3.4

kss/base.py

@@ -144,21 +144,7 @@ def exceptions():
                 year_s.append(f"{i}{j}s")
                 year_s.append(f"{i}{j}S")
 
-        time = []
-        for i in Const.numbers:
-            for j in Const.numbers:
-                for k in Const.single_quotes:
-                    time.append(f"{i}{k}{j}")
-                    time.append(f"{i}{k}{j}")
-                    time.append(f"{i}{k}{j}")
-
-        inch = []
-        for i in Const.numbers + ["."]:
-            for j in Const.numbers:
-                for k in Const.double_quotes:
-                    inch.append(f"{i}{j}{k}")
-
-        return faces + apostrophe + year_s + time + inch
+        return faces + apostrophe + year_s
 
     @staticmethod
     def ec_cases():

kss/kss.py

@@ -132,8 +132,9 @@ def split_sentences(
 
     for _input_for_pp in mp_temp:
         out = "".join(_input_for_pp).replace(" ", "")
-        for special in Const.quotes_or_brackets:
-            out = out.replace(special, "")
+        if use_quotes_brackets_processing:
+            for special in Const.quotes_or_brackets:
+                out = out.replace(special, "")
 
         mp_postprocessing.append(out)
 
@@ -169,8 +170,10 @@ def split_sentences(
     for result in _results:
         mp_temp += result
         out = "".join(mp_temp).replace(" ", "")
-        for special in Const.quotes_or_brackets:
-            out = out.replace(special, "")
+        if use_quotes_brackets_processing:
+            for special in Const.quotes_or_brackets:
+                out = out.replace(special, "")
+            mp_postprocessing = [m.replace("\u200b", "") for m in mp_postprocessing]
 
         if out in mp_postprocessing:
             mp_output_final.append(mp_temp)
@@ -243,8 +246,20 @@ def _split_sentences(
     backend: str,
     recover_step: int = 0,
 ):
-    if text in _cache.dic:
-        return _cache.get(text)
+    if use_quotes_brackets_processing:
+        text = text.replace("\u200b", "")
+
+    cache_key = (
+        text,
+        use_heuristic,
+        use_quotes_brackets_processing,
+        max_recover_step,
+        max_recover_length,
+        backend,
+    )
+
+    if cache_key in _cache.dic:
+        return _cache.get(cache_key)
     else:
         original_text = deepcopy(text)
 
@@ -259,8 +274,9 @@ def _split_sentences(
     text = prep.add_emojis_to_dict(text)
     text = prep.backup(text)
 
-    for s in Const.quotes_or_brackets:
-        text = text.replace(s, f"\u200b{s}\u200b")
+    if use_quotes_brackets_processing:
+        for s in Const.quotes_or_brackets:
+            text = text.replace(s, f"\u200b{s}\u200b")
 
     if use_morpheme:
         eojeols = _morph.pos(text=text, backend=backend)
@@ -583,10 +599,21 @@ def _split_sentences(
     outputs = []
     for s in results:
         s = prep.restore(s)
-        s = s.replace("\u200b", "")
+        if use_quotes_brackets_processing:
+            s = s.replace("\u200b", "")
         outputs.append(s)
 
-    _cache.put(original_text, outputs)
+    _cache.put(
+        (
+            original_text,
+            use_heuristic,
+            use_quotes_brackets_processing,
+            max_recover_step,
+            max_recover_length,
+            backend,
+        ),
+        outputs,
+    )
 
     return outputs
 

setup.py

@@ -17,7 +17,7 @@ def read_file(filename, cb):
 
 setup(
     name="kss",
-    version="3.3.1.1",
+    version="3.4",
     author="Hyunwoong Ko",
     author_email="[email protected]",
     url="https://github.com/hyunwoongko/kss",

tests/test_kss.py

@@ -160,3 +160,26 @@ def test_chunks(self):
             "전하 아니되옵니다. 부디 용서하옵소서. 소인이 큰 죄를 저질렀사옵니다.\n", max_length=6, overlap=True
         )
         print(out)
+
+    def test_yjy2026(self):
+        output = kss.split_sentences("1'1″")
+        print(output)
+
+    def test_hannabros(self):
+        output = kss.split_sentences('분리 할 수 ​​있다.')
+        print(output)
+
+    def test_lifelongeek(self):
+        text = "우리가 타이라는 단어는 원래 동사로 묶다 엮다라는 표현이고요 그다음에 명사로 하게 되면 묶음이라는 표현이죠 그런데"
+        output_1 = kss.split_sentences(text)
+        for i in range(500):
+            output = kss.split_sentences("안녕")
+        text = "우리가 타이라는 단어는 원래 동사로 묶다 엮다라는 표현이고요 그다음에 명사로 하게 되면 묶음이라는 표현이죠 그런데"
+        output_2 = kss.split_sentences(text)
+        print(output_1, output_2)
+        # 음.. 뭘까 잘 되는데
+
+    def test_newdboy(self):
+        tst = '그것이 잘 적혀 있는지 확인해야 한다고 했기 때문이다. EBS 미래교육연구소 최홍규 박사도 그렇게 말했다'
+        output = kss.split_sentences(text=tst, backend='mecab', use_heuristic=True)
+        print(output)