string

	// Model name (default: nomic-embed-text).
	Model string

	// Dimension of embeddings (default: 768 for nomic-embed-text).
	Dimension int

	// Timeout for API requests (default: 30s).
	Timeout time.Duration
}

type OllamaEmbedder struct {
	// contains filtered or unexported fields
}

type OpenAIConfig struct {
	// APIKey for OpenAI API (required).
	APIKey string

	// BaseURL for the API (default: https://api.openai.com/v1).
	BaseURL string

	// Model name (default: text-embedding-3-small).
	Model string

	// Dimension of embeddings (auto-detected from model if 0).
	// For text-embedding-3 models, this maps to the 'dimensions' API parameter.
	Dimension int

	// Timeout for API requests (default: 30s).
	Timeout time.Duration

	// BatchSize for batch embedding (default: 100).
	// Note: OpenAI supports up to 2048 inputs per request, but we use 100 as default
	// to stay within token limits (300,000 tokens total per request).
	BatchSize int

	// EncodingFormat specifies the format to return embeddings in.
	// Values: "float" (default), "base64"
	EncodingFormat string

	// User is a unique identifier representing your end-user.
	// Can help OpenAI monitor and detect abuse.
	User string
}

type OpenAIEmbedder struct {
	// contains filtered or unexported fields
}