Currently, the user doesn't know which debiasing methodologies are going to be run on their model, and cannot edit which methodologies to run.
Instead of the framework running a set of unspecified debiasing methodologies, the "run_framework" method could accept a list of methodologies to run against the models.
You could create a class for a debiasing methodology that takes a name to specify the implementation and a dictionary of input data. That allows the methodologies to be run in a loop over the list provided
Currently, the user doesn't know which debiasing methodologies are going to be run on their model, and cannot edit which methodologies to run. Instead of the framework running a set of unspecified debiasing methodologies, the "run_framework" method could accept a list of methodologies to run against the models.