我实现了一段代码，它从一个队列中获取一个元素，并将相同的对象从队列列表中放入每个队列中。问题是，当我运行特定测试时，我收到 ValueError: task_done() Called too much times 异常。该错误发生在测试代码中，而不是在被测试的代码中。
我正在使用asyncio.Queue并使用协程进行编程。
我将每个 Queue.get 与一个精确的 Queue.task_done 调用相匹配。
我正在使用 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:包含我的类实现的 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。并且它可以正确地完成工作。
测试代码从 10 个队列中的每一个中获取所有元素，并检查它们是否存在于原始队列中。对于第 9 个队列，一切顺利，没有错误，但是当测试尝试从第十个列表中获取第一个元素时，会引发 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'\x15\xad\t\xaf', b'\x8b\xa2M=', 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get都匹配一个task_done。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())
这样做我可以看到，即使引发了许多 ValueError ，元素仍然从队列中检索。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 =============================================================================================
为了确保测试消耗所有列表中的所有元素，我在测试结束时添加了错误断言，强制出现错误:
assert e in c_set
c_set.remove(e)
+ assert False == True
+
结果输出显示从所有列表中检索了所有元素，但最后一个队列上的每个 task_done 都会生成 ValueError。
Queue 7: element: "b'\x9b\xf8m\x02'" number 1 extracted of 3!
Queue 7: element: "b'\x15\xad\t\xaf'" number 2 extracted of 2!
Queue 7: element: "b'\x8b\xa2M='" number 3 extracted of 1!
Queue 7: element: "None" number 4 extracted of 0!
Queue 8: element: "b'\x9b\xf8m\x02'" number 1 extracted of 3!
Queue 8: element: "b'\x15\xad\t\xaf'" number 2 extracted of 2!
Queue 8: element: "b'\x8b\xa2M='" number 3 extracted of 1!
Queue 8: element: "None" number 4 extracted of 0!
Queue 9: element: "b'\x9b\xf8m\x02'" number 1 extracted of 3!
============================================================================================ 1 failed in 1.06 seconds ==
问题是，我是否遗漏了某些内容并且代码中有错误或者我发现了错误？
最佳答案
您的代码有错误。事实上，queue.task_done() 只应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。
但是您的中间件类正在它刚刚使用 .put() 的队列上调用它，这是 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!
当您删除该行时，您的测试就会通过。
您在测试中看到异常的原因是因为只有那时队列才知道task_done()被调用得太频繁。 DistributorMiddleware.distribute() 中的 queue.task_done() 调用会将未完成的任务计数器减 1，但仅当该计数器降至以下时零可以检测到异常。只有当最后一个任务在 test_distribution() 中从队列中取出时，您才会到达这一点，此时未完成的任务计数器至少提前了一步达到 0。
也许这应该是对 self._in.task_done() 的调用？您刚刚在 while 循环中从该队列中获取了一个元素:
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
关于python - Queue.asyncio值错误: task_done() called too many times - Coding error or a bug detected?，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/53218629/