optimize segmentation

app/utils/utils.py

@@ -163,12 +163,27 @@ def str_contains_punctuation(word):
 def split_string_by_punctuations(s):
     result = []
     txt = ""
-    for char in s:
+
+    previous_char = ""
+    next_char = ""
+    for i in range(len(s)):
+        char = s[i]
+        if i > 0:
+            previous_char = s[i - 1]
+        if i < len(s) - 1:
+            next_char = s[i + 1]
+
+        if char == "." and previous_char.isdigit() and next_char.isdigit():
+            # 取现1万，按2.5%收取手续费, 2.5 中的 . 不能作为换行标记
+            txt += char
+            continue
+
         if char not in const.PUNCTUATIONS:
             txt += char
         else:
             result.append(txt.strip())
             txt = ""
+
     return result