Feature: Async/await support for actions and workflows

[FernandoCelmer]

Idea

Add first-class support for async def actions and async workflow execution, enabling modern Python async patterns.

Motivation

Python's async ecosystem is mature and widely adopted. Many I/O-bound operations (HTTP calls, database queries, file operations) benefit from async execution. Dotflow should support async as a first-class citizen.

Proposed API

from dotflow import DotFlow, action

@action
async def fetch_data(context):
    async with aiohttp.ClientSession() as session:
        resp = await session.get("https://api.example.com/data")
        return await resp.json()

@action
async def process_data(context):
    return transform(context.previous_context.storage)

workflow = DotFlow()
workflow.task.add(fetch_data)
workflow.task.add(process_data)
await workflow.start_async(mode="sequential")

Implementation considerations

• Detect if action function is a coroutine (asyncio.iscoroutinefunction)
• Add start_async() method to DotFlow or make start() async-aware
• Execution class should handle both sync and async actions
• Async parallel mode using asyncio.gather() instead of multiprocessing
• Maintain backward compatibility — sync actions must continue to work unchanged
• Async storage providers (e.g., aioboto3 for S3)

Value

• Better performance for I/O-bound workflows
• Modern Python patterns
• Enables asyncio.gather for true concurrent I/O without multiprocessing overhead