Documentation
¶
Overview ¶
Package gemini provides LLM and embedding support for Google's Gemini models.
Gemini is Google's family of multimodal AI models supporting chat completions, embeddings generation, and content generation with image support.
Index ¶
- Variables
- type LLM
- func (g *LLM) Call(ctx context.Context, prompt string, options ...llms.CallOption) (string, error)
- func (g *LLM) EmbedDocuments(ctx context.Context, texts []string) ([][]float32, error)
- func (g *LLM) EmbedQueries(ctx context.Context, texts []string) ([][]float32, error)
- func (g *LLM) EmbedQuery(ctx context.Context, text string) ([]float32, error)
- func (g *LLM) GenerateContent(ctx context.Context, messages []schema.MessageContent, ...) (*schema.ContentResponse, error)
- func (g *LLM) GetDimension(ctx context.Context) (int, error)
- type Option
Constants ¶
This section is empty.
Variables ¶
var ( ErrNoAPIKey = errors.New("gemini: API key is required") ErrInvalidModel = errors.New("gemini: invalid model specified") ErrNoContent = errors.New("gemini: no content generated") ErrSystemMessage = errors.New("gemini: system message must be the first message in the conversation") ErrEmbeddings = errors.New("gemini: failed to generate embeddings") )
Functions ¶
This section is empty.
Types ¶
type LLM ¶
type LLM struct {
// contains filtered or unexported fields
}
LLM implements both the Model and Embedder interfaces for Gemini.
func (*LLM) EmbedDocuments ¶ added in v0.9.0
EmbedDocuments generates embeddings for a slice of texts.
func (*LLM) EmbedQueries ¶ added in v0.9.0
EmbedQueries generates embeddings for multiple queries.
func (*LLM) EmbedQuery ¶ added in v0.9.0
EmbedQuery generates an embedding for a single text query.
func (*LLM) GenerateContent ¶
func (g *LLM) GenerateContent( ctx context.Context, messages []schema.MessageContent, options ...llms.CallOption, ) (*schema.ContentResponse, error)
GenerateContent handles multi-turn conversations and streaming.
func (*LLM) GetDimension ¶ added in v0.9.0
GetDimension returns the embedding dimension of the model. It caches the result after the first successful call. If the first call fails, subsequent calls will retry until a successful result is obtained.
type Option ¶
type Option func(*options)
Option is a function type for configuring the client.
func WithEmbeddingModel ¶ added in v0.9.0
WithEmbeddingModel sets the embedding model name.
func WithLogger ¶
WithLogger sets a custom logger for the client.
Source Files