Is abstract class field in python possible?

How would you design an abstract time series validation class in Python?

• I'm trying to abstract all the work of doing time-series validation, so that I never have to think about it again.  Ideally, I would like to pass the Class a pandas dataframe, a time window, and a function which looks at data in some window and returns a single number (for one step ahead prediction).  It would loop through the data and return various things like predictions, error rates, and so on.  It's ok if it's blindingly slow.  Are there any good design tips to keep in mind when trying to implement this?

• Answer:

  There is one design pattern that may solve your problemas that is the strategy pattern. If you are using OOP you can create a class with you data and the validator would be just another variable. If the solution above is too complicated for the problem you can always create an object with the data you need to validate...