Category: NodeJS

This post covers using the aws-sdk for NodeJS with AWS Translate.

The code examples are written in ES and transpiled with Babel.

Install the AWS SDK

First, install the aws-sdk package in your project using your favourite package manager:

yarn add aws-sdk
# OR
npm i aws-sdk

Ensure There’s a Working AWS Profile

Ensure that you have an AWS profile and configuration properly setup for your user. An AWS Profile is typically stored inside the ~/.aws folder inside your home directory.

Suppose you have a profile named firxworx. An example entry of a useful entry in ~/.aws/config for that profile is:

[profile firxworx]
region = ca-central-1
output = json

A corresponding entry in the ~/.aws/credentials file that specifies credentials for the example firxworx profile looks like this:

[firxworx]
aws_access_key_id=ABCDXXXX
aws_secret_access_key=ABCDXXXX

Refer to the AWS Docs if you need to create a profile and obtain an Access Key ID and Secret Access Key.

Write Your Code

Start by importing the aws-sdk package:

import AWS from 'aws-sdk'

Next, configure AWS by specifying which profile’s credentials to use:

const credentials = new AWS.SharedIniFileCredentials({ profile: 'firxworx' })
AWS.config.credentials = credentials

Specify any other config options. The following line locks AWS to the most current API version (at the time of writing):

AWS.config.apiVersions = {
  translate: '2017-07-01',
}

Reference the AWS Translate homepage and take note of which regions AWS Translate is currently available in. If you need to specify a region that’s different than the default listed in your AWS profile, or you wish for your code to be explicit about which region it’s using, add the following line. Change the region to the valid region that you would like to use:

AWS.config.update({
  region: 'ca-central-1'
})

If you are using any Custom Terminologies, be sure to define them in the same region that you are about to use for AWS Translate. Custom Terminologies are lists of translation overrides that can be uploaded into the AWS Console. They are useful for ensuring that brand names, terms of art, trademarks, etc are translated correctly. Custom Terminology definitions are only available within the region that they were created and saved in.

Next, create an instance of AWS Translate:

const awsTranslate = new AWS.Translate()

At this point everything is setup to write a function that can translate text.

The following implements an async function called awsTranslate(). The function’s params include specifying a hypothetical custom terminology named example-custom-terminology-v1. Do not specify any value in the TerminologyNames array if you are not using any custom terminologies.

A key insight here is the .promise() method in the line containing awsTranslate.translateText(params).promise() which causes the API to return a promise.

async function asyncTranslate(langFrom, langTo, text) {
  const params = {
    SourceLanguageCode: langFrom,
    TargetLanguageCode: langTo,
    Text: text,
    TerminologyNames: [
      'example-custom-terminology-v1'
    ]
  }

  try {
    const translation = await awsTranslate.translateText(params).promise()
    return translation.TranslatedText
  } catch (err) {
    console.log(err, err.stack)
  }
}

The langFrom and langTo must be language codes as understood by AWS Translate. Refer to the docs for a current list of supported language codes: https://docs.aws.amazon.com/translate/latest/dg/what-is.html.

If you had a hypothetical index.js entry point for your NodeJS application and wanted to use the above function, an example invocation could be:

(async () => {

  const translation = await asyncTranslate('en', 'fr', 'Hello World')
  console.log(translation)

})()