Merge 735d721 into 7b9281a

Makefile

@@ -56,7 +56,7 @@ testall:
 	python3 -m pytest --cov=konlpy test/
 
 init_i18n:
-	pip install mock sphinx sphinx-intl tweepy colorama bs4
+	pip install mock sphinx sphinx-intl
 	git submodule init
 	git submodule update
 

konlpy/corpus.py

@@ -21,6 +21,12 @@ class CorpusLoader():
         유구한 역사와 전통에 빛나는 우리 대한국민은 3·1운동으로 건립된 대한민국임시정부의 법통과 불의에 항거한 4·19민주이념을 계승하고, 조국의 민주개혁과 평화적 통일의 사명에 입각하여 정의·인도와 동포애로써 민족의 단결을 공고히 하고, 모든 사회적 폐습과 불의를 타파하며, 자율과 조화를 바 바
     """
 
+    def __init__(self, name=None):
+        if not name:
+            raise Exception("You need to input the name of the corpus")
+        else:
+            self.name = name
+
     def abspath(self, filename=None):
         """Absolute path of corpus file.
         If ``filename`` is *None*, returns absolute path of corpus.
@@ -45,12 +51,6 @@ def open(self, filename):
         """
         return utils.load_txt(self.abspath(filename))
 
-    def __init__(self, name=None):
-        if not name:
-            raise Exception("You need to input the name of the corpus")
-        else:
-            self.name = name
-
 
 kolaw = CorpusLoader('kolaw')
 kobill = CorpusLoader('kobill')

konlpy/downloader.py

@@ -61,6 +61,9 @@ class Downloader(object):
     NOT_INSTALLED = 'not installed'
     STALE = 'corrupt or out of date'
 
+    def __init__(self, download_dir=None):
+        self._download_dir = download_dir
+
     def download(self, id=None, download_dir=None):
         """The KoNLPy data downloader.
         With this module you can download corpora, models and other data packages
@@ -220,8 +223,6 @@ def _get_info(self, id):
         else:
             raise ValueError("Could not find a matching item to download")
 
-    def __init__(self, download_dir=None):
-        self._download_dir = download_dir
 
 # Aliases
 _downloader = Downloader(default_download_dir())

konlpy/tag/_common.py

@@ -2,6 +2,12 @@
 """ Common utility function for tagger classes """
 from __future__ import absolute_import
 from __future__ import unicode_literals
+import sys
+
+
+# For both Python 2 and Python 3 compatibility
+if sys.version_info[0] >= 3:
+    basestring = str
 
 
 def validate_phrase_inputs(phrase):
@@ -11,4 +17,4 @@ def validate_phrase_inputs(phrase):
         phrase (str): phrase input
     """
     msg = "phrase input should be string, not %s" % type(phrase)
-    assert isinstance(phrase, str), msg
+    assert isinstance(phrase, basestring), msg

konlpy/tag/_hannanum.py

@@ -64,6 +64,15 @@ class Hannanum():
     :param max_heap_size: Maximum memory usage limitation (Megabyte) :py:func:`.init_jvm`.
     """
 
+    def __init__(self, jvmpath=None, max_heap_size=1024):
+        if not jpype.isJVMStarted():
+            jvm.init_jvm(jvmpath, max_heap_size)
+
+        jhannanumJavaPackage = jpype.JPackage('kr.lucypark.jhannanum.comm')
+        HannanumInterfaceJavaClass = jhannanumJavaPackage.HannanumInterface
+        self.jhi = HannanumInterfaceJavaClass()  # Java instance
+        self.tagset = utils.read_json('%s/data/tagset/hannanum.json' % utils.installpath)
+
     def analyze(self, phrase):
         """Phrase analyzer.
 
@@ -104,12 +113,3 @@ def morphs(self, phrase):
         """Parse phrase to morphemes."""
 
         return [s for s, t in self.pos(phrase)]
-
-    def __init__(self, jvmpath=None, max_heap_size=1024):
-        if not jpype.isJVMStarted():
-            jvm.init_jvm(jvmpath, max_heap_size)
-
-        jhannanumJavaPackage = jpype.JPackage('kr.lucypark.jhannanum.comm')
-        HannanumInterfaceJavaClass = jhannanumJavaPackage.HannanumInterface
-        self.jhi = HannanumInterfaceJavaClass()  # Java instance
-        self.tagset = utils.read_json('%s/data/tagset/hannanum.json' % utils.installpath)

konlpy/tag/_kkma.py

@@ -39,6 +39,15 @@ class Kkma():
     :param max_heap_size: Maximum memory usage limitation (Megabyte) :py:func:`.init_jvm`.
     """
 
+    def __init__(self, jvmpath=None, max_heap_size=1024):
+        if not jpype.isJVMStarted():
+            jvm.init_jvm(jvmpath, max_heap_size)
+
+        kkmaJavaPackage = jpype.JPackage('kr.lucypark.kkma')
+        KkmaInterfaceJavaClass = kkmaJavaPackage.KkmaInterface
+        self.jki = KkmaInterfaceJavaClass()  # Java instance
+        self.tagset = utils.read_json('%s/data/tagset/kkma.json' % utils.installpath)
+
     def nouns(self, phrase):
         """Noun extractor."""
 
@@ -91,12 +100,3 @@ def sentences(self, phrase):
         sentences = self.jki.morphAnalyzer(phrase)
         if not sentences: return []
         return [sentences.get(i).getSentence() for i in range(sentences.size())]
-
-    def __init__(self, jvmpath=None, max_heap_size=1024):
-        if not jpype.isJVMStarted():
-            jvm.init_jvm(jvmpath, max_heap_size)
-
-        kkmaJavaPackage = jpype.JPackage('kr.lucypark.kkma')
-        KkmaInterfaceJavaClass = kkmaJavaPackage.KkmaInterface
-        self.jki = KkmaInterfaceJavaClass()  # Java instance
-        self.tagset = utils.read_json('%s/data/tagset/kkma.json' % utils.installpath)

konlpy/tag/_komoran.py

@@ -51,6 +51,28 @@ class Komoran():
     :param max_heap_size: Maximum memory usage limitation (Megabyte) :py:func:`.init_jvm`.
     """
 
+    def __init__(self, jvmpath=None, userdic=None, modelpath=None, max_heap_size=1024):
+        if not jpype.isJVMStarted():
+            jvm.init_jvm(jvmpath, max_heap_size)
+
+        if modelpath:
+            self.modelpath = modelpath
+        else:
+            # FIXME: Cannot execute without sudoing
+            # java.lang.NoClassDefFoundErrorPyRaisable: java.lang.NoClassDefFoundError: kr/co/shineware/nlp/komoran/core/analyzer/Komoran
+            self.modelpath = os.path.join(utils.installpath, 'java', 'data', 'models')
+        self.tagset = utils.read_json('%s/data/tagset/komoran.json' % utils.installpath)
+
+        komoranJavaPackage = jpype.JPackage('kr.co.shineware.nlp.komoran.core')
+
+        try:
+            self.jki = komoranJavaPackage.Komoran(self.modelpath)
+        except TypeError:  # Package kr.lucypark.komoran.KomoranInterface is not Callable
+            raise IOError("Cannot access komoran-dic. Please leave an issue at https://github.com/konlpy/konlpy/issues")
+
+        if userdic:
+            self.jki.setUserDic(userdic)
+
     def pos(self, phrase, flatten=True, join=False):
         """POS tagger.
 
@@ -89,25 +111,3 @@ def morphs(self, phrase):
         """Parse phrase to morphemes."""
 
         return [s for s, t in self.pos(phrase)]
-
-    def __init__(self, jvmpath=None, userdic=None, modelpath=None, max_heap_size=1024):
-        if not jpype.isJVMStarted():
-            jvm.init_jvm(jvmpath, max_heap_size)
-
-        if modelpath:
-            self.modelpath = modelpath
-        else:
-            # FIXME: Cannot execute without sudoing
-            # java.lang.NoClassDefFoundErrorPyRaisable: java.lang.NoClassDefFoundError: kr/co/shineware/nlp/komoran/core/analyzer/Komoran
-            self.modelpath = os.path.join(utils.installpath, 'java', 'data', 'models')
-        self.tagset = utils.read_json('%s/data/tagset/komoran.json' % utils.installpath)
-
-        komoranJavaPackage = jpype.JPackage('kr.co.shineware.nlp.komoran.core')
-
-        try:
-            self.jki = komoranJavaPackage.Komoran(self.modelpath)
-        except TypeError:  # Package kr.lucypark.komoran.KomoranInterface is not Callable
-            raise IOError("Cannot access komoran-dic. Please leave an issue at https://github.com/konlpy/konlpy/issues")
-
-        if userdic:
-            self.jki.setUserDic(userdic)

konlpy/tag/_mecab.py

@@ -68,6 +68,26 @@ class Mecab():
     .. _Eunjeon Project: http://eunjeon.blogspot.kr/
     """
 
+    def __init__(self, dicpath='/usr/local/lib/mecab/dic/mecab-ko-dic'):
+        self.dicpath = dicpath
+        try:
+            self.tagger = Tagger('-d %s' % dicpath)
+            self.tagset = utils.read_json('%s/data/tagset/mecab.json' % utils.installpath)
+        except RuntimeError:
+            raise Exception('The MeCab dictionary does not exist at "%s". Is the dictionary correctly installed?\nYou can also try entering the dictionary path when initializing the Mecab class: "Mecab(\'/some/dic/path\')"' % dicpath)
+        except NameError:
+            raise Exception('Install MeCab in order to use it: http://konlpy.org/en/latest/install/')
+
+    def __setstate__(self, state):
+        """just reinitialize."""
+
+        self.__init__(dicpath=state['dicpath'])
+
+    def __getstate__(self):
+        """store arguments."""
+
+        return {'dicpath': self.dicpath}
+
     # TODO: check whether flattened results equal non-flattened
     def pos(self, phrase, flatten=True, join=False):
         """POS tagger.
@@ -103,23 +123,3 @@ def nouns(self, phrase):
 
         tagged = self.pos(phrase)
         return [s for s, t in tagged if t.startswith('N')]
-
-    def __init__(self, dicpath='/usr/local/lib/mecab/dic/mecab-ko-dic'):
-        self.dicpath = dicpath
-        try:
-            self.tagger = Tagger('-d %s' % dicpath)
-            self.tagset = utils.read_json('%s/data/tagset/mecab.json' % utils.installpath)
-        except RuntimeError:
-            raise Exception('The MeCab dictionary does not exist at "%s". Is the dictionary correctly installed?\nYou can also try entering the dictionary path when initializing the Mecab class: "Mecab(\'/some/dic/path\')"' % dicpath)
-        except NameError:
-            raise Exception('Install MeCab in order to use it: http://konlpy.org/en/latest/install/')
-
-    def __setstate__(self, state):
-        """just reinitialize."""
-
-        self.__init__(dicpath=state['dicpath'])
-
-    def __getstate__(self):
-        """store arguments."""
-
-        return {'dicpath': self.dicpath}

konlpy/tag/_okt.py

@@ -46,6 +46,15 @@ class Okt():
     :param max_heap_size: Maximum memory usage limitation (Megabyte) :py:func:`.init_jvm`.
     """
 
+    def __init__(self, jvmpath=None, max_heap_size=1024):
+        if not jpype.isJVMStarted():
+            jvm.init_jvm(jvmpath, max_heap_size)
+
+        oktJavaPackage = jpype.JPackage('kr.lucypark.okt')
+        OktInterfaceJavaClass = oktJavaPackage.OktInterface
+        self.jki = OktInterfaceJavaClass()
+        self.tagset = utils.read_json('%s/data/tagset/twitter.json' % utils.installpath)
+
     def pos(self, phrase, norm=False, stem=False, join=False):
         """POS tagger.
         In contrast to other classes in this subpackage,
@@ -87,12 +96,3 @@ def phrases(self, phrase):
     def normalize(self, phrase):
         text = self.jki.normalize(phrase)
         return text
-
-    def __init__(self, jvmpath=None, max_heap_size=1024):
-        if not jpype.isJVMStarted():
-            jvm.init_jvm(jvmpath, max_heap_size)
-
-        oktJavaPackage = jpype.JPackage('kr.lucypark.okt')
-        OktInterfaceJavaClass = oktJavaPackage.OktInterface
-        self.jki = OktInterfaceJavaClass()
-        self.tagset = utils.read_json('%s/data/tagset/twitter.json' % utils.installpath)

requirements.txt

@@ -1,5 +1,3 @@
 JPype1>=0.7.0
-beautifulsoup4==4.6.0
-colorama
 lxml>=4.1.0
 numpy>=1.6

setup.py

@@ -2,6 +2,7 @@
 # -*- coding: utf-8 -*-
 
 import os
+import sys
 import platform
 from setuptools import find_packages, setup
 
@@ -23,7 +24,11 @@ def _openreq(reqfile):
         with open(os.path.join(os.path.dirname(__file__), reqfile)) as f:
             return f.read().splitlines()
 
-    return _openreq('requirements.txt')
+    req = _openreq('requirements.txt')
+    if sys.version_info[0] < 3:
+        req[0] = req[0].replace(">=", "==")
+
+    return 
 
 
 about = get_about()