Serverless Python Requirements

A Serverless v1.x plugin to automatically bundle dependencies from requirements.txt and make them available in your PYTHONPATH.

Requires Serverless >= v1.12

Install

npm install --save serverless-python-requirements

Add the plugin to your serverless.yml:

plugins:
  - serverless-python-requirements

Cross compiling!

Compiling non-pure-Python modules or fetching their manylinux wheels is supported on non-linux OSs via the use of Docker and the docker-lambda image. To enable docker usage, add the following to your serverless.yml:

custom:
  pythonRequirements:
    dockerizePip: true

To utilize your own Docker container instead of the default, add the following to your serverless.yml:

custom:
  pythonRequirements:
    dockerImage: <image name>:tag

This must be the full image name and tag to use, including the runtime specific tag if applicable.

Dealing with Lambda's size limitations

To help deal with potentially large dependencies (for example: numpy, scipy and scikit-learn) there is support for compressing the libraries. This does require a minor change to your code to decompress them. To enable this add the following to your serverless.yml:

custom:
  pythonRequirements:
    zip: true

and add this to your handler module before any code that imports your deps:

import unzip_requirements

If you want to be able to use sls invoke local and don't have a check for lambda or a try/except ImportError around that import, you can use the following option to make this plugin not delete the unzip_requirements helper:

custom:
  pythonRequirements:
    cleanupZipHelper: false

Omitting Packages

You can omit a package from deployment with the noDeploy option. Note that dependencies of omitted packages must explicitly be omitted too. For example, this will not install the AWS SDKs that are already installed on Lambda:

custom:
  pythonRequirements:
    noDeploy:
      - boto3
      - botocore
      - docutils
      - jmespath
      - python-dateutil
      - s3transfer
      - six

extra pip arguments

You can specify extra arguments to be passed to pip like this:

custom:
  pythonRequirements:
      dockerizePip: true
      pipCmdExtraArgs:
          - --cache-dir
          - .requirements-cache

Customize requirements file name

Some pip workflows involve using requirements files not named requirements.txt. To support these, this plugin has the following option:

custom:
  pythonRequirements:
    fileName: requirements-prod.txt

Customize Python executable

Sometimes your Python executable isn't available on your $PATH as python2.7 or python3.6 (for example, windows or using pyenv). To support this, this plugin has the following option:

custom:
  pythonRequirements:
    pythonBin: /opt/python3.6/bin/python

Manual invocations

The .requirements and requirements.zip(if using zip support) files are left behind to speed things up on subsequent deploys. To clean them up, run sls requirements clean. You can also create them (and unzip_requirements if using zip support) manually with sls requirements install.

Invalidate requirements caches on package

If you are using your own Python library, you have to cleanup .requirements on any update. You can use the following option to cleanup .requirements everytime you package.

custom:
  pythonRequirements:
    invalidateCaches: true

Updating to python 3.6

This requires an update to your serverless.yml:

provider:
  name: aws
  runtime: python3.6

And be sure to clean up .requirements or requirements.zip if they exist as python2.7 and python3.6 code can't coexist.

Credit

This plugin is influenced by serverless-wsgi from @logandk. I however wanted a simpler pip install process. It now also supports bundling packages without the wsgi handler.