openai

README

OpenAI Plugin

This plugin provides a simple interface for using OpenAI's services.

Prerequisites

• Go installed on your system
• An OpenAI API key

Usage

Here's a simple example of how to use the OpenAI plugin:

import (
  // ignoring Genkit imports
  oai "github.com/firebase/genkit/go/plugins/compat_oai/openai"
  "github.com/openai/openai-go"
)
// Initialize the OpenAI plugin with your API key
oai := &oai.OpenAI{APIKey: apiKey}

// Initialize Genkit with the OpenAI plugin
g, err := genkit.Init(ctx,
    genkit.WithDefaultModel("openai/gpt-4o-mini"),
    genkit.WithPlugins(oai),
)
if err != nil {
    // handle errors
}

config := &openai.ChatCompletionNewParams{
    // define optional config fields
}

resp, err = genkit.Generate(ctx, g,
    ai.WithPromptText("Write a short sentence about artificial intelligence."),
    ai.WithConfig(config),
)

Running Tests

First, set your OpenAI API key as an environment variable:

export OPENAI_API_KEY=<your-api-key>

Running All Tests

To run all tests in the directory:

go test -v .

Running Tests from Specific Files

To run tests from a specific file:

# Run only generate_live_test.go tests
go test -run "^TestGenerator"

# Run only openai_live_test.go tests
go test -run "^TestPlugin"

Running Individual Tests

To run a specific test case:

# Run only the streaming test from openai_live_test.go
go test -run "TestPlugin/streaming"

# Run only the Complete test from generate_live_test.go
go test -run "TestGenerator_Complete"

# Run only the Stream test from generate_live_test.go
go test -run "TestGenerator_Stream"

Test Output Verbosity

Add the -v flag for verbose output:

go test -v -run "TestPlugin/streaming"

Note: All live tests require the OPENAI_API_KEY environment variable to be set. Tests will be skipped if the API key is not provided.

Documentation

Index

• type EmbedderRef
• type OpenAI
  • func (o *OpenAI) DefineEmbedder(id string, opts *ai.EmbedderOptions) ai.Embedder
  • func (o *OpenAI) DefineModel(id string, opts ai.ModelOptions) ai.Model
  • func (o *OpenAI) Embedder(g *genkit.Genkit, name string) ai.Embedder
  • func (o *OpenAI) Init(ctx context.Context) []core.Action
  • func (o *OpenAI) ListActions(ctx context.Context) []core.ActionDesc
  • func (o *OpenAI) Model(g *genkit.Genkit, name string) ai.Model
  • func (o *OpenAI) Name() string
  • func (o *OpenAI) ResolveAction(atype core.ActionType, name string) core.Action
• type TextEmbeddingConfig

Types

type EmbedderRef

type EmbedderRef struct {
	Name         string
	ConfigSchema TextEmbeddingConfig // Represents the schema, can be used for default config
	Label        string
	Supports     *ai.EmbedderSupports
	Dimensions   int
}

EmbedderRef represents the main structure for an embedding model's definition.

type OpenAI

type OpenAI struct {
	// APIKey is the API key for the OpenAI API. If empty, the values of the environment variable "OPENAI_API_KEY" will be consulted.
	// Request a key at https://platform.openai.com/api-keys
	APIKey string
	// Optional: Opts are additional options for the OpenAI client.
	// Can include other options like WithOrganization, WithBaseURL, etc.
	Opts []option.RequestOption
	// contains filtered or unexported fields
}

func (*OpenAI) DefineEmbedder

func (o *OpenAI) DefineEmbedder(id string, opts *ai.EmbedderOptions) ai.Embedder

func (*OpenAI) DefineModel

func (o *OpenAI) DefineModel(id string, opts ai.ModelOptions) ai.Model

func (*OpenAI) Embedder

func (o *OpenAI) Embedder(g *genkit.Genkit, name string) ai.Embedder

func (*OpenAI) Init

func (o *OpenAI) Init(ctx context.Context) []core.Action

Init implements genkit.Plugin.

func (*OpenAI) ListActions

func (o *OpenAI) ListActions(ctx context.Context) []core.ActionDesc

func (*OpenAI) Model

func (o *OpenAI) Model(g *genkit.Genkit, name string) ai.Model

func (*OpenAI) Name

func (o *OpenAI) Name() string

Name implements genkit.Plugin.

func (*OpenAI) ResolveAction

func (o *OpenAI) ResolveAction(atype core.ActionType, name string) core.Action

type TextEmbeddingConfig

type TextEmbeddingConfig struct {
	Dimensions     int                                       `json:"dimensions,omitempty"`
	EncodingFormat openaiGo.EmbeddingNewParamsEncodingFormat `json:"encodingFormat,omitempty"`
}