[MRG] label segmentation on whitespace

[shuttle1987]

Related is #214 that deals with Unicode space characters

[shuttle1987]

Interesting that the test fails here with this:

    def test_unicode_segmentation():
        """Test that unicode whitespace characters are correctly handled in segmentation"""
        from persephone.preprocess.labels import segment_into_chars
        no_break_space = "hello\u00A0world"
>       assert segment_into_chars(no_break_space) == "h e l l o w o r l d"
E       AssertionError: assert 'h e l l o \xa0 w o r l d' == 'h e l l o w o r l d'
E         - h e l l o   w o r l d
E         ?           --
E         + h e l l o w o r l d

[shuttle1987]

I've reduced the scope of the PR here since I think this chunk of work about unicode whitespaces is separate to the interface related considerations

[oadams]

Nice stuff, looks good!