Add Solution Quality Info to Concept Learners' Outputs

[Jean-KOUAGOU]

• Currently, h = model.fit(learning_problem=lp).best_hypotheses() returns an OWLClassExpression.
• It would be useful to add a new argument to the fit method, e.g., return_info or verbose. If the latter evaluates to True, then fit returns a tuple: h, info = model.fit(learning_problem=lp, .best_hypotheses()). The returned info can be a list or dictionary containing at least the quality of the computed solution(s).