Lambda used to translate texts

Deploy

You need to install serverless in version 3.X.X

npm install -g serverless

You have then 2 available command line to deploy your lambda fonction

The first one is used to deploy the stack, on case of modification in the serverless.yml

npm run deploy

For production

npm run deploy -- --stage production

The second one is for updating the code of your lambda function

npm run update

For production

npm run update -- --stage production

Test with Localstack

General information on running Localstack can be found on this Notion page.

For testing the lambda function of this repo, you can follow the guide below:

General configuration

Install awslocal

pip install awscli-local

Create a localstack aws profile by adding the following in your ~/.aws/config:

[profile localstack]
region=eu-west-1
output=json
endpoint_url = http://localhost:4566

Add associated aws credentials by adding the following in your ~/.aws/credentials :

[localstack]
aws_access_key_id=test
aws_secret_access_key=test

To use this profile, we’ll add the following option to every command we pass to awslocal:

--profile localstack

Run Localstack in a Docker container

• Use docker-stack-dev to launch localstack container
• You can then access a localstack admin page locally at https://app.localstack.cloud/inst/default/status: From this page you can see your resources and edit some of them (edit secrets in Secrets Manager for example).

Create your AWS resources and invoke your lambda function

• Replace the fake tokens in localstack/payloadGenerator.js by valid tokens.
• Use localstack/payloadGenerator.js to generate a payload for the lambda function. The payload is stored in localstack/payloadOldToken.json and in payloadAccessToken.json.
• Change the OpenAI API key to a correct one in localstack/secretsManagerFixtures/dev-sportheroes-openai.json or on localstack admin page.
• Run sh localstack/runLambdaWithLocalstack.sh. This file will:
  • Create the lambda function in Locastack
  • Define the secrets
  • Invoke the lambda function with an old token and with an access token