fixed issue in bug/#42

JapaneseTokenizer/common/text_preprocess.py

@@ -8,7 +8,9 @@
 import six
 import re
 import unicodedata
-import neologdn
+from JapaneseTokenizer import init_logger
+import logging
+logger = init_logger.init_logger(logging.getLogger(init_logger.LOGGER_NAME))
 __author__ = 'kensuke-mi'
 
 if six.PY2:
@@ -20,6 +22,13 @@ def u(str): return str
     def b(str): return str.encode("utf-8")
     pass
 
+try:
+    import neologdn
+    is_neologdn_valid = True
+except:
+    logger.warning("neologdn package is not installed yet. You could not call neologd dictionary.")
+    is_neologdn_valid = False
+
 STRING_EXCEPTION = set([u('*')])
 
 
@@ -57,8 +66,10 @@ def normalize_text(input_text,
     else:
         without_new_line = new_line_replaced
 
-    if dictionary_mode=='neologd':
+    if dictionary_mode=='neologd' and is_neologdn_valid:
         return neologdn.normalize(normalize_text_normal_ipadic(without_new_line))
+    elif dictionary_mode=='neologd' and is_neologdn_valid == False:
+        raise Exception("You could not call neologd dictionary bacause you do NOT install the package neologdn.")
     else:
         return normalize_text_normal_ipadic(without_new_line, kana=is_kana, ascii=is_ascii, digit=is_digit)
 

JapaneseTokenizer/mecab_wrapper/mecab_wrapper.py

@@ -10,7 +10,6 @@
 import os
 import logging
 import subprocess
-import neologdn
 import six
 from six import text_type
 # typing
@@ -22,6 +21,13 @@
 logger = init_logger.init_logger(logging.getLogger(init_logger.LOGGER_NAME))
 python_version = sys.version_info
 
+try:
+    import neologdn
+    is_neologdn_valid = True
+except:
+    logger.warning("neologdn package is not installed yet. You could not call neologd dictionary.")
+    is_neologdn_valid = False
+
 
 class MecabWrapper(WrapperBase):
     def __init__(self, dictType, pathUserDictCsv='', path_mecab_config=None, string_encoding='utf-8'):
@@ -243,8 +249,10 @@ def tokenize(self, sentence,
             pass
 
         ### decide normalization function depending on dictType
-        if func_normalizer is None and self._dictType == 'neologd':
+        if func_normalizer is None and self._dictType == 'neologd' and is_neologdn_valid:
             normalized_sentence = neologdn.normalize(sentence)
+        elif func_normalizer is None and self._dictType == 'neologd' and is_neologdn_valid == False:
+            raise Exception("You could not call neologd dictionary bacause you do NOT install the package neologdn.")
         elif func_normalizer == normalize_text:
             normalized_sentence = normalize_text(sentence, dictionary_mode=self._dictType)
         elif func_normalizer is None:

setup.py

@@ -38,17 +38,26 @@
     except ImportError:
         logger.error('We failed to install pyknp automatically. Try installing pyknp manually.')
 # --------------------------------------------------------------------------------------------------------
+try:
+    import neologdn
+except ImportError:
+    try:
+        pip.main(['install', 'neologdn'])
+        import neologdn
+    except:
+        logger.error('We failed to install neologdn automatically because of some issues in the package. Try installing pyknp manually.')
+# --------------------------------------------------------------------------------------------------------
 
 if python_version >= (3, 0, 0):
     logger.info(msg='python={}'.format(python_version))
-    install_requires = ['pypandoc', 'future', 'six', 'mecab-python3', 'jaconv>=0.2', 'pip>=8.1.0', 'typing', 'neologdn', 'pexpect']
+    install_requires = ['pypandoc', 'future', 'six', 'mecab-python3', 'jaconv>=0.2', 'pip>=8.1.0', 'typing', 'pexpect']
 elif python_version <= (2, 9, 9):
     logger.info(msg='python={}'.format(python_version))
-    install_requires = ['pypandoc', 'future', 'six', 'mecab-python', 'jaconv>=0.2', 'pip>=8.1.0', 'typing', 'neologdn', 'pexpect']
+    install_requires = ['pypandoc', 'future', 'six', 'mecab-python', 'jaconv>=0.2', 'pip>=8.1.0', 'typing', 'pexpect']
 else:
     raise NotImplementedError()
 
-version = '1.3.6'
+version = '1.3.7'
 name = 'JapaneseTokenizer'
 short_description = '`JapaneseTokenizer` is a package for easy Japanese Tokenization'