AWS Lambdas on FastAPI - LoF

AWS Lambdas on FastAPI (LoF) is a command line tool that helps you fast & easy up & run your Python AWS Lambdas for tests and local development.

Pay attention, that this is only for Python lambdas.

It does not support any other programming languages.

How does it work?

Install

    pip install lof

Now run lof & provide to it path to your template yaml/json file.
Or you can run it from source dir with template.yaml (/.json) without any args

How to use

    lof
    # or if path custom
    lof --template example/template.yaml

You can choose that lambdas exclude from run by passing their names:

    lof --template example/template.yaml --exclude=PostLambda2Function

To pass environment variables to Lambdas, use flag —env, you can pass variables in 2 formats - json format and ‘.env’ format. Both files as examples presented in example/ folder

    lof --env=.env
    # or 
    lof --env=vars.json

Autorizer Lambda

To emulate behaviour of Authorizer lambda use flag —proxy-lambdas, where - LambdaAuthorizer must be changed to your lambda name from cloud formation template. Request will got through this proxy lambdas and only if everything ok it will call target lambda.

Same as in API GAteway - all return values from proxy lambdas will update “requestContext” key in the event.

    lof --proxy-lambdas=LambdaAuthorizer
    # or 
    lof --proxy-lambdas=LambdaAuthorizer,CORS # if you need to or more proxy lambdas

Other settings

    Usage: lof [OPTIONS]
    Options:
  --template TEXT                 Path to AWS Code Deploy template with
                                  lambdas  [default: template.yaml]
  --env TEXT                      Path to file with environment variables
  --exclude TEXT                  Exclude lambdas.FastAPI will not up & run
                                  them. Pass as string with comma. Example:
                                  PostTrafficHook,PretrafficHook.  [default: ]
  --port INTEGER                  Port to run lof  [default: 8000]
  --host TEXT                     Host to run lof  [default: 0.0.0.0]
  --proxy-lambdas TEXT            Lambdas Names that must be used as Handlers
                                  for request. For example, Authorizer Lambda
                                  or CORS Lambds. Each time when you send
                                  request to some lambda - it will go through
                                  those lambdas and populate 'requestContext'
                                  in the event  [default: ]
  --workers INTEGER               Count of unicorn workers to run.If you want
                                  run more when 1 worker LoF will generate
                                  temp FastAPI server code for your lambdas.
                                  [default: 1]
  --debug / --no-debug            Debug flag for Uvicorn  [default: True]
  --reload / --no-reload          Reload flag for Uvicorn  [default: False]
  --help                          Show this message and exit.

This mean, that lof will up & run all lambdas exclude this 2: PostTrafficHook & Roles

Demo

will be added soon

Example

To try how LoF works you can use AWS CloudFormation template.yaml & Lambdas from example/ folder.

Issues & features request

Fill free to open Issues & report bugs. I will solve them as soon as possible. If you have any sugesstions or feature request - also fell free to open the issue.

Problem Context

On my current project I works a lot with AWS Lambdas & tries to up & run them with SAM local.
And there is some issues especially when you work on the project with a big count of lambdas.

Some of them:

1) First of all it does not allow skip some lambdas form config
2) It build lambdas inside each docker container so it takes significant time to build/rebuild & up all containers (and you need up all containers if you want to have fast integration tests)

Both points in the mix make impossible to use SAM in weak developers envs like VDI, for example.

Changelog

v0.5.5

1. One more fix autorizer response population in event context

v0.5.4

1. Fix autorizer response population in event context

v0.5.3

1. Fixed bunch of issues relative to missed pathParams in event & wrong authorizer context providing

v0.5.2

1. Bug fixies.

v0.5.0

1. Added AWS Context object. Part of values filed right now with mock values, values from config will be added in next versions.

v0.4.1
Features:

1. Added option —proxy-lambdas where you can pass a list of lambdas that will be used as middleware request handlers.
   For example, as Authorization Lambda or CORS lambda.

Based on order in that you provided lambdas names in —proxy-lambdas option request will be send through them and populates same as on aws in “requestContext” field of the event.
For example, if you use —proxy-lambdas=CORS,Authorizer this mean request first of all will go to CORS lambda and if all ok (no raise errors) will got to Authorizer lambda and when to target lambda (endpoint that you call).

Fixes:

1. Paths with symbols ‘-.’ now does not cause issue during running with 1 and more worker.

v0.3.0

1. Added Possimility to run multiple workers with flag —workers.
   This helpful if you need speed up your local server or some lambdas need to call another lambdas directly.
2. Added flag —reload to cli if you want auto reload server when code changed (uvicor —reload)
3. Added support for Cloud Formation templates in JSON

v0.2.3

1. Possibility to send port & host to start several instances in same time.

v0.2.2

1. README.md is updated
2. Fixed Issue with lambdas in template, that does not have Events with Path (like S3 triggered lambdas)
3. Fixed issue with status code 204 - now it returns correct answer with no failes.
4. Added some tests

v0.2.1

1. Now LoF do not try/except lambdas errors

v0.2.0

1. Fixed status_code resend from lambda & JSON body response

v0.1.0

1. First version of Lambdas on FastApi.
   Based on AWS CloudFormation template it’s serve lambdas as FastAPI endpoints for local testing.