Python日志(log)记录配置问题

因此，这里有几个问题.首先，__file__在logger.py中的确切位置进行求值.第二个logging.basicConfig被呼叫了两次.第二个呼吁被简单地忽略了.您可以通过从time导入sleep并在调用之前将sleep(3)添加到data.py中的configure_logger来测试这一点.秒数将滚动，从而产生不同的文件名，但只创建第一个文件.最后，您传递的是script_name的完全不同的值，但是，如上所述，第二个调用被忽略了.

未修改的文件名

如果您可以容忍它是小写的，并且扩展名为.py，那么您可以使用可用于日志(log)格式的内置文件名参数.这允许您完全删除logger.py文件以及对getLogger()的调用

# data.py

import logging
from datetime import datetime as dt

# Only put this here if logging needs to happen in imported modules OUTSIDE of functions
log_basename = "./datalog"
log_format = f'%(asctime)s - %(levelname)s - %(filename)s - %(message)s'
logging.basicConfig(filename=f"{log_basename}_{dt.now().strftime('%Y%m%d-%H_%M_%S')}.log", level=logging.INFO, format=log_format)

from login import *
if __name__ == "__main__":
    logging.info(f"This log is from data.py")

# login.py

import logging
logging.info(f"This log is from login.py")

这将提供以下输出:

2023-09-28 04:33:46,382 - INFO - login.py - This log is from login.py
2023-09-28 04:33:46,382 - INFO - data.py - This log is from data.py

正如 comments 所暗示的那样，将日志(log)配置放在导入之前是违反惯例的.如果所有日志(log)记录都发生在函数或主块中，则应该在导入之后设置日志(log)记录.或者，也可以将其移回其自己的文件中，然后再次导入.

import logging

from datetime import datetime as dt
from login import *

def setup_logging():
    log_basename = "./datalog"
    log_format = f'%(asctime)s - %(levelname)s - %(filename)s - %(message)s'
    logging.basicConfig(filename=f"{log_basename}_{dt.now().strftime('%Y%m%d-%H_%M_%S')}.log", level=logging.INFO, format=log_format)

if __name__ == "__main__":
    setup_logging()
    login_foo()
    logging.info(f"This log is from data.py")

# login.py

import logging

def login_foo():
    logging.info(f"This log is from login.py")

带有大小写更改的文件名

如果更改大小写并删除.py扩展名很重要，您可以使用context filter将修改后的文件名变量添加到记录器的字典中.

# data.py

import logging
from datetime import datetime as dt
from login import *

class ContextFilter(logging.Filter):
    def filter(self, record):
        record.FILENAME = record.filename[:-3].upper()
        return True


def setup_logging():
    log_basename = "./datalog"
    log_format = f'%(asctime)s - %(levelname)s - %(FILENAME)s - %(message)s'
    logging.basicConfig(filename=f"{log_basename}_{dt.now().strftime('%Y%m%d-%H_%M_%S')}.log", level=logging.INFO, format=log_format)
    f = ContextFilter()
    logging.getLogger().addFilter(f)


if __name__ == "__main__":
    setup_logging()
    login_foo()
    logging.info(f"This log is from data.py")

这提供了准确的所需输出.

2023-09-28 05:24:56,256 - INFO - LOGIN - This log is from login.py
2023-09-28 05:24:56,256 - INFO - DATA - This log is from data.py