如何在YAML中指定总是使用dictConfig在项目文件夹中创建日志文件？

Question

In my Python program I have the following code:

def main():
    # The file's path
    path = os.path.dirname(os.path.realpath(__file__))
    ...
    # Config file relative to this file
    loggingConf = open('{0}/configs/logging.yml'.format(path), 'r')
    logging.config.dictConfig(yaml.load(loggingConf))
    loggingConf.close()
    logger = logging.getLogger(LOGGER)
    ...

and this is my logging.yml configuration file:

version: 1
formatters:
  default:
    format: '%(asctime)s %(levelname)s %(name)s %(message)s'
handlers:
  console:
    class: logging.StreamHandler
    level: DEBUG
    formatter: default
    stream: ext://sys.stdout
  file:
    class : logging.FileHandler
    formatter: default
    filename: bot.log
loggers:
  cloaked_chatter:
    level: DEBUG
    handlers: [console, file]
    propagate: no

The problem is that the bot.log file is created where the program is launched. I want it to always be created in the project's folder, ie in the same folder as my Python program.

For an example, launching the program with ./bot.py would create the log file in the same folder. But launching it with python3 path/bot.py would create the log file a level above the Python program in the file hierarchy.

How should I write the filename in the config file to solve this? Or do I need to write a custom handler? If so, how? Or is this not possible to solve using dictConfig?

Answer 1

There are a number of ways you can achieve what you want. For example, one way is to make a custom initialiser for your handler:

import os
import yaml

def logmaker():
    path = os.path.dirname(os.path.realpath(__file__))
    path = os.path.join(path, 'bot.log')
    return logging.FileHandler(path)

def main():
    # The file's path
    path = os.path.dirname(os.path.realpath(__file__))

    # Config file relative to this file
    loggingConf = open('{0}/logging.yml'.format(path), 'r')
    logging.config.dictConfig(yaml.load(loggingConf))
    loggingConf.close()
    logger = logging.getLogger('cloaked_chatter')
    logger.debug('Hello, world!')

if __name__ == '__main__':
    main()

Note that I moved the logging.yml to be adjacent to the script. The logmaker is the custom initialiser. Specify it in the YAML as follows:

version: 1
formatters:
  default:
    format: '%(asctime)s %(levelname)s %(name)s %(message)s'
handlers:
  console:
    class: logging.StreamHandler
    level: DEBUG
    formatter: default
    stream: ext://sys.stdout
  file:
    () : __main__.logmaker
    formatter: default
loggers:
  cloaked_chatter:
    level: DEBUG
    handlers: [console, file]
    propagate: no

If you run the Python script, you should find that the bot.log is created adjacent to the script and YAML file. The same message is printed to the console and bot.log :

2013-04-16 11:08:11,178 DEBUG cloaked_chatter Hello, world!

NB The script could be a little tidier, but it illustrates my point.

Update: As per the documentation , the use of () as a key in dictionary indicates that the value is a callable which is essentially a custom constructor for the handler.