python 3.x - Why coroutines cannot be used with run_in_executor?

Question

I want to run a service that requests urls using coroutines and multithread. However I cannot pass coroutines to the workers in the executor. See the code below for a minimal example of this issue:

import time
import asyncio
import concurrent.futures

EXECUTOR = concurrent.futures.ThreadPoolExecutor(max_workers=5)

async def async_request(loop):
    await asyncio.sleep(3)

def sync_request(_):
    time.sleep(3)

async def main(loop):
    futures = [loop.run_in_executor(EXECUTOR, async_request,loop) 
               for x in range(10)]

    await asyncio.wait(futures)

loop = asyncio.get_event_loop()
loop.run_until_complete(main(loop))

Resulting in the following error:

Traceback (most recent call last):
  File "co_test.py", line 17, in <module>
    loop.run_until_complete(main(loop))
  File "/usr/lib/python3.5/asyncio/base_events.py", line 387, in run_until_complete
    return future.result()
  File "/usr/lib/python3.5/asyncio/futures.py", line 274, in result
    raise self._exception
  File "/usr/lib/python3.5/asyncio/tasks.py", line 239, in _step
    result = coro.send(None)
  File "co_test.py", line 10, in main
    futures = [loop.run_in_executor(EXECUTOR, req,loop) for x in range(10)]
  File "co_test.py", line 10, in <listcomp>
    futures = [loop.run_in_executor(EXECUTOR, req,loop) for x in range(10)]
  File "/usr/lib/python3.5/asyncio/base_events.py", line 541, in run_in_executor
    raise TypeError("coroutines cannot be used with run_in_executor()")
TypeError: coroutines cannot be used with run_in_executor()

I know that I could use sync_request funcion instead of async_request, in this case I would have coroutines by means of sending the blocking function to another thread.

I also know I could call async_request ten times in the event loop. Something like in the code below:

loop = asyncio.get_event_loop()
futures = [async_request(loop) for i in range(10)]
loop.run_until_complete(asyncio.wait(futures))

But in this case I would be using a single thread.

How could I use both scenarios, the coroutines working within multithreads? As you can see by the code, I am passing (and not using) the pool to the async_request in the hopes I can code something that tells the worker to make a future, send it to the pool and asynchronously (freeing the worker) waits for the result.

The reason I want to do that is to make the application scalable. Is it an unnecessary step? Should I simply have a thread per url and that is it? Something like:

LEN = len(list_of_urls)
EXECUTOR = concurrent.futures.ThreadPoolExecutor(max_workers=LEN)

is good enough?

Answer 1

You have to create and set a new event loop in the thread context in order to run coroutines:

import asyncio
from concurrent.futures import ThreadPoolExecutor


def run(corofn, *args):
    loop = asyncio.new_event_loop()
    try:
        coro = corofn(*args)
        asyncio.set_event_loop(loop)
        return loop.run_until_complete(coro)
    finally:
        loop.close()


async def main():
    loop = asyncio.get_event_loop()
    executor = ThreadPoolExecutor(max_workers=5)
    futures = [
        loop.run_in_executor(executor, run, asyncio.sleep, 1, x)
        for x in range(10)]
    print(await asyncio.gather(*futures))
    # Prints: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9]


if __name__ == '__main__':
    loop = asyncio.get_event_loop()
    loop.run_until_complete(main())