refactor: set log level to separate out train args

[anhuong]

Description of the change

Separate out python logging and setting log level with setting log level using train_args and altering train_args.log_level to match python logging level.

Related issue number

How to verify the PR

Was the PR tested

• I have added >=1 unit test(s) for every new method I have added.
• I have ensured all unit tests pass

[Abhishek-TAMU]

Nice to have a generic function set_python_log_level to set log level of any logger passed to it from any module of the code. Thank you!

Regarding function set_log_level, according to me, it is called to set root log level in the code based on either CLI using train_argsor ENV variable. Hence, I believe, if we don't want to call set_python_log_level twice in the function set_log_level, we can modify the code in function set_log_level as below:

set_python_log_level(log_level)
train_logger = logging.getLogger()
if logger_name:
    train_logger = logging.getLogger(logger_name)

return train_args, train_logger

Call set_python_log_level(log_level), will set the root log level and train_logger with or without logger_name, will inherit the root log level.