help yourself

https://pipenv.pypa.io/en/latest/#install-pipenv-today

https://www.serverless.com/framework/docs/getting-started/

sls plugin install -n serverless-python-requirements

sls deploy

...however!

when deploying, first set up your credentials. for instance using env vars

AWS_ACCESS_KEY_ID=<your_acces_key_id> AWS_SECRET_ACCESS_KEY=<your_secret_access_key> sls deploy

or

export AWS_ACCESS_KEY_ID=<your_acces_key_id>
export AWS_SECRET_ACCESS_KEY=<your_secret_access_key>
sls deploy

you can create credentials in the aws console, if you haven't already

perform a deploy, familiarize yourself with the lambda in aws console. try calling the lambda using aws console (hint: create a test event) it should fail.

output to s3 by creating a bucket called workshop-aws-output or something similar. update handler.py#24 (_s3_upload(image_out, 'workshop-aws-output', key)) so it matches the bucket name you chose. deploy and verify that test calls produces objects in the bucket.

lambdas can be triggered by events in a bucket. in serverless.yml, uncomment the s3-config: functions.workshop-aws.events.s3. deploy and verify that serverless has created a bucket by name aws-workshop-input. since bucket names apparently must be unique :rolling_eyes:, rename it to whatever you like. put a png-file into the input-bucket, verify that the result is written to output bucket.

lambdas can be triggered by http events. in serverless.yml, uncomment the s3-config: functions.workshop-aws.events.http. nb: this will expose a get-endpoint publicly. figure out how you can adjust the lambda to support http events. hint: you'll want to use url-encoding