README.rst

aws-sfn-builder

AWS Step Functions State Machine builder.

The primary motivation for this was the amount of boilerplate (Next, End) required to compose a valid state machine definition, but soon one got carried away.

Python 3.6+ only.

Installation

pip install aws-sfn-builder

Generate State Machine Definition from Simple Description

from aws_sfn_builder import Machine

Machine.parse([
    "a",
    [
        ["b-10", "b-11"],
        ["b-20"],
    ],
    "c",
]).to_json()

This will generate:

{
    "StartAt": "a",
    "States": {
        "a": {
            "Type": "Task",
            "Next": "d3d52323-137d-4228-9956-d3b77cc43a92"
        },
        "d3d52323-137d-4228-9956-d3b77cc43a92": {
            "Type": "Parallel",
            "Next": "c",
            "Branches": [
                {
                    "StartAt": "b-10",
                    "States": {
                        "b-10": {
                            "Type": "Task",
                            "Next": "b-11"
                        },
                        "b-11": {
                            "Type": "Task",
                            "End": true
                        }
                    }
                },
                {
                    "StartAt": "b-20",
                    "States": {
                        "b-20": {
                            "Type": "Task",
                            "End": true
                        }
                    }
                }
            ]
        },
        "c": {
            "Type": "Task",
            "End": true
        }
    }
}

Parse Existing State Machine Definition

# TODO load state_machine_json_dict

state_machine = Machine.parse(state_machine_json_dict)

Compile State Machine

# TODO initialise state_machine

state_machine.compile()

Test Your State Machine

Work in progress.

# TODO initialise state_machine

runner = Runner()

@runner.resource_provider("arn:aws:lambda:us-east-1:123456789012:function:Foo")
def foo(input):
    return "foo-result"

final_state, output = runner.run(state_machine)