General use case questions for Kedro Datasets. For some datasets the method that eventually uses load_args and/or save_args sometimes expects python objects as arguments. Of course, the yaml file can't support this inherently. I was curious if there were any plans to try and support this natively? And if not if there was a best practice way to handle it. Currently I am making thin wrappers of any dataset this is an issue in and making my own mapping of strings to python objects which works fine. But just curious if there were other recommendations. Thanks!

Awesome thanks, this should also work!