Options
All
  • Public
  • Public/Protected
  • All
Menu

SageMaker A2I Runtime Client - AWS SDK for JavaScript v3

@aws-sdk/client-sagemaker-a2i-runtime

NPM version NPM downloads

Description

AWS SDK for JavaScript SageMakerA2IRuntime Client for Node.js, Browser and React Native.

Amazon Augmented AI (Amazon A2I) adds the benefit of human judgment to any machine learning application. When an AI application can't evaluate data with a high degree of confidence, human reviewers can take over. This human review is called a human review workflow. To create and start a human review workflow, you need three resources: a worker task template, a flow definition, and a human loop.

For information about these resources and prerequisites for using Amazon A2I, see Get Started with Amazon Augmented AI in the Amazon SageMaker Developer Guide.

This API reference includes information about API actions and data types that you can use to interact with Amazon A2I programmatically. Use this guide to:

  • Start a human loop with the StartHumanLoop operation when using Amazon A2I with a custom task type. To learn more about the difference between custom and built-in task types, see Use Task Types . To learn how to start a human loop using this API, see Create and Start a Human Loop for a Custom Task Type in the Amazon SageMaker Developer Guide.

  • Manage your human loops. You can list all human loops that you have created, describe individual human loops, and stop and delete human loops. To learn more, see Monitor and Manage Your Human Loop in the Amazon SageMaker Developer Guide.

Amazon A2I integrates APIs from various AWS services to create and start human review workflows for those services. To learn how Amazon A2I uses these APIs, see Use APIs in Amazon A2I in the Amazon SageMaker Developer Guide.

Installing

To install the this package, simply type add or install @aws-sdk/client-sagemaker-a2i-runtime using your favorite package manager:

  • npm install @aws-sdk/client-sagemaker-a2i-runtime
  • yarn add @aws-sdk/client-sagemaker-a2i-runtime
  • pnpm add @aws-sdk/client-sagemaker-a2i-runtime

Getting Started

Import

The AWS SDK is modulized by clients and commands. To send a request, you only need to import the SageMakerA2IRuntimeClient and the commands you need, for example DeleteHumanLoopCommand:

// ES5 example
const { SageMakerA2IRuntimeClient, DeleteHumanLoopCommand } = require("@aws-sdk/client-sagemaker-a2i-runtime");
// ES6+ example
import { SageMakerA2IRuntimeClient, DeleteHumanLoopCommand } from "@aws-sdk/client-sagemaker-a2i-runtime";

Usage

To send a request, you:

  • Initiate client with configuration (e.g. credentials, region).
  • Initiate command with input parameters.
  • Call send operation on client with command object as input.
  • If you are using a custom http handler, you may call destroy() to close open connections.
// a client can be shared by different commands.
const client = new SageMakerA2IRuntimeClient({ region: "REGION" });

const params = {
  /** input parameters */
};
const command = new DeleteHumanLoopCommand(params);

Async/await

We recommend using await operator to wait for the promise returned by send operation as follows:

// async/await.
try {
  const data = await client.send(command);
  // process data.
} catch (error) {
  // error handling.
} finally {
  // finally.
}

Async-await is clean, concise, intuitive, easy to debug and has better error handling as compared to using Promise chains or callbacks.

Promises

You can also use Promise chaining to execute send operation.

client.send(command).then(
  (data) => {
    // process data.
  },
  (error) => {
    // error handling.
  }
);

Promises can also be called using .catch() and .finally() as follows:

client
  .send(command)
  .then((data) => {
    // process data.
  })
  .catch((error) => {
    // error handling.
  })
  .finally(() => {
    // finally.
  });

Callbacks

We do not recommend using callbacks because of callback hell, but they are supported by the send operation.

// callbacks.
client.send(command, (err, data) => {
  // process err and data.
});

v2 compatible style

The client can also send requests using v2 compatible style. However, it results in a bigger bundle size and may be dropped in next major version. More details in the blog post on modular packages in AWS SDK for JavaScript

import * as AWS from "@aws-sdk/client-sagemaker-a2i-runtime";
const client = new AWS.SageMakerA2IRuntime({ region: "REGION" });

// async/await.
try {
  const data = await client.deleteHumanLoop(params);
  // process data.
} catch (error) {
  // error handling.
}

// Promises.
client
  .deleteHumanLoop(params)
  .then((data) => {
    // process data.
  })
  .catch((error) => {
    // error handling.
  });

// callbacks.
client.deleteHumanLoop(params, (err, data) => {
  // process err and data.
});

Troubleshooting

When the service returns an exception, the error will include the exception information, as well as response metadata (e.g. request id).

try {
  const data = await client.send(command);
  // process data.
} catch (error) {
  const { requestId, cfId, extendedRequestId } = error.$$metadata;
  console.log({ requestId, cfId, extendedRequestId });
  /**
   * The keys within exceptions are also parsed.
   * You can access them by specifying exception names:
   * if (error.name === 'SomeServiceException') {
   *     const value = error.specialKeyInException;
   * }
   */
}

Getting Help

Please use these community resources for getting help. We use the GitHub issues for tracking bugs and feature requests, but have limited bandwidth to address them.

To test your universal JavaScript code in Node.js, browser and react-native environments, visit our code samples repo.

Contributing

This client code is generated automatically. Any modifications will be overwritten the next time the @aws-sdk/client-sagemaker-a2i-runtime package is updated. To contribute to client you can check our generate clients scripts.

License

This SDK is distributed under the Apache License, Version 2.0, see LICENSE for more information.