RAG 实践：基于腾讯云 ES 与混元，十分钟构建你的专属 AI 助手

PUT /index-name
{
  "mappings": {
    "properties": {
    "title": {
        "type": "keyword"
      },
      "content": {
        "type": "text"
      },
      "url": {
        "type": "keyword"
      }
    }
  }
}

PUT /_ingest/pipeline/index-name-pipeline
{
  "description": "Text embedding pipeline",
  "processors": [
    {
      "inference": {
        "model_id": ".multilingual-e5-small_linux-x86_64",
        "input_output": [
          {
            "input_field": "content",
            "output_field": "content_embedding"
          },
          {
            "input_field": "title",
            "output_field": "title_embedding"

          }
        ]
      }
    }
  ]
}

POST index-name/_bulk?pipeline=index-name-pipeline&refresh
{ "index" : {} }
{ "title" : "标题 1","content": "内容 1","url": "https：url1"  }
{ "index" : {} }
{ "title" : "标题 2","content": "内容 2","url": "https：url2"  }

import json
from tencentcloud.common import credential
from tencentcloud.common.profile.client_profile import ClientProfile
from tencentcloud.common.profile.http_profile import HttpProfile
from tencentcloud.common.exception.tencent_cloud_sdk_exception import TencentCloudSDKException
from tencentcloud.hunyuan.v20230901 import hunyuan_client, models

def hunyuan_gpt(system_prompt, content):
    try:
        # 密钥可前往官网控制台 https://console.cloud.tencent.com/cam/capi 进行获取
        cred = credential.Credential("SecretId", "SecretKey")        
        # 实例化一个http选项，可选的，没有特殊需求可以跳过
        httpProfile = HttpProfile()
        httpProfile.endpoint = "hunyuan.tencentcloudapi.com"

        # 实例化一个client选项，可选的，没有特殊需求可以跳过
        clientProfile = ClientProfile()
        clientProfile.httpProfile = httpProfile
        
        # 实例化要请求产品的client对象,clientProfile是可选的
        client = hunyuan_client.HunyuanClient(cred, "", clientProfile)

        # 实例化一个请求对象,每个接口都会对应一个request对象
        req = models.ChatCompletionsRequest()
        params = {
            "Model": "hunyuan-standard",
            "Messages": [
                {"Role": "system", "Content": system_prompt},
                {"Role": "user", "Content": content}
            ]
        }
        req.from_json_string(json.dumps(params))

        # 返回的resp是一个ChatCompletionsResponse的实例，与请求对象对应
        resp = client.ChatCompletions(req)
    
        return resp
        
    except TencentCloudSDKException as err:
        return str(err)

pip install streamlit

import streamlit as st
from elasticsearch import Elasticsearch
from hunyuan import hunyuan_gpt

es_client = Elasticsearch(
    "ES 集群访问地址",basic_auth=("用户名", "密码"))

def get_elasticsearch_results(query):
    es_query = {
        "knn": {
            "field": "content_embedding",
            "num_candidates": 100,
            "query_vector_builder": {
                "text_embedding": {
                    "model_id": ".multilingual-e5-small_linux-x86_64",
                    "model_text": query
                }
            }
        },
        "query":{
            "match":{
                "content":query
            }
        },
        "rank":{
            "rrf":{
                "window_size":100,
                "rank_constant":20
            }
        }
    }
    result = es_client.search(index="index-name", body=es_query)
    return result["hits"]["hits"]

def create_hunyuan_prompt(results,question):
    context = ""
    for hit in results:
        source_field = hit["_source"]["content"]
        hit_context = source_field
        context += f"{hit_context}\n"
    prompt = f"""
Instructions:

回答此问题：{question}
回答时，你只能参考文档{context} 生成答案，如果无法生成答案，请回复：对不起，该问题我无法回答

"""
    return prompt

def generate_hunyuan_completion(user_prompt, question):
    response = hunyuan_gpt(user_prompt, question)
    return response.Choices[0].Message.Content

def format_result(hit):
    title = hit["_source"]["title"]
    return f'[{title}]'

def main():
    st.title("我的专属AI助手")

    # 创建输入框和查询按钮
    question = st.text_input("请输入您的问题：")
    if st.button("查询并生成答案"):
        # 获取Elasticsearch结果并创建HunYuan提示
        elasticsearch_results = get_elasticsearch_results(question)
        user_prompt = create_hunyuan_prompt(elasticsearch_results,question)
        system_prompt = f"""
你是一个问答任务的助手。使用呈现的上下文真实、实事求是地回答问题。
"""

        # 使用HunYuan模型生成回答
        openai_completion = generate_hunyuan_completion(system_prompt, user_prompt)

        # 显示结果
        st.write(openai_completion)

        # 展示Elasticsearch查询结果
        st.write("大模型参考的文档：")
        for hit in elasticsearch_results:
            st.markdown(format_result(hit))

if __name__ == "__main__":
    main()

streamlit run demo.py