Typed Python interface to AWS Simple Workflow Service.
- Type annotations
- Explicit exceptions
- Execution state construction
- Consistent method/attribute/parameter names (see below)
- Consistent model struture
- Automatic flattening of paged-list responses
- next-page calls are run concurrently and on-demand
- Better execution filtering
The most interesting part is state construction: given an execution history, this
library can build a full state of the execution and all of its tasks with all details;
see swf_typed.build_state
. The rest of the API simply wraps and closely matches AWS's SWF API.
- py-swf - typed and object-oriented interface layer
- mypy-boto3-swf - type-annotated layer
- python-simple-workflow - higher-level interface layer
pip install swf-typed
import swf_typed
execution = swf_typed.ExecutionId(id="spam", run_id="abcd1234")
execution_details = swf_typed.describe_execution(execution, domain="eggs")
print(execution_details.configuration)
events = swf_typed.get_execution_history(execution, domain="eggs")
state = swf_typed.build_state(events)
for task in state.tasks:
print(task.status)
This library has a slight change in terminology from AWS SDKs/APIs/docs:
- Workflow type -> workflow
- Workflow execution -> execution
- Workflow execution
workflowId
-> execution ID - Activity type -> activity
- Activity task -> task
- Activity worker -> worker
- Activity task
activityId
-> task ID
This is to simplify symbol names.