''' 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 # 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'] ) # 初始化DeepSeek客户端 client = OpenAI(api_key=config.DEEPSEEK_API_KEY) def search_related_data(query): """搜索向量数据和原始相关数据""" # 向量搜索 vector = text_to_embedding(query) vector_results = es.search( index='knowledge_base', body={ "size": 5, "query": { "script_score": { "query": {"match_all": {}}, "script": { "source": "cosineSimilarity(params.query_vector, 'vector') + 1.0", "params": {"query_vector": vector} } } }, "_source": ["text"] } ) # 文本精确搜索 text_results = es.search( index='raw_texts', body={ "query": { "multi_match": { "query": query, "fields": ["text"], "type": "best_fields" } }, "size": 5 } ) # 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']]) 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}""" 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 if __name__ == "__main__": #user_query = input("请输入您的查询要求:") user_query = "整理云南省初中在校生情况文档" report = process_query(user_query) print("\n=== 生成的报告 ===\n") print(report)