Conversation
qlib/log.py
Outdated
| super().__init__(name, bases, dic) | ||
|
|
||
| def __new__(cls, name, bases, dict): | ||
| wrapper_dict = type(logging.getLogger("MetaLogger")).__dict__.copy() |
There was a problem hiding this comment.
Why not use Logger directly?
There was a problem hiding this comment.
type(logging.Logger) will give the result of <class 'type'>, while calling type(logger object) will give the result <class 'logging.Logger'>. The aim of wrapper dict here aimed to get the type dictionary of logger object which needs a logger instance here (since we need the information and messages in the dict for making help work).
There was a problem hiding this comment.
Will it be different by replacing type(logging.getLogger("MetaLogger")) with logging.Logger?
Why should we create a Logger instance here?
qlib/log.py
Outdated
| def __new__(cls, name, bases, dict): | ||
| wrapper_dict = type(logging.getLogger("MetaLogger")).__dict__.copy() | ||
| wrapper_dict.update(dict) | ||
| wrapper_dict["__doc__"] = logging.getLogger("MetaLogger").__doc__ |
There was a problem hiding this comment.
Why not use logging.Logger.__doc__ ?
There was a problem hiding this comment.
Yes, we can use logging.Logger.__doc__ here, which will have the same results as doing so with an instance.
There was a problem hiding this comment.
Why should we create a Logger instance here instead of using the class?
Using the class directly takes fewer steps and the code looks simpler
Update qlib logger
2 participants
Description
The first commit solves the problem of pickling logger object under certain python version. But, I'm still figuring out how to maintain the original docstring of the logger.
First trial: now use
helpfunction will print all the docstrings ofQlibLogger, but the docstring hints won't show up while using IDEs such as VSCode.Motivation and Context
How Has This Been Tested?
pytest qlib/tests/test_all_pipeline.pyunder upper directory ofqlib.Screenshots of Test Results (if appropriate):
Types of changes