probcomp / iventure-aws-lambda

aws lambda integration for iventure

Summary

When using only the stock Python libraries, creating and deploying AWS Lambda functions is straightforward. However, when you need a third-party library, you'll need to package the library and any other dependencies into the deployment zip. Unfortunately, the documentation on how to do this is sparse and spread across AWS, StackOverflow, and blogs.

The default AWS AMI image is based on RHEL, so we use a CentOS docker image to build and package all dependencies. Then, we pull the relevant Python site packages (ignoring internal Python packages) and the AWS Lambda handler into a ZIP archive. We complete the process by deploying to AWS Lambda.

Usage

Usage: aws-lambda [OPTIONS] COMMAND [ARGS]...

Options:
  --role TEXT     Role ARN
  --profile TEXT  AWS credential profile
  --bucket TEXT   S3 bucket to upload deployment
  --help          Show this message and exit.

Commands:
  all      Build-package-deploy-run in one step
  build    Build image with dependencies
  deploy   Deploy to AWS Lambda
  package  Create AWS Lambda package
  run      Run deployed AWS Lambda function

Definitions for subcommand arguments

FUNCTION

Name of the AWS Lambda function to be invoked by the client.

PACKAGE

A standalone Python source file (aka the AWS Lambda handler). This is where the actual work happens.

METHOD

The entrypoint to the handler. When combined with `PACKAGE`, the combination of `PACKAGE.METHOD` is the name of the handler configured in AWS Lambda.

ROLE

The AWS resource name (ARN) for the role. It can also be defined using the `AWS_ROLE_ARN` environment variable.

BUCKET

The name of the S3 bucket used for deployment. It can also be defined using the `AWS_S3_BUCKET` environment variable.

Examples

AWS_ROLE_ARN=xxxxxxxxxx ./aws-lambda all HelloWorld hello.py execute