3

How can I loop over a list of objects async and call their function. For example:

class Cat:
    def talk():
        print("Meow")

class Dog:
    def talk():
        print("Woof")


cat = Cat()
dog = Dog()

animal_list = [cat, dog]

# How would I do these async?
for animal in animal_list:
    animal.talk()

This thread, How to use an async for loop to iterate over a list?, recommends using asyncio, but doesn't example how I can have an object call it's own function such as animal.talk()

superdee73
  • 353
  • 2
  • 13
  • Are you ok with making `talk` function an async function? It's easier that way. – Hurried-Helpful Mar 03 '20 at 00:38
  • Yes, how would that look? – superdee73 Mar 03 '20 at 04:06
  • Can you describe in more detail what you would like to achieve? Specifically, what's wrong with the current code, and how you'd like the new code to behave. Note that the linked question doesn't just "recommend" the use of asyncio based on the use case, it is about `async for` to begin with. Unless you understand what `async for` does (hint: it doesn't automatically parallelize your loop, one could almost say that it does the opposite - see [here](https://stackoverflow.com/a/52856646/1600898) or [here](https://stackoverflow.com/a/56162461/1600898)), it doesn't make sense to use it. – user4815162342 Mar 03 '20 at 10:01

2 Answers2

1

Make the talk functions async otherwise there's no point using asyncio.

class Cat:
    async def talk():
        print("Meow")

class Dog:
    async def talk():
        print("Woof")


cat = Cat()
dog = Dog()

animal_list = [cat, dog]

Create a list (Iterable) of coroutines returned by animal.talk().

Either coroutines = map(lambda animal : animal.talk(), animal_list) or coroutines = [animal.talk() for animal in animal_list] will do.

Then finally scheduled the list of coroutines for execution.

# This returns the results of the async functions together.
results = await asyncio.gather(coroutines)

# This returns the results one by one.
for future in asyncio.as_completed(coroutines):
    result = await future

cat.talk() and dog.talk() will be executed asynchronously, which means the order of their execution is not guaranteed, and may be run on different threads. But here the talk function is so simple that it will look like it's run synchronously, and confers no real benefits.

But if talk involved making a network request or a long, heavy computation, and the animal_list were very long, doing it this way can help with the performance.

Hurried-Helpful
  • 1,850
  • 5
  • 15
0

I like @Hurried-Helpful's post on this. Making their example work might look as follows.

import asyncio
import random
import time

class Cat:
    @staticmethod
    async def talk():
        await asyncio.sleep(random.choice([1,2]))
        print("Meow")

class Dog:
    @staticmethod
    async def talk():
        await asyncio.sleep(random.choice([0,1]))
        print("Woof")

async def main():
    print(f"Started at: {time.strftime('%X')}")
    cat = Cat()
    dog = Dog()
    animal_list = [cat, dog]
    res = await asyncio.gather(
        *[animal.talk() for animal in animal_list]
    )
    print(f"Ended at: {time.strftime('%X')}")

asyncio.run(main())

The Python docs are much better on async functions now, and also contain good examples: https://docs.python.org/3/library/asyncio-task.html#running-tasks-concurrently

ross-spencer
  • 1