使用 asyncio 创建两个并发异步任务

Question

我需要创建一个同时从 Web 套接字和管道接收的软件，并在另一个通道上发送消息（它从套接字接收，创建一个新线程并发送到管道。以同样的方式从管道，创建一个新线程并发送到套接字）。

我遇到了多线程问题，在程序启动时我必须启动方法socket_receiver 和pipe_receiver，但我只能启动pipe_receiver。我尝试删除所有代码并仅保留socket_receiver 和pipe_receiver，但它只输入pipe_receiver 的while True。

import asyncio
import sys
import json
from concurrent.futures.thread import ThreadPoolExecutor
import websockets

# make the Pool of workers
executor = ThreadPoolExecutor(max_workers=10)
# Make connection to socket and pipe
header = {"Authorization": r"Basic XXXX="}
connection = websockets.connect('wss://XXXXXXXX', extra_headers=header)


async def socket_receiver():
    """Listening from web socket"""
    async with connection as web_socket:
        while True:
            message = await web_socket.recv()
            # send the message to the pipe in a new thread
            executor.submit(send_to_pipe(message))


async def pipe_receiver():
    """Listening from pipe"""
    while True:
        message = sys.stdin.readline()
        if not message:
            break
        executor.submit(send_to_socket(message))
        # jsonValue = json.dump(str(line), file);
        sys.stdout.flush()


def send_to_pipe(message):
    # Check if message is CAM or DENM
    json_message = json.loads(message)
    type = int(json_message["header"]["messageID"])
    # 1 is DENM message, 2 is CAM message
    if type == 1  or type == 2:
        # send the message to the pipe
        sys.stdout.print(json_message);


async def send_to_socket(message):
     async with connection as web_socket:
        json_message = json.dumps(message)
        await web_socket.send(json_message)


asyncio.get_event_loop().run_until_complete(
    asyncio.gather(socket_receiver(),pipe_receiver()))

这个程序被一个子进程调用，父进程通过连接到stdout和stdin的管道与其通信。

更新：我收到@Martijn Pieters 代码的异常

Traceback (most recent call last):
  File "X", line 121, in <module>
    main()
  File "X", line 119, in main
    loop.run_until_complete(asyncio.gather(socket_coro, pipe_coro))
  File "X\AppData\Local\Programs\Python\Python37-32\lib\asyncio\base_events.py", line 568, in run_until_complete
    return future.result()
  File "X", line 92, in connect_pipe
    reader, writer = await stdio()
  File "X", line 53, in stdio
    lambda: asyncio.StreamReaderProtocol(reader), sys.stdin)
  File "X/AppData\Local\Programs\Python\Python37-32\lib\asyncio\base_events.py", line 1421, in connect_read_pipe
    transport = self._make_read_pipe_transport(pipe, protocol, waiter)
  File "X/AppData\Local\Programs\Python\Python37-32\lib\asyncio\base_events.py", line 433, in _make_read_pipe_transport
    raise NotImplementedError
NotImplementedError

Answer 1

您没有正确使用ThreadPoolExecutor，并且您真的不想在这里使用它。相反，您需要设置消费者和生产者来处理您的套接字和管道，并使用队列在它们之间发送消息。

• 对于每种连接类型，创建一个创建连接的协程，然后将该单个连接传递给该连接的消费者和生产者任务（使用asyncio.create_task() 创建）。使用asyncio.wait() 和return_when=asyncio.FIRST_COMPLETED 一起运行这两个任务，因此当两个任务之一“提前”完成（例如失败）时，您可以取消任何仍在运行的任务。

• 使用queue 将消息从一个连接的消费者传递到另一个连接的生产者。

• sys.stdin 和sys.stdout 是阻塞 流，不要只是读写它们！有关尝试设置非阻塞 STDIO 流的要点，请参阅 https://gist.github.com/nathan-hoad/8966377，以及要求提供非阻塞流功能的 this asyncio issue。

• 不要使用全局套接字连接，当然不要使用两个单独的async with 语句。您的send_to_socket() 方法实际上会关闭 套接字，因为async with connection as web_socket: 上下文管理器在发送第一条消息时退出，这会导致socket_receiver 代码出现问题，该代码假定套接字无限期保持打开状态.

• 不要在这里使用线程！您的连接完全由 asyncio 管理，线程将主要影响这一点。

• asyncio.Executor() instances 只能与常规可调用对象一起使用，不能与协程一起使用。 Executor.submit() 声明它需要一个可调用的，通过 executor.submit(send_to_pipe(message)) 或 executor.submit(send_to_socket(message)) 传入协程将导致引发异常，因为协程不是可调用的。您可能没有看到异常消息，因为该异常是在另一个线程中引发的。

  这就是你的 socket_receiver() 协程失败的原因；它肯定开始，但尝试发送消息失败。当我针对本地模拟 websocket 服务器运行您的代码时，会打印一条警告：

  RuntimeWarning: coroutine 'send_to_socket' was never awaited
    executor.submit(send_to_socket(message))
  

  当没有等待协程时，该协程中的代码永远不会执行。将协程包装在一个打印出 stderr (try: callable(), except Exception: traceback.print_exc(file=sys.stderr))) 异常的协程中，您将得到：

  Traceback (most recent call last):
    File "soq52219672.py", line 15, in log_exception
      callable()
  TypeError: 'coroutine' object is not callable
  

Executors 应该只用于集成无法转换为使用协程的代码；执行程序管理该代码以与asyncio 任务并行运行而不受干扰。如果该代码想要与asyncio 任务交互，则应小心，请始终使用asyncio.run_coroutine_threadsafe() 或asyncio.call_soon_threadsafe() 跨边界调用。请参阅Concurrency and multithreading section。

这是我如何重写代码以使用消费者/生产者模式的示例，stdio() 基于 Nathan Hoad gist on the subject，加上 Windows 的后备support for treating stdio as pipes is limited：

import asyncio
import json
import os
import sys

import websockets

async def socket_consumer(socket, outgoing):
    # take messages from the web socket and push them into the queue
    async for message in socket:
        await outgoing.put(message)

async def socket_producer(socket, incoming):
    # take messages from the queue and send them to the socket
    while True:
        message = await incoming.get()
        jsonmessage = json.dumps(message)
        await socket.send(jsonmessage)

async def connect_socket(incoming, outgoing):
    header = {"Authorization": r"Basic XXXX="}
    uri = 'wss://XXXXXXXX'
    async with websockets.connect(uri, extra_headers=header) as websocket:
        # create tasks for the consumer and producer. The asyncio loop will
        # manage these independently
        consumer_task = asyncio.create_task(socket_consumer(websocket, outgoing))
        producer_task = asyncio.create_task(socket_producer(websocket, incoming))

        # start both tasks, but have the loop return to us when one of them
        # has ended. We can then cancel the remainder
        done, pending = await asyncio.wait(
            [consumer_task, producer_task],
            return_when=asyncio.FIRST_COMPLETED
        )
        for task in pending:
            task.cancel()
        # force a result check; if there was an exception it'll be re-raised
        for task in done:
            task.result()


# pipe support
async def stdio(loop=None):
    if loop is None:
        loop = asyncio.get_running_loop()

    if sys.platform == 'win32':
        # no support for asyncio stdio yet on Windows, see https://bugs.python.org/issue26832
        # use an executor to read from stdio and write to stdout
        class Win32StdinReader:
            def __init__(self):
                self.stdin = sys.stdin.buffer 
            async def readline():
                # a single call to sys.stdin.readline() is thread-safe
                return await loop.run_in_executor(None, self.stdin.readline)

        class Win32StdoutWriter:
            def __init__(self):
                self.buffer = []
                self.stdout = sys.stdout.buffer
            def write(self, data):
                self.buffer.append(data)
            async def drain(self):
                data, self.buffer = self.buffer, []
                # a single call to sys.stdout.writelines() is thread-safe
                return await loop.run_in_executor(None, sys.stdout.writelines, data)

        return Win32StdinReader(), Win32StdoutWriter()

    reader = asyncio.StreamReader()
    await loop.connect_read_pipe(
        lambda: asyncio.StreamReaderProtocol(reader),
        sys.stdin
    )

    writer_transport, writer_protocol = await loop.connect_write_pipe(
        asyncio.streams.FlowControlMixin,
        os.fdopen(sys.stdout.fileno(), 'wb')
    )
    writer = asyncio.streams.StreamWriter(writer_transport, writer_protocol, None, loop)

    return reader, writer

async def pipe_consumer(pipereader, outgoing):
    # take messages from the pipe and push them into the queue
    while True:
        message = await pipereader.readline()
        if not message:
            break
        await outgoing.put(message.decode('utf8'))

async def pipe_producer(pipewriter, incoming):
    # take messages from the queue and send them to the pipe
    while True:
        jsonmessage = await incoming.get()
        message = json.loads(jsonmessage)
        type = int(message.get('header', {}).get('messageID', -1))
        # 1 is DENM message, 2 is CAM message
        if type in {1, 2}:
            pipewriter.write(jsonmessage.encode('utf8') + b'\n')
            await pipewriter.drain()

async def connect_pipe(incoming, outgoing):
    reader, writer = await stdio()
    # create tasks for the consumer and producer. The asyncio loop will
    # manage these independently
    consumer_task = asyncio.create_task(pipe_consumer(reader, outgoing))
    producer_task = asyncio.create_task(pipe_producer(writer, incoming))

    # start both tasks, but have the loop return to us when one of them
    # has ended. We can then cancel the remainder
    done, pending = await asyncio.wait(
        [consumer_task, producer_task],
        return_when=asyncio.FIRST_COMPLETED
    )
    for task in pending:
        task.cancel()
    # force a result check; if there was an exception it'll be re-raised
    for task in done:
        task.result()

async def main():
    pipe_to_socket = asyncio.Queue()
    socket_to_pipe = asyncio.Queue()

    socket_coro = connect_socket(pipe_to_socket, socket_to_pipe)
    pipe_coro = connect_pipe(socket_to_pipe, pipe_to_socket)

    await asyncio.gather(socket_coro, pipe_coro)

if __name__ == '__main__':
    asyncio.run(main())

然后从两个任务开始，一个管理套接字，另一个管理 STDIO 管道。两者都为他们的消费者和生产者启动了另外 2 个任务。有两个队列将消息从一个的消费者发送到另一个的生产者。