更新环境变量配置，调整模型名称获取方式，新增Dify API相关配置，删除无用的脚本文件，优化意图识别逻辑，添加LLM提取词条逻辑

rag2_0/dify/AnalysisDifyAnswer.py

@@ -3,19 +3,15 @@ import json
 
 from regex import search
 
-import ijson
+import sys
+import os
+sys.path.append(os.getcwd())
+from rag2_0.dify.dify_tool import DifyTool
 
-df = pd.read_excel("data/excel/已分析数据汇总（第一轮）.xlsx")
-df=df[df["评价"]=="dislike"]
+dify_tool = DifyTool()
+
+df = pd.read_excel("data/excel/0714提问数据汇总（已分析）_软件.xlsx")
 
-msg_id_list = df["msg_id"].tolist()
-msg_debug_list = {}
-# 流式解析 JSON 数组
-with open("data/excel/msg_debug_list.json", "r", encoding="utf-8") as f:
-    # 使用ijson.items直接获取顶层键值对
-    for msg_id, data in ijson.kvitems(f, ''):
-        if msg_id in msg_id_list:
-            msg_debug_list[msg_id] = data
 
 def get_rewrite_query(intent_node_execution_info)->str:
     outputs_result =json.loads(intent_node_execution_info['outputs'])
@@ -28,7 +24,7 @@ def judge_error_node_and_reason(intent_node_execution_info, knowledge_filter_nod
     
     outputs_result =json.loads(intent_node_execution_info['outputs'])
     result["问题改写结果"] = outputs_result['optimize_query']
-    if outputs_result['is_complete'] == False:
+    if outputs_result['is_complete'] == False and outputs_result["has_slot_filling"] == True:
         result["错误环节"] = "槽点填充"
         result["错误原因"] = f"槽点缺失"
         result["具体描述"] = f"缺失内容：{outputs_result['missing_slots']}"
@@ -80,6 +76,8 @@ for index, row in df.iterrows():
         answer = row["回答"]
         query = row["提问"]
         rating = row["评价"]
+        if rating != "dislike":
+            continue
         class_type = row["问题分类"]
         dislike_reason = row["点踩原因"]
         if dislike_reason is None or pd.isna(dislike_reason):
@@ -87,7 +85,8 @@ for index, row in df.iterrows():
         
         answer_wiki_name = row["关联词条"]
         search_wiki = row["检索到的词条"]
-        node_executions_info = msg_debug_list[msg_id]
+        msg_debug_info = dify_tool.get_message_debug_info_by_id(msg_id)
+        node_executions_info = msg_debug_info["workflow_node_executions_info"]
         intent_node_execution_info = [node_execution_info for node_execution_info in node_executions_info 
                                     if node_execution_info["title"] == "意图识别结果解析"]
         
@@ -109,7 +108,7 @@ for index, row in df.iterrows():
         print(f"msg_id: {msg_id} 处理失败: {e}")
         continue
 
-df.to_excel("data/excel/已分析数据汇总（第一轮）_分析.xlsx", index=False)
+df.to_excel("data/excel/0714提问数据汇总（已分析）_软件_分析.xlsx", index=False)