Use `.apply` to change categories' dtype in `handle_float_categories_for_catboost`

From https://github.com/pandas-dev/pandas/issues/51074 using `apply(str)` can be used to set the float categories to be strings, and we can try to see if that lets us use the actual float categories. But we should also be able to use that to transform float integers to integers, and it seems like that should be faster.