Currently the only methods that I wrap in during the metaclass class generation are the magic methods. I wonder if this should be expanded to all methods.
If that's the case, then would the __getattribute__ and _wrap ever get called?
Currently _wrap creates a new Method for each instance since it is doing it on the fly.
The only reason I could see in not pre-wrapping all pandas methods is that the UserFrame.dict would be huge. though it's already pretty big.
Actualy, I should add an intermediary class before UserFrame/UserSeries that only contains the wrapped functions. Then UserFrame.dict would only contain the UserFrame specific methods...
Currently the only methods that I wrap in during the metaclass class generation are the magic methods. I wonder if this should be expanded to all methods.
If that's the case, then would the
__getattribute__and_wrapever get called?Currently
_wrapcreates a new Method for each instance since it is doing it on the fly.The only reason I could see in not pre-wrapping all pandas methods is that the UserFrame.dict would be huge. though it's already pretty big.
Actualy, I should add an intermediary class before UserFrame/UserSeries that only contains the wrapped functions. Then UserFrame.dict would only contain the UserFrame specific methods...