Making the Most of asyncio: Adding Tasks to Event Loops

The asyncio module in Python provides infrastructure for writing asynchronous code using the async/await syntax. At the heart of asyncio is the event loop which runs the asynchronous tasks.

To leverage concurrency and parallelism in Python, you need to understand how to properly add tasks to asyncio event loops. Here are some key points:

Queue Up Tasks with loop.create_task()

The event loop provides a .create_task() method that schedules a coroutine to run. For example:

import asyncio

async def my_coro():
    print('Hello world')

loop = asyncio.get_event_loop()
task = loop.create_task(my_coro())
loop.run_until_complete(task)

This queues up my_coro() to execute on the loop. Important: The task does not run until loop.run_until_complete() is called!

Submit Tasks with ensure_future()

You can also submit tasks using asyncio.ensure_future(). Under the hood, this still calls loop.create_task(), but it's more concise:

import asyncio

async def my_coro():
   print('Hello again!')
   
loop = asyncio.get_event_loop()
asyncio.ensure_future(my_coro())  
loop.run_forever()

Watch Out for Blocking Calls

Care must be taken to avoid blocking calls which pause execution waiting on long-running I/O. These will stall the event loop. Use the provided asyncio I/O methods like .read() instead.

In summary, asyncio event loops manage task execution. Enqueue jobs with loop.create_task() or ensure_future(). Take advantage of asyncio's asynchronous programming to speed up I/O-bound Python code.