improve __str__ and __repr__

[eiennohito]

Based on WorksApplications/SudachiPy#166
Fixes #122

Closes #SudachiPy/166

There is a slight difference in the proposed format caused by WordId formatting, the implemented version uses (dic_id, word_id)

>>> d = sudachipy.Dictionary()
>>> tok = d.create(sudachipy.SplitMode.A)
>>> mrs = tok.tokenize("外国人参政権")
>>> mrs
<MorphemeList[
  <Morpheme(外国, 0:2, (0, 375175))>,
  <Morpheme(人, 2:3, (0, 284079))>,
  <Morpheme(参政, 3:5, (0, 331513))>,
  <Morpheme(権, 5:6, (0, 522170))>,
]>
>>> str(mrs)
'外国 人 参政 権'
>>> mrs[0]
<Morpheme(外国, 0:2, (0, 375175))>
>>> str(mrs[0])
'外国'

Remaining question:
Should strings be naked as they are now or should we put them into quotes? (<Morpheme('外国', 0:2, (0, 375175))>)