Decorator: Specify supported/required fields for e.g. parameters by a function

[euronion]

Description

We are going to have a number of functions that work on self.data and require specific columns or row entries of the data dataframe.

As long as we only have functions that operate on Technologies this is not a problem.
I'm thinking that we could have a decorator to check whether the required columns (or row entries) exist and warn/fail if they don't.

This would solve to challenges that I anticipate:

• If we have other types of objects that expose self.data with transformations, we test to make sure that the data passed to the function is compatible with the function
• If someone wants to use our methods on their own data or starts modifying self.data, then our functions may fail if the required columns are not present
• It helps us to document and make it clearer, which function requires which columns/fields.

Example

E.g.

• the decorator could be called @requires_columns
• we wrap adjust_scale(..) as

@requires_columns(on_dataframe=self.data, columns=["value","unit","scale","scale_unit"])
def adjust_scale(...)

• The decorator then checks before adjust_scale(...) is called whether the columns exist in the dataframe on_dataframe
• Fail if they don't exist
• Pass to the function and execute normally if they exist

Notes, ideas, potential issues

• I mentioned row entries because I think this could also be implemented with a similar decorator for rows. E.g. every pair of [source, technology, year, region] would need on entry with a certain [parameter] for the method to work properly. Technologies / row groups that do not fulfill this requirement would be excluded from the function automatically. But let's keep it simple for now and not implement this.
• This could already be useful for e.g. the currency conversion functions and the archiving functions?

Happy to hear your thoughts @finozzifa

[euronion]

@finozzifa fyi