precision,recall,f1_score for non binary classification

[SamanHushi]

Is your feature request related to a problem?

When I need the results from a non binary classifier you need to iterate over every unique positive class to evaluate it

Desired solution

example on recall()
model.recall() should also accept None for positive_class and then return results for every unique value in the target column,
That way it can also handel non binary classification

it could be handelt in a table with head ["value", "recall"] and show in the columns the results for each positive class

Possible alternatives (optional)

No response

Screenshots (optional)

No response

Additional Context (optional)

No response