从事件流中运行异步协程

Question

I'm having issues where my call to api.say() fails. I understand its a coroutine and needs to be yielded from but is that valid if my print_all is a standard function.

What is the correct way to structure this example?

Please assume that the Client isn't changeable, only the code in my example.

from .client import Client
import asyncio

api = Client()
login = ('', '')

def print_all(b=None, m=None):
    print("Buffer!", b)
    print("Message", m)

    if b and m:
        if b.name == 'bat':
            print("-sending to", b)
            api.say(b, "Hey yo."):    # <----


def main():
    api.login(*login)
    api.register_message_callback(print_all)
    api.register_state_callback(print_all)

    loop = asyncio.get_event_loop()
    loop.run_until_complete(api.run())
    loop.close()

if __name__ == '__main__':
    main()

Answer 1

The Client API might handle callbacks that are coroutines too ie, you can convert print_all() into a coroutine (prepend async before def , add await before api.say() ).

Otherwise, you could call asyncio.ensure_future(api.say(..)) to schedule the coroutine. It assumes that loop.run_until_complete(api.run()) won't return before api.say() is done ie, there is loop.run_forever() equivalent in your program or you wait for all tasks -- either collected explicitly or implicitly ( asyncio.Task.all_tasks() ).