✏️ typo

nipunsadvilkar · Jul 5, 2020 · d306e7c · d306e7c
1 parent 5fb5cc3
commit d306e7c
Show file tree

Hide file tree

Showing 3 changed files with 19 additions and 19 deletions.
diff --git a/pysbd/exclaimation_words.py b/pysbd/exclaimation_words.py
diff --git a/pysbd/exclamation_words.py b/pysbd/exclamation_words.py
@@ -0,0 +1,17 @@
+# -*- coding: utf-8 -*-
+import re
+from pysbd.punctuation_replacer import replace_punctuation
+
+
+class ExclamationWords(object):
+    """
+    Searches for exclamation points that are part of words
+    and not ending punctuation and replaces them.
+    """
+    EXCLAMATION_WORDS = "!Xũ !Kung ǃʼOǃKung !Xuun !Kung-Ekoka ǃHu ǃKhung ǃKu ǃung ǃXo ǃXû ǃXung ǃXũ !Xun Yahoo! Y!J Yum!".split()
+    EXCLAMATION_REGEX = r"|".join(re.escape(w) for w in EXCLAMATION_WORDS)
+
+    @classmethod
+    def apply_rules(cls, text):
+        return re.sub(ExclamationWords.EXCLAMATION_REGEX, replace_punctuation,
+                      text)
diff --git a/pysbd/processor.py b/pysbd/processor.py
@@ -3,7 +3,7 @@
 import spacy
 from pysbd.utils import Text, TextSpan
 from pysbd.lists_item_replacer import ListItemReplacer
-from pysbd.exclaimation_words import ExclaimationWords
+from pysbd.exclamation_words import ExclamationWords
 from pysbd.between_punctuation import BetweenPunctuation
 from pysbd.abbreviation_replacer import AbbreviationReplacer
 
@@ -162,7 +162,7 @@ def check_for_punctuation(self, txt):
     def process_text(self, txt):
         if txt[-1] not in self.lang.Punctuations:
             txt += 'ȸ'
-        txt = ExclaimationWords.apply_rules(txt)
+        txt = ExclamationWords.apply_rules(txt)
         txt = self.between_punctuation(txt)
         # handle text having only doublepunctuations
         if not re.match(self.lang.DoublePunctuationRules.DoublePunctuation, txt):