'commit'

dsSchoolBuddy/ElasticSearch/T5_SelectByTags.py

@@ -16,22 +16,14 @@ es = Elasticsearch(
     verify_certs=False
 )
 
-# 2. 直接在代码中指定要查询的标签
-query_tag = ["MATH_1"]  # 可以修改为其他需要的标签
+# 2. 直接在代码中指定要查询的关键字
+query_keyword = "混凝土"
 
 # 3. 构建查询条件
 query = {
     "query": {
-        "bool": {
-            "should": [
-                {
-                    "terms": {
-                        "tags.tags": query_tag
-                    }
-                }
-            ],
-            "minimum_should_match": 1
-            
+        "match": {
+            "user_input": query_keyword
         }
     },
     "size": 1000
@@ -39,41 +31,23 @@ query = {
 
 # 4. 执行查询并处理结果
 try:
-    response = es.search(index="knowledge_base", body=query)
-    hits = response['hits']['hits']
-    
-    if not hits:
-        print(f"未找到标签为 '{query_tag}' 的数据。")
-    else:
-        print(f"找到 {len(hits)} 条标签为 '{query_tag}' 的数据：")
-        for i, hit in enumerate(hits, 1):
+    results = es.search(index=ES_CONFIG['index_name'], body=query)
+    print(f"查询关键字 '{query_keyword}' 结果：")
+    if results['hits']['hits']:
+        for i, hit in enumerate(results['hits']['hits'], 1):
+            doc = hit['_source']
             print(f"{i}. ID: {hit['_id']}")
-            print(f"   内容: {hit['_source'].get('user_input', '')}")
-            print(f"   标签: {hit['_source']['tags']['tags']}")
+            print(f"   标签: {doc['tags']['tags']}")
+            print(f"   用户问题: {doc['user_input']}")
+            print(f"   时间: {doc['timestamp']}")
+            print(f"   向量: {doc['embedding'][:5]}...")
             print("-" * 50)
+    else:
+        print(f"未找到包含 '{query_keyword}' 的数据。")
 except Exception as e:
-    print(f"查询执行失败，错误详情: {str(e)}")
+    print(f"查询失败: {e}")
     print(f"查询条件: {query}")
     if hasattr(e, 'info') and isinstance(e.info, dict):
         print(f"Elasticsearch错误详情: {e.info}")
     if hasattr(e, 'status_code'):
-        print(f"HTTP状态码: {e.status_code}")
-    print(f"查询出错: {str(e)}")
-
-# 4. 执行查询
-try:
-    results = es.search(index=ES_CONFIG['index_name'], body=query)
-    print(f"查询标签 '{query_tag}' 结果：")
-    if results['hits']['hits']:
-        for hit in results['hits']['hits']:
-            doc = hit['_source']
-            print(f"ID: {hit['_id']}")
-            print(f"标签: {doc['tags']['tags']}")
-            print(f"用户问题: {doc['user_input']}")
-            print(f"时间: {doc['timestamp']}")
-            print(f"向量: {doc['embedding'][:5]}...")
-            print("-" * 40)
-    else:
-        print(f"未找到标签为 '{query_tag}' 的数据。")
-except Exception as e:
-    print(f"查询失败: {e}")
+        print(f"HTTP状态码: {e.status_code}")