Class OpenAI<CallOptions>

Wrapper around OpenAI large language models.

To use you should have the openai package installed, with the OPENAI_API_KEY environment variable set.

To use with Azure you should have the openai package installed, with the AZURE_OPENAI_API_KEY, AZURE_OPENAI_API_INSTANCE_NAME, AZURE_OPENAI_API_DEPLOYMENT_NAME and AZURE_OPENAI_API_VERSION environment variable set.

Remarks

Any parameters that are valid to be passed to openai.createCompletion can be passed through modelKwargs, even if not explicitly available on this class.

Example

const model = new OpenAI({
  modelName: "gpt-4",
  temperature: 0.7,
  maxTokens: 1000,
  maxRetries: 5,
});

const res = await model.call(
  "Question: What would be a good company name for a company that makes colorful socks?\nAnswer:"
);
console.log({ res });

Type Parameters

Hierarchy

Implements

Constructors

constructor

Properties

CallOptions ParsedCallOptions batchSize caller frequencyPenalty maxTokens modelName n presencePenalty streaming temperature topP verbose azureOpenAIApiDeploymentName? azureOpenAIApiInstanceName? azureOpenAIApiKey? azureOpenAIApiVersion? azureOpenAIBasePath? bestOf? cache? callbacks? logitBias? metadata? modelKwargs? openAIApiKey? organization? stop? tags? timeout? user?

Accessors

callKeys

Methods

batch bind call completionWithRetry generate generatePrompt getNumTokens identifyingParams invocationParams invoke map pipe predict predictMessages serialize stream streamLog toJSON toJSONNotImplemented transform withConfig withFallbacks withRetry deserialize isRunnable

Constructors

constructor

Properties

CallOptions

CallOptions: CallOptions

ParsedCallOptions

ParsedCallOptions: Omit<CallOptions, never>

batchSize

batchSize: number

Batch size to use when passing multiple documents to generate

caller

caller: AsyncCaller

The async caller should be used by subclasses to make any async calls, which will thus benefit from the concurrency and retry logic.

frequencyPenalty

frequencyPenalty: number

Penalizes repeated tokens according to frequency

maxTokens

maxTokens: number

Maximum number of tokens to generate in the completion. -1 returns as many tokens as possible given the prompt and the model's maximum context size.

modelName

modelName: string

Model name to use

n

n: number

Number of completions to generate for each prompt

presencePenalty

presencePenalty: number

Penalizes repeated tokens

streaming

streaming: boolean

Whether to stream the results or not. Enabling disables tokenUsage reporting

temperature

temperature: number

Sampling temperature to use

topP

topP: number

Total probability mass of tokens to consider at each step

verbose

verbose: boolean

Whether to print out response text.

Optional azureOpenAIApiDeploymentName

azureOpenAIApiDeploymentName?: string

Azure OpenAI API deployment name to use for completions when making requests to Azure OpenAI. This is the name of the deployment you created in the Azure portal. e.g. "my-openai-deployment" this will be used in the endpoint URL: https://{InstanceName}.openai.azure.com/openai/deployments/my-openai-deployment/

Optional azureOpenAIApiInstanceName

azureOpenAIApiInstanceName?: string

Azure OpenAI API instance name to use when making requests to Azure OpenAI. this is the name of the instance you created in the Azure portal. e.g. "my-openai-instance" this will be used in the endpoint URL: https://my-openai-instance.openai.azure.com/openai/deployments/{DeploymentName}/

Optional azureOpenAIApiKey

azureOpenAIApiKey?: string

API key to use when making requests to Azure OpenAI.

Optional azureOpenAIApiVersion

azureOpenAIApiVersion?: string

API version to use when making requests to Azure OpenAI.

Optional azureOpenAIBasePath

azureOpenAIBasePath?: string

Custom endpoint for Azure OpenAI API. This is useful in case you have a deployment in another region. e.g. setting this value to "https://westeurope.api.cognitive.microsoft.com/openai/deployments" will be result in the endpoint URL: https://westeurope.api.cognitive.microsoft.com/openai/deployments/{DeploymentName}/

Optional bestOf

bestOf?: number

Generates bestOf completions server side and returns the "best"

Optional cache

cache?: BaseCache<Generation[]>

Optional callbacks

callbacks?: Callbacks

Optional logitBias

logitBias?: Record<string, number>

Dictionary used to adjust the probability of specific tokens being generated

Optional metadata

metadata?: Record<string, unknown>

Optional modelKwargs

modelKwargs?: Record<string, any>

Holds any additional parameters that are valid to pass to openai.createCompletion that are not explicitly specified on this class.

Optional openAIApiKey

openAIApiKey?: string

API key to use when making requests to OpenAI. Defaults to the value of OPENAI_API_KEY environment variable.

Optional organization

organization?: string

Optional stop

stop?: string[]

List of stop words to use when generating

Optional tags

tags?: string[]

Optional timeout

timeout?: number

Timeout to use when making requests to OpenAI.

Optional user

user?: string

Unique string identifier representing your end-user, which can help OpenAI to monitor and detect abuse.

Accessors

callKeys

  • get callKeys(): string[]

  • Keys that the language model accepts as call options.

    Returns string[]

Methods

batch

  • Default implementation of batch, which calls invoke N times. Subclasses should override this method if they can batch more efficiently.

    Parameters

    • inputs: BaseLanguageModelInput[]

      Array of inputs to each batch call.

    • Optional options: Partial<CallOptions> | Partial<CallOptions>[]

      Either a single call options object to apply to each batch call or an array for each call.

    • Optional batchOptions: RunnableBatchOptions & {
          returnExceptions?: false;
      }

    Returns Promise<string[]>

    An array of RunOutputs, or mixed RunOutputs and errors if batchOptions.returnExceptions is set

  • Parameters

    Returns Promise<(string | Error)[]>

  • Parameters

    Returns Promise<(string | Error)[]>

bind

  • Bind arguments to a Runnable, returning a new Runnable.

    Parameters

    • kwargs: Partial<CallOptions>

    Returns Runnable<BaseLanguageModelInput, string, CallOptions>

    A new RunnableBinding that, when invoked, will apply the bound args.

call

  • Convenience wrapper for generate that takes in a single string prompt and returns a single string output.

    Parameters

    • prompt: string
    • Optional options: string[] | CallOptions
    • Optional callbacks: Callbacks

    Returns Promise<string>

completionWithRetry

  • Calls the OpenAI API with retry logic in case of failures.

    Parameters

    • request: CompletionCreateParamsStreaming

      The request to send to the OpenAI API.

    • Optional options: OpenAICoreRequestOptions

      Optional configuration for the API call.

    Returns Promise<AsyncIterable<Completion>>

    The response from the OpenAI API.

  • Parameters

    • request: CompletionCreateParamsNonStreaming
    • Optional options: OpenAICoreRequestOptions

    Returns Promise<Completion>

generate

  • Run the LLM on the given prompts and input, handling caching.

    Parameters

    • prompts: string[]
    • Optional options: string[] | CallOptions
    • Optional callbacks: Callbacks

    Returns Promise<LLMResult>

generatePrompt

  • This method takes prompt values, options, and callbacks, and generates a result based on the prompts.

    Parameters

    • promptValues: BasePromptValue[]

      Prompt values for the LLM.

    • Optional options: string[] | CallOptions

      Options for the LLM call.

    • Optional callbacks: Callbacks

      Callbacks for the LLM call.

    Returns Promise<LLMResult>

    An LLMResult based on the prompts.

getNumTokens

  • Parameters

    Returns Promise<number>

identifyingParams

  • Get the identifying parameters for the model

    Returns Omit<CompletionCreateParams, "prompt"> & {
        model_name: string;
    } & ClientOptions

invocationParams

  • Get the parameters used to invoke the model

    Parameters

    • Optional options: Omit<CallOptions, never>

    Returns Omit<CompletionCreateParams, "prompt">

invoke

  • This method takes an input and options, and returns a string. It converts the input to a prompt value and generates a result based on the prompt.

    Parameters

    Returns Promise<string>

    A string result based on the prompt.

map

pipe

  • Create a new runnable sequence that runs each individual runnable in series, piping the output of one runnable into another runnable or runnable-like.

    Type Parameters

    • NewRunOutput

    Parameters

    • coerceable: RunnableLike<string, NewRunOutput>

      A runnable, function, or object whose values are functions or runnables.

    Returns RunnableSequence<BaseLanguageModelInput, Exclude<NewRunOutput, Error>>

    A new runnable sequence.

predict

  • This method is similar to call, but it's used for making predictions based on the input text.

    Parameters

    • text: string

      Input text for the prediction.

    • Optional options: string[] | CallOptions

      Options for the LLM call.

    • Optional callbacks: Callbacks

      Callbacks for the LLM call.

    Returns Promise<string>

    A prediction based on the input text.

predictMessages

  • This method takes a list of messages, options, and callbacks, and returns a predicted message.

    Parameters

    • messages: BaseMessage[]

      A list of messages for the prediction.

    • Optional options: string[] | CallOptions

      Options for the LLM call.

    • Optional callbacks: Callbacks

      Callbacks for the LLM call.

    Returns Promise<BaseMessage>

    A predicted message based on the list of messages.

serialize

stream

  • Stream output in chunks.

    Parameters

    Returns Promise<IterableReadableStream<string>>

    A readable stream that is also an iterable.

streamLog

  • Stream all output from a runnable, as reported to the callback system. This includes all inner runs of LLMs, Retrievers, Tools, etc. Output is streamed as Log objects, which include a list of jsonpatch ops that describe how the state of the run has changed in each step, and the final state of the run. The jsonpatch ops can be applied in order to construct state.

    Parameters

    • input: BaseLanguageModelInput
    • Optional options: Partial<CallOptions>
    • Optional streamOptions: Omit<LogStreamCallbackHandlerInput, "autoClose">

    Returns AsyncGenerator<RunLogPatch, any, unknown>

toJSON

toJSONNotImplemented

transform

  • Default implementation of transform, which buffers input and then calls stream. Subclasses should override this method if they can start producing output while input is still being generated.

    Parameters

    Returns AsyncGenerator<string, any, unknown>

withConfig

withFallbacks

withRetry

Static deserialize

Static isRunnable

Settings

Member Visibility

Theme

On This Page

Generated using TypeDoc