Difficulty in retriving AutoML settings if the settings only are provided via AutoML.fit

I am dealing with an AutoML use case written in a similar way as this [example](https://github.com/microsoft/FLAML/blob/main/test/automl/test_classification.py#L388).

During development, I want to retrieve the AutoML settings in a function of the AutoML instance other than the `fit` function, for example, retrieving the original "log_file_name" in `retrain_from_log`. Currently, there is no convenient way for one to retrieve them (unless the settings are provided when the AutoML instance is constructed). Is this correct?
If so, I suggest recording the user-specified settings in the `fit` function as well (as in the AutoMl [constructor](https://github.com/microsoft/FLAML/blob/main/flaml/automl/automl.py#L722) (line 722-765)).