Prepare an S3 bucket (srcBucket) where you will upload the images, and another different S3 bucket (dstBucket) to where new images will be resized and copied over.

Create/update an env.yml file in the root directory, copy and paste the content of env.sample.yml in it, and fill up the respective values.

Install docker. The purpose of docker is to install the correct image cropping binary for the lamdba node 8.10 environment that lambda uses instead of that of your local machine.

When running the deployment script, it will install the docker image prior to uploading to lambda.

Time to deploy!

To deploy to AWS using named profile:

npm run deploy

This project is meant to be plug and play. In the event you need something more than what is set up, you can develop and debug.

Run npm run installsharp to cleanly install the correct sharp for your local machine.

Before you start development, create a pair of src and dst buckets for development purpose. Upload an image in your src bucket. This is because during development, the copying of files from src bucket to dst will be executed on the AWS.

In the event.json file, replace <BUCKET_NAME> with your src bucket name and <PATH_TO_IMAGE> with the path to the image in the src bucket. The json inside event.json follows the exact JSON structure of a S3 put event.

Then run this command to debug:

npm run installsharp # ensures installs binary
sls invoke local -f onUpload -p event.json

The first command will download the sharp folder that fits your local machine. It can be skipped if you have run it at least once after your lattest deploy to production.

The second command will call the onUpload function using the content in event.json as the event param.

In the env.yml file, append widths that you want to crop the image according to to the widths variable, separated by comma. Follow the syntax as shown in env.sample.yml file.

Uses sharp to resize and produce produce extra image of the webp format. This will require lambda to have the correct binary. That is easily installed with the sharp package with npm and easily uploaded to lambda using docker.

• code in a way that can use other cloud providers