Change logs¶
Version 0.6.0¶
Released on Jan 2, 2022.
Version 0.5.2¶
Released on Dec 3, 2019.
Added data acquisition features from Koshort [#211 by nyanye]
Updated Mecab to mecab-0.996-ko-0.9.2 and mecab-ko-dic-2.1.1-20180720 [#214 by rickiepark]
Added Python 2 warning at the install.rst [#277 by minhoryang]
Added methods to make Mecab picklable [#234 by rickiepark] [#258 by rickiepark]
Added tests for coverage [#261 by minhoryang] [#262 by minhoryang]
Fixed Komoran not to POS tag empty sentences [#201 by lovit]
Fixed JPype usage by adding numpy as dependency [#246 by e9t]
Fixed to use tweepy 3.7.0+ to avoid collision with Python 3.7+ [#243 by shurain]
Fixed to use JPype 0.7.0+ to remove warning message [#245 by e9t]
Fixed to use lxml 4.1.0+ to avoid installation errors [#242 by shurain]
Fixed by removing .decode in Mecab.pos for Python3 usage [#108 by ty91]
Updated Mecab installation script [#158 by HaebinShin], [#247 by e9t], [#255 by HaebinShin], [#277 by minhoryang]
Version 0.5.1¶
Released on Aug 3, 2018.
Version 0.5.0¶
Released on Aug 1, 2018.
Added join parameter to POS taggers [#135 by pinetree408]
Fixed JPype-related installation error [#94 by shaynekang]
Moved description.py to konlpy/about.py [#194]
Updated Komoran from 2.4 to 3.0 and add userdic [#198 by lovit]
Warning
Previous dicpath in Komoran’s API is now modelpath. The name was changed to prevent confusion with the newly added userdic.
Version 0.4.4¶
Released on Oct 25, 2015.
Included tagsets for each morpheme analyzer [26a39d7]
Version 0.4.3¶
Released on Feb 27, 2015.
Fixed conditional requirement bug for pip3 [#50]
Version 0.4.2¶
Released on Feb 25, 2015.
Updated Korean documents (i.e., include missing .mo files)
Version 0.4.1¶
Released on Feb 25, 2015.
Version 0.4.0¶
Released on Jan 18, 2015.
Added
data
,downloader
[#4]Added
utils.csvread()
,utils.csvwrite()
Added
utils.read_txt()
Added nonflattened results for POS taggers
Added Komoran module with KOMORAN 2.4
Change dependency version of JPype1 to 0.5.7+
Change dependency version of mecab-python from 0.993 to 0.996 [#19 with comments by Yong-woon Lee]
Fixed concordance bug for Python2 [6caa929]
Fixed pprint bug on Windows [#37]
Version 0.3.3¶
Released on Sep 7, 2014.
Version 0.3.2¶
Released on Sep 4, 2014.
Version 0.3.1¶
Released on Sep 4, 2014.
Added MeCab installer script
Fixed Morph modules to handle strings with whitespaces only
Fixed data inclusion error for Hannanum
Modified tagger filenames with underscore prefixes
Modified concordance function not to print results by default
Modified Hannanum morph method to analyze
Uploaded KoNLPy to PyPI
Version 0.3.0¶
Released on Aug 25, 2014.
Changed API by automatically initializing JVM for Hannanum, Kkma
Added Kkma module with Kkma 2.0
Added documents using Sphinx and Read the Docs
Added license: GPL v3 or above
Added pretty print function for Unicode
Added noun extractor to Mecab
Enabled installation via
pip
Fixed Hannanum, Kkma module bug where it couldn’t handle empty input strings
Warning
The versions below do not have documents available, and are not backwards-compatible.
Version 0.2¶
Released on Aug 1, 2014.
Changed API by explicitly initializing JVM for Hannanum
Added Mecab module with MeCab-0.996-ko-0.9.1
Added unit tests
Added test automation with Travis CI
Fixed Hannanum module parsing error when ‘/’, ‘+’ are in text
Fixed Hannanum module text indexing error (where results get truncated)
Version 0.1¶
Released on Jun 15, 2014. Initial release of KoNLPy.
Inspired by Heewon Jeon’s KoNLP project, a wrapper of the Hannanum analyzer for R. The name KoNLPy, comes from this project.
Added Hannanum module with JHannanum 0.8.3