重写地区识别模糊匹配

src/agent/code.py

@@ -91,9 +91,14 @@ class AreaCodeTool:
             
         # 模糊匹配
         if not results:
-            mask = self.df['name'].str.contains(area_name, na=False)
-            matches = self.df[mask]
-            results.extend([(row['name'], row['code']) for _, row in matches.iterrows()])
+            for name in self.full_name_map.keys():
+                if self.is_subsequence(name,area_name):
+                    results.append((area_name, self.full_name_map[name]))
+                    return results
+
+            # mask = self.df['name'].str.contains(area_name, na=False)
+            # matches = self.df[mask]
+            # results.extend([(row['name'], row['code']) for _, row in matches.iterrows()])
             
         return results
 
@@ -113,18 +118,37 @@ class AreaCodeTool:
             return matches.iloc[0]['name']
         return None
 
+    def is_subsequence(self,source, target):
+        # 初始化两个索引，分别指向源字符串和目标字符串的开头
+        source_index = 0
+        target_index = 0
+
+        # 遍历源字符串，直到找到目标字符串的所有字符或者遍历完源字符串
+        while source_index < len(source) and target_index < len(target):
+            # 如果当前源字符串的字符等于目标字符串的字符，则移动目标字符串的索引
+            if source[source_index] == target[target_index]:
+                target_index += 1
+            # 无论是否匹配，都移动源字符串的索引
+            source_index += 1
+
+        # 如果目标字符串的索引已经遍历完，说明找到了完整的子序列
+        return target_index == len(target)
+
 # 使用示例
 def example_usage():
     tool = AreaCodeTool()
     
     # 测试不同类型的查询
     test_cases = [
-        "安徽省",
-        "安庆市",
-        "迎江区",
-        "安徽省安庆市",
-        "安徽省安庆市迎江区",
-        "安庆"  # 模糊查询
+        "贵州省",
+        "贵阳市",
+        "云岩区",
+        "贵州省贵阳市",
+        "贵州省贵阳市南明区",
+        "贵阳",  # 模糊查询
+        "贵州贵阳" , # 模糊查询
+        "贵州贵阳南明",  # 模糊查询
+        "贵州贵阳市南明" , # 模糊查询
     ]
     
     for query in test_cases: