我实现了一段代码，该代码从一个队列中获取一个元素，并将相同的对象放入队列列表中的每个队列中。问题是，当我运行一个特定的测试时，我会得到一个ValueError: task_done() called too many times异常。此错误发生在测试代码中，而不是正在测试的代码中。
我正在使用asyncio.Queue，并使用协程进行编程。我将每个Queue.get与一个完全相同的Queue.task_done调用进行匹配。我正在用pytest测试代码。
我正在使用以下库：
- Python 3.7
- pytest==3.10.0
- pytest asyncio==0.9.0
我有两个文件：包含类实现的middleware.py和实现pytest测试的test_middleware.py。
文件middlware.py:
import asyncio
class DistributorMiddleware:
def __init__(self, in_queue, out_list_queue):
self._in = in_queue
self._out = out_list_queue
async def distribute(self):
while True:
ele = await self._in.get()
count=0
for queue in self._out:
await queue.put(ele)
count+=1
print(f'inserted ele in {count}')
queue.task_done()
if ele == None:
break
for queue in self._out:
await queue.join()
文件test_middleware.py:
import pytest
import asyncio
from asyncio import Queue
from middleware import DistributorMiddleware
import random
import os
@pytest.mark.asyncio
async def test_distribution(request, event_loop):
q_list = [ Queue() for _ in range(10) ]
_in = Queue()
distrib = DistributorMiddleware(_in, q_list)
event_loop.create_task(distrib.distribute())
num_ele = random.randint(1, 10)
ele_set = set()
for _ in range(num_ele):
ele = os.urandom(4)
ele_set.add(ele)
await _in.put(ele)
await _in.put(None)
await asyncio.sleep(1)
for i,q in enumerate(q_list):
assert q.qsize() == num_ele + 1
c_set = ele_set.copy()
count= 0
while True:
e = await q.get()
count+=1
print(f'Queue {i}: element: "{e}" number {count} extracted of {q.qsize()}!')
q.task_done()
if e == None:
break
assert e in c_set
c_set.remove(e)
在测试中，中间件应该从输入队列中获取元素，并将它们放入列表中的10个队列中。它正确地完成了工作。
测试代码从10个队列中的每个队列中获取所有元素，并检查它们是否存在于原始队列中。对于9个第一个队列，一切都很顺利，没有错误，但当测试试图从第十个列表中获取第一个元素时，会引发一个ValueError：
request = <FixtureRequest for <Function 'test_distribution'>>, event_loop = <_UnixSelectorEventLoop running=False closed=False debug=False>
@pytest.mark.asyncio
async def test_distribution(request, event_loop):
q_list = [ Queue() for _ in range(10) ]
_in = Queue()
distrib = DistributorMiddleware(_in, q_list)
event_loop.create_task(distrib.distribute())
num_ele = random.randint(1, 10)
ele_set = set()
for _ in range(num_ele):
ele = os.urandom(4)
ele_set.add(ele)
await _in.put(ele)
await _in.put(None)
await asyncio.sleep(1)
for i,q in enumerate(q_list):
assert q.qsize() == num_ele + 1
c_set = ele_set.copy()
count= 0
while True:
e = await q.get()
count+=1
print(f'Queue {i}: element: "{e}" number {count} extracted of {q.qsize()}!')
> q.task_done()
test_middlewares.py:34:
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
self = <Queue at 0x7f7af5b9d828 maxsize=0 _queue=[b'x15xadtxaf', b'x8bxa2M=', None]>
def task_done(self):
"""Indicate that a formerly enqueued task is complete.
Used by queue consumers. For each get() used to fetch a task,
a subsequent call to task_done() tells the queue that the processing
on the task is complete.
If a join() is currently blocking, it will resume when all items have
been processed (meaning that a task_done() call was received for every
item that had been put() into the queue).
Raises ValueError if called more times than there were items placed in
the queue.
"""
if self._unfinished_tasks <= 0:
> raise ValueError('task_done() called too many times')
E ValueError: task_done() called too many times
/usr/lib/python3.7/asyncio/queues.py:202: ValueError
每个CCD_ 10与一个CCD_。我可以验证对test_middlware.py文件进行以下修改：
- q.task_done()
+ try:
+ q.task_done()
+ except ValueError as err:
+ print(f'Value Error: {err}')
+ print(q.qsize())
通过这样做，我可以看到，即使引发了许多ValueError，也会不断从队列中检索元素。测试成功：
platform linux -- Python 3.7.1, pytest-3.10.0, py-1.7.0, pluggy-0.8.0
rootdir: /tmp/stack, inifile:
plugins: asyncio-0.9.0
collected 1 item
test_middlewares.py . [100%]
============================================================================================ 1 passed in 1.04 seconds =============================================================================================
为了确保测试消耗了所有列表中的所有元素，我在测试结束时添加了一个错误断言：
assert e in c_set
c_set.remove(e)
+ assert False == True
+
结果输出显示，所有元素都是从所有列表中检索的，但最后一个队列上的每个task_done都会生成一个ValueError。
Queue 7: element: "b'x9bxf8mx02'" number 1 extracted of 3!
Queue 7: element: "b'x15xadtxaf'" number 2 extracted of 2!
Queue 7: element: "b'x8bxa2M='" number 3 extracted of 1!
Queue 7: element: "None" number 4 extracted of 0!
Queue 8: element: "b'x9bxf8mx02'" number 1 extracted of 3!
Queue 8: element: "b'x15xadtxaf'" number 2 extracted of 2!
Queue 8: element: "b'x8bxa2M='" number 3 extracted of 1!
Queue 8: element: "None" number 4 extracted of 0!
Queue 9: element: "b'x9bxf8mx02'" number 1 extracted of 3!
============================================================================================ 1 failed in 1.06 seconds ==
问题是，我是遗漏了什么，代码中有错误，还是发现了错误？
您的代码中有一个错误。实际上，queue.task_done()应该只在从队列中取出元素时调用，而不是在将它们放入队列时调用。
但是您的中间件类在刚刚使用.put()的队列上调用它，作为self._out列表中的最后一个队列；从DistributorMiddleware.distribute():中删除queue.task_done()调用
async def distribute(self):
while True:
ele = await self._in.get()
count=0
for queue in self._out:
await queue.put(ele)
count+=1
print(f'inserted ele in {count}')
queue.task_done()
# ^^^^^ you didn't take anything from the queue here!
当你去掉那条线时，你的测试就通过了。
您在测试中看到异常的原因是，只有这样队列才知道task_done()被调用得太频繁了。DistributorMiddleware.distribute()中的queue.task_done()调用将未完成任务计数器递减1，但只有当该计数器降至零以下时，才能检测到异常。只有当最后一个任务从test_distribution()中的队列中取出时，您才能达到这一点，此时未完成的任务计数器提前了至少一步到达0。
也许这是对self._in.task_done()的一个调用？您刚刚在while循环中从队列中获得了一个元素：
async def distribute(self):
while True:
ele = await self._in.get()
# getting an element from self._in
count=0
for queue in self._out:
await queue.put(ele)
count+=1
print(f'inserted ele in {count}')
self._in.task_done()
# done with ele, so decrement the self._in unfinished tasks counter