python日志记录确保只添加一次处理程序

Question

我有一段代码正在初始化一个记录器,如下所示.

logger = logging.getLogger()
hdlr = logging.FileHandler('logfile.log')
formatter = logging.Formatter('%(asctime)s %(levelname)s %(message)s')
hdlr.setFormatter(formatter)
logger.addHandler(hdlr) 
logger.setLevel(logging.DEBUG)

不幸的是这个代码被多次调用,有什么方法可以检查处理程序是否已经存在 - 我更愿意实现它而不必使用Singleton.

编辑:对不起,忘了提到这是在python 2.5 - 干杯,理查德

Answer 1

作为@offbyone注释,可以将冗余处理程序添加到记录器的同一实例中.用于记录的python文档说 -

"使用相同名称多次调用getLogger()将返回对同一记录器对象的引用."

所以我们不必担心将实现变成单例,就像它已经存在一样.

遗憾的是,对于与记录器的同一实例关联的处理程序,情况并非如此.有可能是连接重复处理.

例-

1. 复制此代码并将其保存在main.py中

   import logging
   print 'inside main.py',
   print '-'*50
   def logger():
   
         print 'initializing logger....'
         logPath = '.'
         fileName = 'temp'
   
         # configure log formatter
         logFormatter = logging.Formatter("%(asctime)s [%(filename)s] [%(funcName)s] [%(levelname)s] [%(lineno)d] %(message)s")
   
         # configure file handler
         fileHandler = logging.FileHandler("{0}/{1}.log".format(logPath, fileName))
         fileHandler.setFormatter(logFormatter)
   
         # configure stream handler
         consoleHandler = logging.StreamHandler()
         consoleHandler.setFormatter(logFormatter)
   
         # get the logger instance
         logger = logging.getLogger(__name__)
   
         # set the logging level
         logger.setLevel(logging.DEBUG)
   
         print 'adding handlers- '
   
         #if not len(logger.handlers):
         logger.addHandler(fileHandler)
         logger.addHandler(consoleHandler)
   
         print 'logger initialized....\n'
         print 'associated handlers - ', len(logger.handlers)
         for handler in logger.handlers:
               print handler
         print
         return logger
   
   main_logger = logger()
   main_logger.info('utilizing main.py logger.')
   print 'exiting main.py',
   print '-'*50
   

2. 以及sub.py中的以下代码

   print 'inside sub.py',
   print '-'*50
   print 'importing main.py'
   import main
   print 'imported main.py'
   import logging
   print 'getting logger instance in sub'
   sub_logger = main.logger()
   print 'got logger instance in sub'
   sub_logger.info("utilizing sub_logger")
   print 'exiting sub.py',
   print '-'*50
   

3. 运行sub.py

   narayan@y510p:~/code/so$ python sub.py
   inside sub.py --------------------------------------------------
   importing main.py
   inside main.py --------------------------------------------------
   initializing logger....
   adding handlers- 
   logger initialized....
   
   associated handlers -  2
   <logging.FileHandler object at 0x7f7158740c90>
   <logging.StreamHandler object at 0x7f7158710b10>
   
   2015-08-04 07:41:01,824 [main.py] [<module>] [INFO] [41] utilizing main.py logger.
   exiting main.py --------------------------------------------------
   imported main.py
   getting logger instance in sub
   initializing logger....
   adding handlers- 
   logger initialized....
   
   associated handlers -  4 # <===== 4 handlers (duplicates added)
   <logging.FileHandler object at 0x7f7158740c90>
   <logging.StreamHandler object at 0x7f7158710b10>
   <logging.FileHandler object at 0x7f7158710bd0>
   <logging.StreamHandler object at 0x7f7158710c10>
   
   got logger instance in sub
   2015-08-04 07:41:01,824 [sub.py] [<module>] [INFO] [10] utilizing sub_logger
   2015-08-04 07:41:01,824 [sub.py] [<module>] [INFO] [10] utilizing sub_logger
   exiting sub.py --------------------------------------------------
   

因此,对返回同一记录器的方法的多次调用添加了重复的处理程序.

现在,对于你的问题 -

有什么方法可以检查处理程序是否已经存在

就在这里-

logger.handlers返回与给定关联的所有处理程序的列表logger.

在将处理程序添加到记录器实例之前,请确保不要添加重复的处理程序在main.py中,只需取消注释所说的行if not len(logger.handlers):并正确缩进以下两行 -

if not len(logger.handlers):
    logger.addHandler(fileHandler)
    logger.addHandler(consoleHandler)

现在再次运行sub.py

narayan@y510p:~/code/so$ python sub.py
inside sub.py --------------------------------------------------
importing main.py
inside main.py --------------------------------------------------
initializing logger....
adding handlers- 
logger initialized....

associated handlers -  2
<logging.FileHandler object at 0x7fd67a891c90>
<logging.StreamHandler object at 0x7fd67a862b10>

2015-08-04 08:14:45,620 [main.py] [<module>] [INFO] [41] utilizing main.py logger.
exiting main.py --------------------------------------------------
imported main.py
getting logger instance in sub
initializing logger....
adding handlers- 
logger initialized....

associated handlers -  2 # <===== Still 2 handlers (no duplicates)
<logging.FileHandler object at 0x7fd67a891c90>
<logging.StreamHandler object at 0x7fd67a862b10>

got logger instance in sub
2015-08-04 08:14:45,620 [sub.py] [<module>] [INFO] [10] utilizing sub_logger
exiting sub.py --------------------------------------------------

此外,如果要限制要添加到记录器实例的处理程序类型,可以执行以下操作 -

    print 'adding handlers- '
    # allows to add only one instance of file handler and stream handler
    if len(logger.handlers) > 0:
        print 'making sure we do not add duplicate handlers'
        for handler in logger.handlers:
              # add the handlers to the logger
              # makes sure no duplicate handlers are added

              if not isinstance(handler, logging.FileHandler) and not isinstance(handler, logging.StreamHandler):
                    logger.addHandler(fileHandler)
                    print 'added file handler'
                    logger.addHandler(consoleHandler)
                    print 'added stream handler'
    else:
        logger.addHandler(fileHandler)
        logger.addHandler(consoleHandler)
        print 'added handlers for the first time'

希望这可以帮助!

编辑:

遗憾的是,对于与记录器的同一实例关联的处理程序,情况并非如此.有可能是连接重复处理.

事实证明,上述陈述并非完全正确.

假设我们已经在主模块中创建并配置了一个名为"main_logger"的记录器(它只是配置记录器,不会返回任何内容).

# get the logger instance
logger = logging.getLogger("main_logger")
# configuration follows
...

现在在子模块中,如果我们在命名层次结构'main_logger.sub_module_logger'之后创建子记录器,我们不需要在子模块中配置它.只需在命名层次结构后创建记录器就足够了.

# get the logger instance
logger = logging.getLogger("main_logger.sub_module_logger")
# no configuration needed
# it inherits the configuration from the parent logger
...

并且它也不会添加重复的处理程序.

参考 - 使用多个模块中的日志记录

Answer 2

如果处理程序已经存在,logger.addHandler()将不会添加处理程序.要检查处理程序是否已经存在,您可以检查logger.handlers列表:

logger = logging.getLogger()
hdlr = logging.FileHandler('logfile.log')
formatter = logging.Formatter('%(asctime)s %(levelname)s %(message)s')
hdlr.setFormatter(formatter)
logger.addHandler(hdlr) 
logger.setLevel(logging.DEBUG)
print logger.handlers
# [<logging.FileHandler object at 0x14542d0>]
logger.addHandler(hdlr)
print logger.handlers
# [<logging.FileHandler object at 0x14542d0>]

除此之外,如果您有一个或在__init__.py您的包的文件中,我建议将此代码放在您的main()函数中,以便每次都不必调用它.我还建议您使用命名记录器,不要使用根记录器.像这样的东西:

logger = logging.getLogger(__name__)
...

希望这有用:)

Answer 3

您还可以检查处理程序列表是否为空。这是我最终得到的解决方案：

def setup_logging(self, logfile):
    self._logger = logging.getLogger('TestSuite')
    self._logger.setLevel(logging.INFO)
    host = socket.gethostname().split('.')[0]
    if self._logger.handlers == []:
        fh = logging.handlers.RotatingFileHandler(logfile,
                                                  maxBytes=10*1024*1024,
                                                  backupCount=5)
        strfmt = "%" + "(asctime)s [%s] " % host + "%" + "(message)s"
        fmt = logging.Formatter(strfmt, datefmt="%Y.%m%d %H:%M:%S")
        fh.setFormatter(fmt)

        self._logger.addHandler(fh)
    self._logger.info('-' * (55 - len(host)))

我看到处理程序添加了多次，因此每个日志消息都被多次写入日志文件，这解决了这个问题。