tag Package

Note

Initial runs of each class method may require some time to load dictionaries (< 1 min). Second runs should be faster.

Hannanum Class

class konlpy.tag._hannanum.Hannanum(jvmpath=None, max_heap_size=1024)

Wrapper for JHannanum.

JHannanum is a morphological analyzer and POS tagger written in Java, and developed by the Semantic Web Research Center (SWRC) at KAIST since 1999.

>>> from konlpy.tag import Hannanum
>>> hannanum = Hannanum()
>>> print(hannanum.analyze(u'롯데마트의 흑마늘 양념 치킨이 논란이 되고 있다.'))
[[[('롯데마트', 'ncn'), ('의', 'jcm')], [('롯데마트의', 'ncn')], [('롯데마트', 'nqq'), ('의', 'jcm')], [('롯데마트의', 'nqq')]], [[('흑마늘', 'ncn')], [('흑마늘', 'nqq')]], [[('양념', 'ncn')]], [[('치킨', 'ncn'), ('이', 'jcc')], [('치킨', 'ncn'), ('이', 'jcs')], [('치킨', 'ncn'), ('이', 'ncn')]], [[('논란', 'ncpa'), ('이', 'jcc')], [('논란', 'ncpa'), ('이', 'jcs')], [('논란', 'ncpa'), ('이', 'ncn')]], [[('되', 'nbu'), ('고', 'jcj')], [('되', 'nbu'), ('이', 'jp'), ('고', 'ecc')], [('되', 'nbu'), ('이', 'jp'), ('고', 'ecs')], [('되', 'nbu'), ('이', 'jp'), ('고', 'ecx')], [('되', 'paa'), ('고', 'ecc')], [('되', 'paa'), ('고', 'ecs')], [('되', 'paa'), ('고', 'ecx')], [('되', 'pvg'), ('고', 'ecc')], [('되', 'pvg'), ('고', 'ecs')], [('되', 'pvg'), ('고', 'ecx')], [('되', 'px'), ('고', 'ecc')], [('되', 'px'), ('고', 'ecs')], [('되', 'px'), ('고', 'ecx')]], [[('있', 'paa'), ('다', 'ef')], [('있', 'px'), ('다', 'ef')]], [[('.', 'sf')], [('.', 'sy')]]]
>>> print(hannanum.morphs(u'롯데마트의 흑마늘 양념 치킨이 논란이 되고 있다.'))
['롯데마트', '의', '흑마늘', '양념', '치킨', '이', '논란', '이', '되', '고', '있', '다', '.']
>>> print(hannanum.nouns(u'다람쥐 헌 쳇바퀴에 타고파'))
['다람쥐', '쳇바퀴', '타고파']
>>> print(hannanum.pos(u'웃으면 더 행복합니다!'))
[('웃', 'P'), ('으면', 'E'), ('더', 'M'), ('행복', 'N'), ('하', 'X'), ('ㅂ니다', 'E'), ('!', 'S')]
Parameters:
  • jvmpath – The path of the JVM passed to init_jvm().
  • max_heap_size – Maximum memory usage limitation (Megabyte) init_jvm().
analyze(phrase)

Phrase analyzer.

This analyzer returns various morphological candidates for each token. It consists of two parts: 1) Dictionary search (chart), 2) Unclassified term segmentation.

morphs(phrase)

Parse phrase to morphemes.

nouns(phrase)

Noun extractor.

pos(phrase, ntags=9, flatten=True, join=False)

POS tagger.

This tagger is HMM based, and calculates the probability of tags.

Parameters:
  • ntags – The number of tags. It can be either 9 or 22.
  • flatten – If False, preserves eojeols.
  • join – If True, returns joined sets of morph and tag.

Kkma Class

class konlpy.tag._kkma.Kkma(jvmpath=None, max_heap_size=1024)

Wrapper for Kkma.

Kkma is a morphological analyzer and natural language processing system written in Java, developed by the Intelligent Data Systems (IDS) Laboratory at SNU.

>>> from konlpy.tag import Kkma
>>> kkma = Kkma()
>>> print(kkma.morphs(u'공부를 하면할수록 모르는게 많다는 것을 알게 됩니다.'))
['공부', '를', '하', '면', '하', 'ㄹ수록', '모르', '는', '것', '이', '많', '다는', '것', '을', '알', '게', '되', 'ㅂ니다', '.']
>>> print(kkma.nouns(u'대학에서 DB, 통계학, 이산수학 등을 배웠지만...'))
['대학', '통계학', '이산', '이산수학', '수학', '등']
>>> print(kkma.pos(u'다 까먹어버렸네요?ㅋㅋ'))
[('다', 'MAG'), ('까먹', 'VV'), ('어', 'ECD'), ('버리', 'VXV'), ('었', 'EPT'), ('네요', 'EFN'), ('?', 'SF'), ('ㅋㅋ', 'EMO')]
>>> print(kkma.sentences(u'그래도 계속 공부합니다. 재밌으니까!'))
['그래도 계속 공부합니다.', '재밌으니까!']

Warning

There are reports that Kkma() is weak for long strings with no spaces between words. See issue #73 for details.

Parameters:
  • jvmpath – The path of the JVM passed to init_jvm().
  • max_heap_size – Maximum memory usage limitation (Megabyte) init_jvm().
morphs(phrase)

Parse phrase to morphemes.

nouns(phrase)

Noun extractor.

pos(phrase, flatten=True, join=False)

POS tagger.

Parameters:
  • flatten – If False, preserves eojeols.
  • join – If True, returns joined sets of morph and tag.
sentences(phrase)

Sentence detection.

Komoran Class

class konlpy.tag._komoran.Komoran(jvmpath=None, userdic=None, modelpath=None, max_heap_size=1024)

Wrapper for KOMORAN.

KOMORAN is a relatively new open source Korean morphological analyzer written in Java, developed by Shineware, since 2013.

>>> cat /tmp/dic.txt  # Place a file in a location of your choice
코모란     NNP
오픈소스    NNG
바람과 함께 사라지다     NNP
>>> from konlpy.tag import Komoran
>>> komoran = Komoran(userdic='/tmp/dic.txt')
>>> print(komoran.morphs(u'우왕 코모란도 오픈소스가 되었어요'))
['우왕', '코모란', '도', '오픈소스', '가', '되', '었', '어요']
>>> print(komoran.nouns(u'오픈소스에 관심 많은 멋진 개발자님들!'))
['오픈소스', '관심', '개발자']
>>> print(komoran.pos(u'혹시 바람과 함께 사라지다 봤어?'))
[('혹시', 'MAG'), ('바람과 함께 사라지다', 'NNP'), ('보', 'VV'), ('았', 'EP'), ('어', 'EF'), ('?', 'SF')]
Parameters:
  • jvmpath – The path of the JVM passed to init_jvm().
  • userdic –

    The path to the user dictionary.

    This enables the user to enter custom tokens or phrases, that are mandatorily assigned to tagged as a particular POS. Each line of the dictionary file should consist of a token or phrase, followed by a POS tag, which are delimited with a <tab> character.

    An example of the file format is as follows:

    바람과 함께 사라지다 NNG
바람과 함께      NNP
자연어 NNG

    If a particular POS is not assigned for a token or phrase, it will be tagged as NNP.

  • modelpath – The path to the Komoran HMM model.
  • max_heap_size – Maximum memory usage limitation (Megabyte) init_jvm().
morphs(phrase)

Parse phrase to morphemes.

nouns(phrase)

Noun extractor.

pos(phrase, flatten=True, join=False)

POS tagger.

Parameters:
  • flatten – If False, preserves eojeols.
  • join – If True, returns joined sets of morph and tag.

Mecab Class

Warning

Mecab() is not supported on Windows 7.

class konlpy.tag._mecab.Mecab(dicpath='/usr/local/lib/mecab/dic/mecab-ko-dic')

Wrapper for MeCab-ko morphological analyzer.

MeCab, originally a Japanese morphological analyzer and POS tagger developed by the Graduate School of Informatics in Kyoto University, was modified to MeCab-ko by the Eunjeon Project to adapt to the Korean language.

In order to use MeCab-ko within KoNLPy, follow the directions in optional-installations.

>>> # MeCab installation needed
>>> from konlpy.tag import Mecab
>>> mecab = Mecab()
>>> print(mecab.morphs(u'영등포구청역에 있는 맛집 좀 알려주세요.'))
['영등포구', '청역', '에', '있', '는', '맛집', '좀', '알려', '주', '세요', '.']
>>> print(mecab.nouns(u'우리나라에는 무릎 치료를 잘하는 정형외과가 없는가!'))
['우리', '나라', '무릎', '치료', '정형외과']
>>> print(mecab.pos(u'자연주의 쇼핑몰은 어떤 곳인가?'))
[('자연', 'NNG'), ('주', 'NNG'), ('의', 'JKG'), ('쇼핑몰', 'NNG'), ('은', 'JX'), ('어떤', 'MM'), ('곳', 'NNG'), ('인가', 'VCP+EF'), ('?', 'SF')]
Parameters:dicpath – The path of the MeCab-ko dictionary.
morphs(phrase)

Parse phrase to morphemes.

nouns(phrase)

Noun extractor.

pos(phrase, flatten=True, join=False)

POS tagger.

Parameters:
  • flatten – If False, preserves eojeols.
  • join – If True, returns joined sets of morph and tag.

Twitter Class

See also

Korean POS tags comparison chart

Compare POS tags between several Korean analytic projects. (In Korean)