Lightweight library for AWS SWF.
Garcon deals with easy going clients and kitchens. It takes orders from clients (deciders), and send them to the kitchen (activities). Difficult clients and kitchens can be handled directly by the restaurant manager.
Requirements
- Python 3.5, 3.6, 3.7, 3.8 (tested)
- Boto3 (tested)
Goal
The goal of this library is to allow the creation of Amazon Simple Workflow without the need to worry about the orchestration of the different activities and building out the different workers. This framework aims to help simple workflows. If you have a more complex case, you might want to use directly boto.
Code sample
The code sample shows a workflow that has 4 activities. It starts with activity_1, which after being completed schedule activity_2 and activity_3 to be ran in parallel. The workflow ends after the completion of activity_4 which requires activity_2 and activity_3 to be completed.
import boto3
from garcon import activity
from garcon import runner
client = boto3.client('swf', region_name='us-east-1')
domain = 'dev'
name = 'workflow_sample'
create = activity.create(client, domain, name)
test_activity_1 = create(
name='activity_1',
run=runner.Sync(
lambda activity, context: print('activity_1')))
test_activity_2 = create(
name='activity_2',
requires=[test_activity_1],
run=runner.Async(
lambda activity, context: print('activity_2_task_1'),
lambda activity, context: print('activity_2_task_2')))
test_activity_3 = create(
name='activity_3',
requires=[test_activity_1],
run=runner.Sync(
lambda activity, context: print('activity_3')))
test_activity_4 = create(
name='activity_4',
requires=[test_activity_3, test_activity_2],
run=runner.Sync(
lambda activity, context: print('activity_4')))
Application architecture
. βββ cli.py # Instantiate the workers βββ flows # All your application flows. β βββ __init__.py β βββ example.py # Should contain a structure similar to the code sample. βββ tasks # All your tasks β βββ __init__.py β βββ s3.py # Task that focuses on s3 files. βββ task_example.py # Your different tasks.
Contributors
- Michael Ortali (Author)
- Adam Griffiths
- Raphael Antonmattei
- John Penner