Fix mecab parse problem

[swyo]

If input string has a strange word, mecab does not tokenize it.

This is because konlpy.tag._mecab.parse triggers Error as follows.

In [1]: from konlpy.tag import Mecab

In [2]: mc = Mecab()

In [3]: mc.morphs('_\u2028')

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-3-928619c63e3f> in <module>
----> 1 mc.morphs('_\u2028')

/usr/local/lib/python3.6/dist-packages/konlpy/tag/_mecab.py in morphs(self, phrase)
     84         """Parse phrase to morphemes."""
     85
---> 86         return [s for s, t in self.pos(phrase)]
     87
     88     def nouns(self, phrase):

/usr/local/lib/python3.6/dist-packages/konlpy/tag/_mecab.py in pos(self, phrase, flatten)
     76             if flatten:
     77                 result = self.tagger.parse(phrase)
---> 78                 return parse(result)
     79             else:
     80                 return [parse(self.tagger.parse(eojeol).decode('utf-8'))

/usr/local/lib/python3.6/dist-packages/konlpy/tag/_mecab.py in pos(self, phrase, flatten)
     76             if flatten:
     77                 result = self.tagger.parse(phrase)
---> 78                 return parse(result)
     79             else:
     80                 return [parse(self.tagger.parse(eojeol).decode('utf-8'))\

/usr/local/lib/python3.6/dist-packages/konlpy/tag/_mecab.py in parse(result, allattrs)
     30         return (s, t.split(',', 1)[0])
     31
---> 32     return [split(elem) for elem in result.splitlines()[:-1]]
     33
     34 class Mecab():

/usr/local/lib/python3.6/dist-packages/konlpy/tag/_mecab.py in <listcomp>(.0)
     30         return (s, t.split(',', 1)[0])
     31
---> 32     return [split(elem) for elem in result.splitlines()[:-1]]
     33
     34 class Mecab():

/usr/local/lib/python3.6/dist-packages/konlpy/tag/_mecab.py in split(elem)
     27     def split(elem):
     28         if not elem: return ('','SY')
---> 29         s, t = elem.split('\t')
     30         return (s, t.split(',', 1)[0])
     31

ValueError: not enough values to unpack (expected 2, got 1)

To fix this issue, I suggest this PR.

[minhoryang]

LGTM