pyorchestration is a celery based orchestration engine that allows to create orchestration flows through a minimal configuration.

It provides the following features:

• Automatic retries of orchestration
• Skippable steps
• Multiple orchestrations
• Injectable configurations

There are three main concepts:

• An orchestration is a sequence of functions being executed
• A step is a function being executed within an orchestration
• A State is an object being shared throughout the steps of the orchestration

First, define the orchestration state, this is an object that will be shared throughout the orchestration:

You need to override the from_dict method to parse the incoming request payload:

from pyorchestration.core.decorators import step
from pyorchestration.core.commons import OrchestrationState
from pyorchestration.start import celery_app

class MyState(OrchestrationState):
    def __init__(self, name):
        super().__init__()
        self.name = name

    @classmethod
    def from_dict(cls, payload):
        return cls(payload['name'])

Then, you simply define your orchestration using decorators:

@step(number=1, state_class=MyState)
def say_hi(state: MyState):
    logging.info(f'hi {state.name}')


@step(number=2, state_class=MyState)
def say_bye(state: MyState):
    logging.info(f'bye {state.name}')

That's it! You can launch the celery worker with celery command:

celery --app app.celery_app worker

to trigger an orchestration you can use standard celery approach, you need to specify two arguments:

• orchestration_name (default is: default_orchestration)
• payload object (can be blank if not needed)

    celery_app.send_task(
        "orchestration.tasks.orchestrate",
        args=[
            "default_orchestration",
            json.dumps({
                "name": "Andrea"
            })
        ]
    )

The worker will then pick up the new task and begin the orchestration.