Python异步IO任务列表进度日志记录

Question

我需要处理大量的数据行，只有异步处理才有意义

我需要查看列表处理状态，即Done processing 1/3，但当我增加计数器时，它始终保持在1。这是有意义的，因为我将计数器发送到函数中。我需要这样做，因为如果没有它，我会：

UnboundLocalError: local variable 'processed' referenced before assignment

使用Python 3.8

任何帮助都将不胜感激

这里有一个测试链接：https://ideone.com/gRjrf2

我将代码抽象如下：

    import os, logging
    import asyncio
     
    logging.basicConfig(level=logging.INFO, format='%(asctime)s %(levelname)-8s [%(filename)s:%(lineno)d] %(message)s', datefmt='%d-%b-%y %H:%M:%S')
    logger = logging.getLogger(__name__)
     
    items = [{"name": "A"}, {"name": "B"}, {"name": "C"}]
     
    processed = 0
     
    async def increment(item):
        count = item.get('count', 0)
        count += 1
        return count
     
    async def get_and_update(item, processed):
        item['count'] = await increment(item)
        # Show progress now, but how?
        processed += 1
        logger.info(f"You can't see me {processed}")
     
    async def run():
        logger.info(f"Processing {len(items)} items...")
        await asyncio.gather(*[
            asyncio.create_task(
                get_and_update(item, processed)
            ) for item in items
        ])
     
    loop = asyncio.get_event_loop()
    loop.run_until_complete(run())

我得到的结果是：

28-Aug-20 11:19:22 INFO     [prog.py:23] Processing 3 items...
28-Aug-20 11:19:22 INFO     [prog.py:20] You can't see me 1
28-Aug-20 11:19:22 INFO     [prog.py:20] You can't see me 1
28-Aug-20 11:19:22 INFO     [prog.py:20] You can't see me 1

Answer 1

您的基本问题是，通过将processed声明为get_and_update的参数，您正在隐藏全局processed变量。您需要删除该参数，然后在该函数中将processed声明为全局，如下所示：

import os, logging
import asyncio
 
logging.basicConfig(level=logging.INFO, format='%(asctime)s %(levelname)-8s [%(filename)s:%(lineno)d] %(message)s', datefmt='%d-%b-%y %H:%M:%S')
logger = logging.getLogger(__name__)
 
items = [{"name": "A"}, {"name": "B"}, {"name": "C"}]
 
processed = 0
 
async def increment(item):
    count = item.get('count', 0)
    count += 1
    return count
 
async def get_and_update(item):
    global processed

    item['count'] = await increment(item)
    # Show progress now, but how?
    processed += 1
    logger.info(f"You can't see me {processed}")
 
async def run():
    logger.info(f"Processing {len(items)} items...")
    await asyncio.gather(*[
        asyncio.create_task(
            get_and_update(item)
        ) for item in items
    ])
 
loop = asyncio.get_event_loop()
loop.run_until_complete(run())

上面的输出是：

28-Aug-20 08:15:00 INFO     [asynctest.py:25] Processing 3 items...
28-Aug-20 08:15:00 INFO     [asynctest.py:22] You can't see me 1
28-Aug-20 08:15:00 INFO     [asynctest.py:22] You can't see me 2
28-Aug-20 08:15:00 INFO     [asynctest.py:22] You can't see me 3