anthropic

Documentation

Index

• Variables
• func GetChatPrompt(chat []Message) string
• func GetPrompt(userQuestion string) string
• type Client
  • func NewClient(apiKey string) (*Client, error)
  • func (c *Client) Complete(req *CompletionRequest, callback StreamCallback) (*CompletionResponse, error)
• type CompletionRequest
• type CompletionResponse
• type Message
• type Model
• type StreamCallback
• type StreamResponse

Variables

var (
	ErrAnthropicInvalidRequest = errors.New("invalid request: there was an issue with the format or content of your request")
	ErrAnthropicUnauthorized   = errors.New("unauthorized: there's an issue with your API key")
	ErrAnthropicForbidden      = errors.New("forbidden: your API key does not have permission to use the specified resource")
	ErrAnthropicRateLimit      = errors.New("your account has hit a rate limit")
	ErrAnthropicInternalServer = errors.New("an unexpected error has occurred internal to Anthropic's systems")

	ErrAnthropicApiKeyRequired = errors.New("apiKey is required")
)

Functions

func GetChatPrompt

func GetChatPrompt(chat []Message) string

GetChatPrompt constructs a prompt string from a series of messages in a chat conversation.

func GetPrompt

func GetPrompt(userQuestion string) string

GetPrompt returns a prompt string that can be used to complete a user question.

Types

type Client

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

Client represents the Anthropic API client and its configuration.

func NewClient

func NewClient(apiKey string) (*Client, error)

NewClient initializes a new Anthropic API client with the required headers.

func (*Client) Complete

func (c *Client) Complete(req *CompletionRequest, callback StreamCallback) (*CompletionResponse, error)

Complete sends a completion request to the API and returns a single completion or a stream of completions.

type CompletionRequest

type CompletionRequest struct {
	Prompt            string   `json:"prompt"`
	Model             Model    `json:"model"`
	MaxTokensToSample int      `json:"max_tokens_to_sample"`
	StopSequences     []string `json:"stop_sequences,omitempty"` // optional
	Stream            bool     `json:"stream,omitempty"`         // optional
	Temperature       float64  `json:"temperature,omitempty"`    // optional
	TopK              int      `json:"top_k,omitempty"`          // optional
	TopP              float64  `json:"top_p,omitempty"`          // optional
}

CompletionRequest is the request to the Anthropic API for a completion.

type CompletionResponse

type CompletionResponse struct {
	Completion string `json:"completion"`
	StopReason string `json:"stop_reason"`
	Stop       string `json:"stop"`
}

CompletionResponse is the response from the Anthropic API for a completion request.

type Message

type Message struct {
	Sender  string // The sender's name (e.g., "Human" or a username)
	Content string // The content of the message
}

Message represents a single message in a chat conversation.

type Model

type Model string

Model represents a Claude model.

const (
	// Superior performance on tasks that require complex reasoning.
	ClaudeV2 Model = "claude-2"

	// Our largest model, ideal for a wide range of more complex tasks.
	ClaudeV1 Model = "claude-v1"

	// An enhanced version of ClaudeV1 with a 100,000 token context window.
	ClaudeV1_100k Model = "claude-v1-100k"

	// A smaller model with far lower latency, sampling at roughly 40 words/sec!
	ClaudeInstantV1 Model = "claude-instant-v1"

	// An enhanced version of ClaudeInstantV1 with a 100,000 token context window.
	ClaudeInstantV1_100k Model = "claude-instant-v1-100k"

	// More robust against red-team inputs, better at precise instruction-following,
	// better at code, and better and non-English dialogue and writing.
	ClaudeV1_3 Model = "claude-v1.3"

	// An enhanced version of ClaudeV1_3 with a 100,000 token context window.
	ClaudeV1_3_100k Model = "claude-v1.3-100k"

	// An improved version of ClaudeV1, slightly improved at general helpfulness,
	// instruction following, coding, and other tasks. It is also considerably
	// better with non-English languages.
	ClaudeV1_2 Model = "claude-v1.2"

	// An earlier version of ClaudeV1.
	ClaudeV1_0 Model = "claude-v1.0"

	// Latest version of ClaudeInstantV1. It is better at a wide variety of tasks
	// including writing, coding, and instruction following. It performs better on
	// academic benchmarks, including math, reading comprehension, and coding tests.
	ClaudeInstantV1_1 Model = "claude-instant-v1.1"

	// An enhanced version of ClaudeInstantV1_1 with a 100,000 token context window.
	ClaudeInstantV1_1_100k Model = "claude-instant-v1.1-100k"

	// An earlier version of ClaudeInstantV1.
	ClaudeInstantV1_0 Model = "claude-instant-v1.0"
)

type StreamCallback

type StreamCallback func(*CompletionResponse) error

StreamCallback is a function that handles a stream of completions.

type StreamResponse

type StreamResponse struct {
	Completion string `json:"completion"`
	StopReason string `json:"stop_reason"`
	Model      string `json:"model"`
	Stop       string `json:"stop"`
	LogID      string `json:"log_id"`
}

StreamResponse is the response from the Anthropic API for a stream of completions.