Eino: Embedding guide

type Embedder interface {
    EmbedStrings(ctx context.Context, texts []string, opts ...Option) ([][]float64, error)
}

type Options struct {
    // Model is the name of the model used to generate vectors
    Model *string
}

// Set model name
WithModel(model string) Option

import "github.com/cloudwego/eino-ext/components/embedding/openai"

embedder, _ := openai.NewEmbedder(ctx, &openai.EmbeddingConfig{
    APIKey:     accessKey,
    Model:      "text-embedding-3-large",
    Dimensions: &defaultDim,
    Timeout:    0,
})

vectorIDs, _ := embedder.EmbedStrings(ctx, []string{"hello", "how are you"})

// Use in Chain
chain := compose.NewChain[[]string, [][]float64]()
chain.AppendEmbedding(embedder)

// Use in Graph
graph := compose.NewGraph[[]string, [][]float64]()
graph.AddEmbeddingNode("embedding_node", embedder)

// Use options (example for standalone usage)
vectors, err := embedder.EmbedStrings(ctx, texts,
    embedding.WithModel("text-embedding-3-small"),
)

import (
    "github.com/cloudwego/eino/callbacks"
    "github.com/cloudwego/eino/components/embedding"
    "github.com/cloudwego/eino/compose"
    callbacksHelper "github.com/cloudwego/eino/utils/callbacks"
    "github.com/cloudwego/eino-ext/components/embedding/openai"
)

handler := &callbacksHelper.EmbeddingCallbackHandler{
    OnStart: func(ctx context.Context, runInfo *callbacks.RunInfo, input *embedding.CallbackInput) context.Context {
       log.Printf("input access, len: %v, content: %s\n", len(input.Texts), input.Texts)
       return ctx
    },
    OnEnd: func(ctx context.Context, runInfo *callbacks.RunInfo, output *embedding.CallbackOutput) context.Context {
       log.Printf("output finished, len: %v\n", len(output.Embeddings))
       return ctx
    },
}

callbackHandler := callbacksHelper.NewHandlerHelper().Embedding(handler).Handler()

chain := compose.NewChain[[]string, [][]float64]()
chain.AppendEmbedding(embedder)

// Compile and run
runnable, _ := chain.Compile(ctx)
vectors, _ = runnable.Invoke(ctx, []string{"hello", "how are you"},
    compose.WithCallbacks(callbackHandler))

log.Printf("vectors in chain: %v", vectors)

// Define the Option struct
type MyEmbeddingOptions struct {
    BatchSize int
    MaxRetries int
    Timeout time.Duration
}

// Define the Option function
func WithBatchSize(size int) embedding.Option {
    return embedding.WrapEmbeddingImplSpecificOptFn(func(o *MyEmbeddingOptions) {
        o.BatchSize = size
    })
}

// CallbackInput is the input for the embedding callback
type CallbackInput struct {
    // Texts are the list of texts to be converted
    Texts []string
    // Config is the configuration information for generating vectors
    Config *Config
    // Extra is additional information for the callback
    Extra map[string]any
}

// CallbackOutput is the output for the embedding callback
type CallbackOutput struct {
    // Embeddings are the list of generated vectors
    Embeddings [][]float64
    // Config is the configuration information for generating vectors
    Config *Config
    // TokenUsage is the token usage information
    TokenUsage *TokenUsage
    // Extra is additional information for the callback
    Extra map[string]any
}

// TokenUsage is the token usage information
type TokenUsage struct {
    // PromptTokens is the number of tokens in the prompt
    PromptTokens int
    // CompletionTokens is the number of tokens in the completion
    CompletionTokens int
    // TotalTokens is the total number of tokens
    TotalTokens int
}

type MyEmbedder struct {
    model string
    batchSize int
}

func NewMyEmbedder(config *MyEmbedderConfig) (*MyEmbedder, error) {
    return &MyEmbedder{
        model: config.DefaultModel,
        batchSize: config.DefaultBatchSize,
    }, nil
}

func (e *MyEmbedder) EmbedStrings(ctx context.Context, texts []string, opts ...embedding.Option) ([][]float64, error) {
    // 1. Handle the options
    options := &MyEmbeddingOptions{
        Options: &embedding.Options{},
        BatchSize: e.batchSize,
    }
    options.Options = embedding.GetCommonOptions(options.Options, opts...)
    options = embedding.GetImplSpecificOptions(options.Options, opts...)
    
    // 2. Get the callback manager
    cm := callbacks.ManagerFromContext(ctx)
    
    // 3. Trigger the pre-generation callback
    ctx = cm.OnStart(ctx, info, &embedding.CallbackInput{
        Texts: texts,
        Config: &embedding.Config{
            Model: e.model,
        },
    })
    
    // 4. Execute the vector generation logic
    vectors, tokenUsage, err := e.doEmbed(ctx, texts, options)
    
    // 5. Handle errors and trigger the completion callback
    if err != nil {
        ctx = cm.OnError(ctx, info, err)
        return nil, err
    }
    
    ctx = cm.OnEnd(ctx, info, &embedding.CallbackOutput{
        Embeddings: vectors,
        Config: &embedding.Config{
            Model: e.model,
        },
        TokenUsage: tokenUsage,
    })
    
    return vectors, nil
}

func (e *MyEmbedder) doEmbed(ctx context.Context, texts []string, opts *MyEmbeddingOptions) ([][]float64, *TokenUsage, error) {
    // Implementation logic
    return vectors, tokenUsage, nil
}