Survey rules logic should allow the user to reference components of questions similar to piping

[rbyh]

Currently, methods for applying logical rules to a Survey are based on answers to questions, e.g.:

        s = Survey(questions=[q0, q1, q2])
        s = s.add_rule(q0, "q0 == 'Paris'", q2)

For consistency with piping, it would be nice to refer to an answer in the same way, e.g.:

        s = Survey(questions=[q0, q1, q2])
        s = s.add_rule(q0, "q0.answer == 'Paris'", q2)

It would also be nice to be able to reference other components of questions as you can with piping, e.g.:

        s = Survey(questions=[q0, q1, q2])
        s = s.add_rule(q0, "'Paris' in q0.comment", q2)