Using Amazon Translate with Amazon DynamoDB

This example shows you how to translate a product review and store it in Amazon DynamoDB. If you request the same review later, DynamoDB returns it without Amazon Translate needing to translate it again.

In this example, you:

• Use AWS CloudFormation to create DynamoDB tables to store the translation and a Lambda function that calls the TranslateText operation.

• Test the function using the AWS Lambda console.

To run the example

1. Copy the contents of example.py, which you can find in Python Lambda function, to a file named example.py. example.py is a Lambda function that calls the TranslateText operation. Compress the file to a zip archive named example.zip. Store it in an S3 bucket in the same AWS Region where you want to run the function.

2. Create a new file named template.yaml. Copy the AWS CloudFormation template code, which you can find in AWS CloudFormation template, into the file. AWS CloudFormation uses the template to create resources for the sample application. Change BUCKET_NAME to the name of the S3 bucket that contains example.zip. Save the file in a local directory.

3. Sign in to the AWS Management Console and open the AWS CloudFormation console at https://console.aws.amazon.com/cloudformation.

4. Choose Create new stack.

5. Choose Upload a template to Amazon S3, and then choose Choose file. Choose template.yaml, that you created in Step 2, then Next.

6. Type a name for the stack, then choose Next.

7. On the Options page, choose Next.

8. Choose I acknowledge that AWS CloudFormation might create IAM resources and I acknowledge that AWS CloudFormation might create IAM resources with custom names. For more information, see Controlling Access with AWS Identity and Access Management in the AWS CloudFormation User Guide.

9. Choose Create Change Set.

10. After AWS CloudFormation creates the change set, choose Execute. Wait until AWS CloudFormation creates the stack.

11. Sign in to the AWS Management Console and open the AWS Lambda console at https://console.aws.amazon.com/lambda/.

12. Choose the new function. Its name starts with TestTranslate-ReviewTranslate.

13. On the function detail page, choose Test.

14. For Event name, type TestTranslate. For Configure test event, replace the JSON with the following:

    {
       "review": "hello world",
       "target_language": "es",
       "source_language": "en",
       "review_id": "1"
    }

    Choose Create.

15. Make sure that TestTranslate is selected, then choose Test. When the test finishes, you receive the following message:

    

Example code

Use the following code to create the example.

Python Lambda function

The following is the contents of the Python Lambda function. The Lambda function call the TranslateText operation and passes the review, the source language, and the target language to get the translated review. Save this file as example.py and them compress it in a .zip archive called example.zip. Save the file in an S3 bucket in the same region that you are running the example..

import logging
import json
import boto3
import os

translate = boto3.client('translate')
dynamodb = boto3.client('dynamodb')
firehose = boto3.client('firehose')

TABLE_NAME = os.getenv('TABLE_NAME')

logger = logging.getLogger()
logger.setLevel(logging.INFO)

def lambda_handler(event, context):

    logger.info(event)

    if 'source_language' in event and 'target_language' in event and 'review' in event and 'review_id' in event:
        review_id = event['review_id']
        source_language = event['source_language']
        target_language = event['target_language']
        review = event['review']

        try:
            # The Lambda function queries the Amazon DynamoDB table to check whether 
            # the review has already been translated. If the translated review 
            # is already stored in Amazon DynamoDB, the function returns it.
            response = dynamodb.get_item(
                TableName=TABLE_NAME,
                Key={
                    'review_id': {
                        'N': review_id,
                    },
                    'language': {
                        'S': target_language,
                    },
                }
            )
            logger.info(response)
            if 'Item' in response:
                return response['Item']['review']['S']
        except Exception as e:
            logger.error(response)
            raise Exception("[ErrorMessage]: " + str(e))

        try:
            # The Lambda function calls the TranslateText operation and passes the 
            # review, the source language, and the target language to get the 
            # translated review. 
            result = translate.translate_text(Text=review, SourceLanguageCode=source_language, TargetLanguageCode=target_language)
            logging.info("Translation output: " + str(result))
        except Exception as e:
            logger.error(response)
            raise Exception("[ErrorMessage]: " + str(e))

        try:
            # After the review is translated, the function stores it using
            # the Amazon DynamoDB putItem operation. Subsequent requests
            # for this translated review are returned from Amazon DynamoDB.
            response = dynamodb.put_item(
            TableName=TABLE_NAME,
            Item={
                'review_id': {
                    'N': review_id,
                },
                'language': {
                    'S': target_language,
                },
                'review': {
                    'S': result.get('TranslatedText')
                }
                }
            )
            logger.info(response)
        except Exception as e:
                logger.error(e)
                raise Exception("[ErrorMessage]: " + str(e))
        return result.get('TranslatedText')
    else:
        logger.error(e)
        raise Exception("[ErrorMessage]: Invalid input ")

AWS CloudFormation template

The following is the template file that you use with AWS CloudFormation to create and configure the Lambda function and the DynamoDB tables. Use this file when you create the AWS CloudFormation stack for the example. Update BUCKET_NAME to the name of the S3 bucket that contains the example.zip file and then save it to a local directory as template.yaml.

AWSTemplateFormatVersion: '2010-09-09'
Transform: 'AWS::Serverless-2016-10-31'
Resources:  ReviewTranslate:
    Type: 'AWS::Serverless::Function'
    Properties:
        Handler: example.lambda_handler
        Runtime: python2.7
        CodeUri:
          Bucket: BUCKET_NAME
          Key: example.zip
        Policies:
         - AWSLambdaFullAccess
         - TranslateReadOnly
        Environment:
          Variables:
            TABLE_NAME: !Ref ReviewTable  
        Tracing: "Active"
  ReviewTable:
    Type: 'AWS::DynamoDB::Table'
    Properties:
        AttributeDefinitions: 
            - AttributeName: "review_id"
              AttributeType: "N"
            - AttributeName: "language"
              AttributeType: "S"
        KeySchema:
            - AttributeName: "review_id"
              KeyType: "HASH"
            - AttributeName: "language"
              KeyType: "RANGE"
        ProvisionedThroughput: 
            ReadCapacityUnits: 5
            WriteCapacityUnits: 5