Translate Dynamic and Static Text using AWS Lambda.

Translate Dynamic and Static text into any language using AWS Lambda and DynamoDB

This uses Zappa and Python

For demo purposes only Arabic is supported, but more languages can be added. To do this, just add the language iso code to the validLangs list. The list of these codes can be found here: https://docs.aws.amazon.com/translate/latest/dg/what-is.html#what-is-languages

This Lambda function can translate text in 2 ways:

1. In bulk such as a List or an Array of strings (useful for static translations/static web pages)
2. Single line of text (useful when there is dynamic text such as form validation)

Ensure that you have a dynamodb table called Translations created beforehand.

Initially since no translation exist, the first call takes a little longer (few seconds) than every other subsequent call.

If the translation does not exist, then the text is translated using AWS translate_text method. The hash of the original string and the translated text are then stored in a dynamodb table called Translations.

In this table, the unique key is the SrcString which is the MD5 hash of the source string. You may need this value when translating an array of texts.

The TranslatedText column represents the translated text.

---

To run locally

1. Create a virtual environment called venv
2. source venv/bin/activate
3. pip install -r requirements.txt
4. export FLASK_APP=translation.py
5. export ApiKey='123' (Note: this is just used for testing. An actual Auth header value must be generated prior to a production release)
6. flask run

This should start a local wsgi development server. The application will be running on port 5000 by default.

You can now send requests to this endpoint.

To Translate a single line of text:

Request:

curl --location --request POST 'http://127.0.0.1:5000/translate' \
--header 'Authorization: Bearer 123' \
--header 'Content-Type: application/json' \
--data-raw '{
    "language": "ar",
    "txt": "hello world! please translate me!"
}'

Response:

{
    "data": "مرحبا بالعالم! الرجاء ترجمة لي!",
    "isSuccess": true
}

To translate a list/array of texts:

Request:

curl --location --request POST 'http://127.0.0.1:5000/translate' \
--header 'Authorization: Bearer 123' \
--header 'Content-Type: application/json' \
--data-raw '{
    "language": "ar",
    "TextList": ["hello world! please translate me!", "Hello World!"]
}'

Response:

{
    "data": [
        {
            "strHash": "a6b1e80a7a066436a369943c38f56f56",
            "translatedText": "عالم مرحبا!"
        },
        {
            "strHash": "03c53061de8fb588d9b407115a50402c",
            "translatedText": "مرحبا بالعالم! الرجاء ترجمة لي!"
        }
    ],
    "isSuccess": true
}

Notice how the response is different. The frontend application can utilize the strHash to map the original string with the translated string.

Deploy this application to lambda

Before proceeding ensure that you have appropriate AWS IAM permissions.

1. Update the following values in the zappa_settings.json file:

   aws_region

   profile_name

   s3_bucket

2. zappa deploy dev

If you have already deployed it and want to run an update, simply run

zappa update dev