2

The question of how to define a logger in one location was asked here: Using Python logging in multiple modules

However the two top answers use basic config or a file config. I don't want to use a file config because I want to be able to dynamically adjust the level with command line arguments. I don't want to use a basic config because, well, my config won't be basic.

This is not library code so I'm not concerned with whatever might happen if the module is imported. My code also has a single entry point.

This is (a simplified) version of what I have at the moment:

temp1.py :

import logging
import sys
import temp2

formatter = logging.Formatter(
        fmt = '%(asctime)s - %(module)s - %(levelname)s ===> %(message)s'
        )
logger = logging.getLogger('root')
logger.setLevel(logging.DEBUG)
handler = logging.StreamHandler()
handler.setFormatter(formatter)
logger.addHandler(handler)

logger.info("THIS IS AN INFO MESSAGE IN TEMP 1")

temp2.do_something()

temp2.py:

import logging

logger = logging.getLogger(__name__)

logger.info('In temp 2')

def do_something():
    logger.critical('doing someting in temp 2')

This is the output:

2018-11-23 18:31:40,543 - temp1 - INFO ===> THIS IS AN INFO MESSAGE IN TEMP 1
doing something in temp 2

The output I desire would: 1) Format the output in the file temp2.py as well and 2) Also log the module_level logging message in temp2.py

I've been through the docs but I can't figure this one out. What do I need to do?

Neil
  • 3,020
  • 4
  • 25
  • 48
  • Couldn't the file config be set to some default level that you then override with your command line parameters, i.e. read the file config _then_ read command line parameters, overriding as necessary? – Jonah Bishop Nov 23 '18 at 16:41
  • Maybe, but I then don't see how that override would propagate through to the other files. So I'd still be in the same position. – Neil Nov 23 '18 at 16:43
  • 1
    Do you have control over the other modules? One method I've used is passing the `logger` itself into the other modules as arguments. I concede it doesn't feel very pythonic but it does work. – r.ook Nov 23 '18 at 16:56
  • I could do that I suppose. – Neil Nov 23 '18 at 16:57
  • 1
    Another way is you can separate the logger itself entirely as a separate module, and then in each of your module `from my_logger import logger`. You might even `try: logger.debug(f'In {__name__}'); except NameError: from my_logger import logger;` to check if a logger already exists, if not, load in your custom logger – r.ook Nov 23 '18 at 17:00
  • Ohkay yes that does seem to work, and is somewhat less ugly than the first suggestion. I'm happy to do that. Thanks. – Neil Nov 23 '18 at 17:04

0 Answers0