Adding penalization to identical contiguous voicing

[napulen]

Dealing with #9.

Slight penalization on repeated voicings (cost = 2).

Tested with this example (a slight variation of the original example, 3 repeated harmonies):

parser.set_defaults(
        key="B-",
        chord_progression="I I IV IV ii V V I",
        durations="1 1/2 1 1/2 1 1/2 1/2 1",
        time_signature="6/8",
)

Without penalization:

With penalization:

Notice that the third repeated pair, V and V, was voiced identically (not always changing the voicing, which may be desired), but the first two changed.

Second version sounds much better in my opinion.