Best way to use a single logger in my python project, overwrite root logger or replace import everywhere?

Title is a bit vague but I think my code examples make it clear. I want to start using a custom logger in a large project of mine.

Currently, in every file that I use logging, I import logging and use that like this:

# file A
import logging

logging.info(...)

I do this in many files. I want to use my own specific logger:

# __init__.py

logger = logging.getLogger('my_logger!')

There are two approaches I think to using this logger everywhere in my project:

Option 1:
Importing the logger defined above in all of my files instead of the logging module:

# file A
from ... import logger

logger.info(...)

Option 2: Overriding the root logger, so that I do not have to replace anything

# __init_
logger = logging.getLogger('my_logger!')
logging.root = logger

Now I do not have to replace the imports in all my files. What is the best option and why?

getLogger('foo') always returns the same Logger instance, no matter where it is called. (Loggers have their own global "namespace" managed by the module itself.)

Wherever you want to use my_logger, just use

import logging

logger = logging.getLogger('my_logger')

It's the job of the application to ensure that the logger named my_logger is configured correctly on startup.