lambdex turns pex files into aws lambda functions.
pex is a tool that simplifies packaging python environments and is ideally suited for aws lambda. lambdex takes pex files and turns them into aws lambda functions, allowing you to more easily run complex applications in the cloud.
aws lambda documentation and concepts can be found here.
The lambdex cli has two subcommands: build
and test
. build
further has two possible modes of operation: by specifying
an entry point that already exists within the pex file (-e
) or by specifying an external script and handler to embed within
the pex file (-s/-H
).
First you must package a pex file. Assuming you already have the pex
tool
and a requirements.txt, you can simply run
pex -r requirements.txt -o lambda_function.pex
to produce a pex file containing the requirements. If you must build a pex file with platform-specific extensions, see the tips section below for more information about building Amazon Linux-specific extensions.
This can be done one of two ways, depending on where your code lives.
If you have a handler function named 'handler' in package 'mymodule.myapp' that is already contained within lambda_function.pex, then you can simply run
lambdex build -e mymodule.myapp:handler lambda_function.pex
If you have a script function.py with a lambda handler named my_handler
, you would instead run
lambdex build -s function.py -H my_handler lambda_function.pex
This bundles function.py within the pex environment and instructs lambdex to
call the python function my_handler
when being invoked by AWS.
Once you have created a lambdex file, you can test it as if it were being invoked by Amazon using lambdex test
.
Given a lambdex package lambda_function.pex
, you can either send it an empty json event using
lambdex test --empty lambda_function.pex
You can alternately supply a list of files containing json structs e.g.
lambdex test lambda_function.pex event1.json event2.json ...
You can create/update lambda functions via the AWS Console, or you can do it
via the CLI using aws lambda create-function
or aws lambda update-function-code
respectively.
NOTE: When creating the function, you must specify the AWS Lambda handler as
lambdex_handler.handler
. Via the CLI, this is the --handler
flag. This
is the wrapper injected by lambdex that manages invocation of your code.
Do not confuse this with the -H
option to lambdex build
.
Most simple dependencies have no platform-specific extensions and thus can be built anywhere. However there are a number of popular packages (e.g. numpy, scipy, matplotlib, PIL, ...) that require building C extensions that can prove tricky to get packaged correctly.
Amazon provides an amazonlinux docker image which can be useful for building platform-specific extensions to run on AWS Lambda. See documentation for information about that image.
The minimum Dockerfile to produce can environment that can build Amazon Linux-specific pex files can be found here