Rework and expand quorum calculations

[polettif]

With this PR quorum calculations in biproportional() and pukelsheim() are reworked. Instead of the parameters quorum_districts and quorum_total there's a new parameter quorum which is best used with the new quorum_all()/quorum_any() functions. This closes #3.

Instead of

biproporz(votes_matrix, seats, quorum_districts = 0.1, quorum_total = 0.05)

You'd write:

biproporz(votes_matrix, seats, quorum = quorum_all(any_district = 0.1, total = 0.05))

That way the previously implicit assumption that both quorums must be reached to be elegible for seats must be explicitly stated with quorum_all. If it's enough to reach at least one quorum (e.g. either by district or total), quorum_any should be used. If you don't want any quorum calculations to be done, the parameter can be ignored. In general, you can no longer set a quorum threshold of 0.

With this approach it should also be easier to add new quorum calculations without modifying biproportional too much. It might be a bit convoluted (a function creating a list of functions as a function argument) but quorums can be applied in a number of combinations which are difficult to cover just by parameters alone.

Other changes:

• biprop_quorum() was removed
• The separated functions reached_quorum_any_district() and reached_quorum_total() were added
• reached_quorums() allows a "lower-level" access to quorum calculations
• the quorum parameter is now in third position in pukelsheim() (instead of after new_seats_col), like biproportional()