优化与调整

photo_review/util/data_util.py

@@ -6,7 +6,12 @@ from datetime import datetime
 def handle_decimal(string):
     if not string:
         return ""
-    return re.sub(r'[^0-9.]', '', string)
+    string = re.sub(r'[^0-9.]', '', string)
+    front, back = string.rsplit('.', 1)
+    front = front.replace(".", "")
+    if back:
+        back = "." + back
+    return front + back
 
 
 # 处理日期类数据
@@ -14,8 +19,13 @@ def handle_date(string):
     if not string:
         return ""
 
-    string = string.replace("年", "-").replace("月", "-").replace("日", "")
+    string = string.replace("年", "-").replace("月", "-").replace("日", "").replace("/", "-").replace(".", "-")
     string = re.sub(r'[^0-9-]', '', string)
+    length = len(string)
+    if length > 8 and "-" not in string:
+        string = string[:8]
+    elif length > 10 and "-" in string:
+        string = string[:10]
     if is_valid_date_format(string):
         return string
     else:
@@ -48,3 +58,20 @@ def is_valid_date_format(date_str):
             pass
 
     return False
+
+
+def handle_department(string):
+    result = []
+    if not string:
+        return result
+    result.append(string)
+    string_without_num = re.sub(r'\d|一|二|三|四|五|六|七|八|九|十', '', string)
+    if string_without_num != string:
+        result.append(string_without_num)
+    string_without_brackets = re.sub(r'\([^()]*\)|\[[^\[\]]*\]|\{[^\{\}]*\}|（[^（）]*）', "", string_without_num)
+    if string_without_brackets != string_without_num:
+        result.append(string_without_brackets)
+    pure_string = string_without_brackets.split("科")[0] + "科"
+    if pure_string != string_without_brackets:
+        result.append(pure_string)
+    return result