From e1d2472ebaa0146e375eb7bf84d2ddb9c5715ce4 Mon Sep 17 00:00:00 2001
From: HuangHai <10402852@qq.com>
Date: Mon, 23 Jun 2025 20:28:58 +0800
Subject: [PATCH] 'commit'

---
 dsRag/T9_RAG.py | 121 +++++++++++++++++++++++++++---------------------
 1 file changed, 68 insertions(+), 53 deletions(-)

diff --git a/dsRag/T9_RAG.py b/dsRag/T9_RAG.py
index 3e219b76..1d14c448 100644
--- a/dsRag/T9_RAG.py
+++ b/dsRag/T9_RAG.py
@@ -3,50 +3,94 @@ pip install openai
 '''
 from elasticsearch import Elasticsearch
 from Util.EmbeddingUtil import text_to_embedding
-import Config.Config as config
 from openai import OpenAI
-import json
-
+from Config import Config
 # 初始化ES连接
 es = Elasticsearch(
-    hosts=config.ES_CONFIG['hosts'],
-    basic_auth=config.ES_CONFIG['basic_auth'],
-    verify_certs=config.ES_CONFIG['verify_certs']
+    hosts=Config.ES_CONFIG['hosts'],
+    basic_auth=Config.ES_CONFIG['basic_auth'],
+    verify_certs=Config.ES_CONFIG['verify_certs']
 )
 
 # 初始化DeepSeek客户端
-client = OpenAI(api_key=config.DEEPSEEK_API_KEY)
+client = OpenAI(
+    api_key=Config.DEEPSEEK_API_KEY,
+    base_url=Config.DEEPSEEK_URL
+)
+
+def generate_report(query, context):
+    """使用DeepSeek生成报告"""
+    prompt = f"""根据以下关于'{query}'的相关信息，整理一份结构化的报告：
+要求：
+1. 分章节组织内容
+2. 包含关键数据和事实
+3. 语言简洁专业
+
+相关信息：
+{context}"""
+    
+    try:
+        response = client.chat.completions.create(
+            model="deepseek-chat",
+            messages=[
+                {"role": "system", "content": "你是一个专业的文档整理助手"},
+                {"role": "user", "content": prompt}
+            ],
+            temperature=0.3,
+            stream=True
+        )
+        
+        # 流式输出处理
+        full_response = ""
+        for chunk in response:
+            if chunk.choices[0].delta.content:
+                content = chunk.choices[0].delta.content
+                print(content, end="", flush=True)
+                full_response += content
+        
+        return full_response
+    except Exception as e:
+        print(f"生成报告时出错: {str(e)}")
+        return ""
+
+def process_query(query):
+    """处理用户查询并生成报告"""
+    print(f"正在搜索与'{query}'相关的数据...")
+    context = search_related_data(query)
+    print(f"找到{len(context.split(chr(10)+chr(10)))}条相关数据")
+    
+    print("正在生成报告...")
+    report = generate_report(query, context)
+    
+    return report
 
 def search_related_data(query):
-    """搜索向量数据和原始相关数据"""
+    """搜索与查询相关的数据"""
     # 向量搜索
-    vector = text_to_embedding(query)
+    query_vector = text_to_embedding(query)
     vector_results = es.search(
-        index='knowledge_base',
+        index=Config.ES_CONFIG['default_index'],
         body={
-            "size": 5,
             "query": {
                 "script_score": {
                     "query": {"match_all": {}},
                     "script": {
                         "source": "cosineSimilarity(params.query_vector, 'vector') + 1.0",
-                        "params": {"query_vector": vector}
+                        "params": {"query_vector": query_vector}
                     }
                 }
             },
-            "_source": ["text"]
+            "size": 5
         }
     )
     
     # 文本精确搜索
     text_results = es.search(
-        index='raw_texts',
+        index="raw_texts",
         body={
             "query": {
-                "multi_match": {
-                    "query": query,
-                    "fields": ["text"],
-                    "type": "best_fields"
+                "match": {
+                    "text.keyword": query
                 }
             },
             "size": 5
@@ -54,43 +98,14 @@ def search_related_data(query):
     )
     
     # 合并结果
-    all_results = [hit['_source']['text'] for hit in vector_results['hits']['hits']]
-    all_results.extend([hit['_source']['text'] for hit in text_results['hits']['hits']])
+    context = ""
+    for hit in vector_results['hits']['hits']:
+        context += f"向量相似度结果(score={hit['_score']}):\n{hit['_source']['text']}\n\n"
     
-    return "\n\n".join(all_results)
-
-def generate_report(query, context):
-    """使用DeepSeek生成报告"""
-    prompt = f"""根据以下关于'{query}'的相关信息，整理一份结构化的报告：
-要求：
-1. 分章节组织内容
-2. 包含关键数据和事实
-3. 语言简洁专业
-
-相关信息：
-{context}"""
+    for hit in text_results['hits']['hits']:
+        context += f"文本精确匹配结果(score={hit['_score']}):\n{hit['_source']['text']}\n\n"
     
-    response = client.chat.completions.create(
-        model="deepseek-chat",
-        messages=[
-            {"role": "system", "content": "你是一个专业的文档整理助手"},
-            {"role": "user", "content": prompt}
-        ],
-        temperature=0.3
-    )
-    
-    return response.choices[0].message.content
-
-def process_query(query):
-    """处理用户查询并生成报告"""
-    print(f"正在搜索与'{query}'相关的数据...")
-    context = search_related_data(query)
-    print(f"找到{len(context.split(chr(10)+chr(10)))}条相关数据")
-    
-    print("正在生成报告...")
-    report = generate_report(query, context)
-    
-    return report
+    return context
 
 if __name__ == "__main__":
     #user_query = input("请输入您的查询要求：")