Add a method that returns all the original features.
Kyubyong opened this issue · comments
Hi,
The original mecab-ko returns many features, but this repo and konlpy extract only the first one, POS tag.
I guess it's useful if you add a method that returns all the features like this:
def analyze(self, sentence):
lattice = _create_lattice(sentence)
if not self.tagger.parse(lattice):
raise MeCabError(self.tagger.what())
return [
(node.surface, node.feature)
for node in lattice
]
Hello!
I added parse(sentence)
method and it returns a Feature
namedtuple
.
Please refer to the example snippet below!
import mecab
mecab = mecab.MeCab()
mecab.parse('즐거운 하루 보내세요!')
# [
# ('즐거운', Feature(
# pos='VA+ETM', semantic=None, has_jongseong=True, reading='즐거운',
# type='Inflect', start_pos='VA', end_pos='ETM',
# expression='즐겁/VA/*+ᆫ/ETM/*')),
# ('하루', Feature(
# pos='NNG', semantic=None, has_jongseong=False, reading='하루',
# type=None, start_pos=None, end_pos=None,
# expression=None)),
# ('보내', Feature(
# pos='VV', semantic=None, has_jongseong=False, reading='보내',
# type=None, start_pos=None, end_pos=None,
# expression=None)),
# ('세요', Feature(
# pos='EP+EF', semantic=None, has_jongseong=False, reading='세요',
# type='Inflect', start_pos='EP', end_pos='EF',
# expression='시/EP/*+어요/EF/*')),
# ('!', Feature(
# pos='SF', semantic=None, has_jongseong=None, reading=None,
# type=None, start_pos=None, end_pos=None,
# expression=None))
# ]
Thanks. Awesome!