RAG通常指的是"Retrieval-Augmented Generation"，即“檢索增強(qiáng)的生成”。這是一種結(jié)合了檢索（Retrieval）和生成（Generation）的機(jī)器學(xué)習(xí)模型，通常用于自然語言處理任務(wù)，如文本生成、問答系統(tǒng)等。

我們通過一下幾個步驟來完成一個基于京東云官網(wǎng)文檔的RAG系統(tǒng)

數(shù)據(jù)收集

建立知識庫

向量檢索

提示詞與模型

數(shù)據(jù)收集

數(shù)據(jù)的收集再整個RAG實(shí)施過程中無疑是最耗人工的，涉及到收集、清洗、格式化、切分等過程。這里我們使用京東云的官方文檔作為知識庫的基礎(chǔ)。文檔格式大概這樣：

{
    "content": "DDoS IP高防結(jié)合Web應(yīng)用防火墻方案說明n=======================nnnDDoS IP高防+Web應(yīng)用防火墻提供三層到七層安全防護(hù)體系，應(yīng)用場景包括游戲、金融、電商、互聯(lián)網(wǎng)、政企等京東云內(nèi)和云外的各類型用戶。nnn部署架構(gòu)n====nnn[!["部署架構(gòu)"]("https://jdcloud-portal.oss.cn-north-1.jcloudcs.com/cn/image/Advanced%20Anti-DDoS/Best-Practice02.png")]("https://jdcloud-portal.oss.cn-north-1.jcloudcs.com/cn/image/Advanced%20Anti-DDoS/Best-Practice02.png")  nnDDoS IP高防+Web應(yīng)用防火墻的最佳部署架構(gòu)如下：nnn* 京東云的安全調(diào)度中心，通過DNS解析，將用戶域名解析到DDoS IP高防CNAME。n* 用戶正常訪問流量和DDoS攻擊流量經(jīng)過DDoS IP高防清洗，回源至Web應(yīng)用防火墻。n* 攻擊者惡意請求被Web應(yīng)用防火墻過濾后返回用戶源站。n* Web應(yīng)用防火墻可以保護(hù)任何公網(wǎng)的服務(wù)器，包括但不限于京東云，其他廠商的云，IDC等nnn方案優(yōu)勢n====nnn1. 用戶源站在DDoS IP高防和Web應(yīng)用防火墻之后，起到隱藏源站IP的作用。n2. CNAME接入，配置簡單，減少運(yùn)維人員工作。nnn",
    "title": "DDoS IP高防結(jié)合Web應(yīng)用防火墻方案說明",
    "product": "DDoS IP高防",
    "url": "https://docs.jdcloud.com/cn/anti-ddos-pro/anti-ddos-pro-and-waf"
}

每條數(shù)據(jù)是一個包含四個字段的json，這四個字段分別是"content"：文檔內(nèi)容；"title"：文檔標(biāo)題；"product"：相關(guān)產(chǎn)品；"url"：文檔在線地址

向量數(shù)據(jù)庫的選擇與Retriever實(shí)現(xiàn)

向量數(shù)據(jù)庫是RAG系統(tǒng)的記憶中心。目前市面上開源的向量數(shù)據(jù)庫很多，那個向量庫比較好也是見仁見智。本項(xiàng)目中筆者選擇則了clickhouse作為向量數(shù)據(jù)庫。選擇ck主要有一下幾個方面的考慮：

ck再langchain社區(qū)的集成實(shí)現(xiàn)比較好，入庫比較平滑

向量查詢支持sql，學(xué)習(xí)成本較低，上手容易

京東云有相關(guān)產(chǎn)品且有專業(yè)團(tuán)隊(duì)支持，用著放心

文檔向量化及入庫過程

為了簡化文檔向量化和檢索過程，我們使用了longchain的Retriever工具集
首先將文檔向量化，代碼如下：

from libs.jd_doc_json_loader import JD_DOC_Loader
from langchain_community.document_loaders import DirectoryLoader

root_dir = "/root/jd_docs"
loader = DirectoryLoader(
    '/root/jd_docs', glob="**/*.json", loader_cls=JD_DOC_Loader)
docs = loader.load()

langchain 社區(qū)里并沒有提供針對特定格式的裝載器，為此，我們自定義了JD_DOC_Loader來實(shí)現(xiàn)加載過程

import json
import logging
from pathlib import Path
from typing import Iterator, Optional, Union

from langchain_core.documents import Document

from langchain_community.document_loaders.base import BaseLoader
from langchain_community.document_loaders.helpers import detect_file_encodings

logger = logging.getLogger(__name__)


class JD_DOC_Loader(BaseLoader):
    """Load text file.


    Args:
        file_path: Path to the file to load.

        encoding: File encoding to use. If `None`, the file will be loaded
        with the default system encoding.

        autodetect_encoding: Whether to try to autodetect the file encoding
            if the specified encoding fails.
    """

    def __init__(
        self,
        file_path: Union[str, Path],
        encoding: Optional[str] = None,
        autodetect_encoding: bool = False,
    ):
        """Initialize with file path."""
        self.file_path = file_path
        self.encoding = encoding
        self.autodetect_encoding = autodetect_encoding

    def lazy_load(self) -> Iterator[Document]:
        """Load from file path."""
        text = ""
        from_url = ""
        try:
            with open(self.file_path, encoding=self.encoding) as f:
                doc_data = json.load(f)
                text = doc_data["content"]
                title = doc_data["title"]
                product = doc_data["product"]
                from_url = doc_data["url"]

                # text = f.read()
        except UnicodeDecodeError as e:
            if self.autodetect_encoding:
                detected_encodings = detect_file_encodings(self.file_path)
                for encoding in detected_encodings:
                    logger.debug(f"Trying encoding: {encoding.encoding}")
                    try:
                        with open(self.file_path, encoding=encoding.encoding) as f:
                            text = f.read()
                        break
                    except UnicodeDecodeError:
                        continue
            else:
                raise RuntimeError(f"Error loading {self.file_path}") from e
        except Exception as e:
            raise RuntimeError(f"Error loading {self.file_path}") from e
        # metadata = {"source": str(self.file_path)}
        metadata = {"source": from_url, "title": title, "product": product}
        yield Document(page_content=text, metadata=metadata)

以上代碼功能主要是解析json文件，填充Document的page_content字段和metadata字段。

接下來使用langchain 的 clickhouse 向量工具集進(jìn)行文檔入庫

import langchain_community.vectorstores.clickhouse as clickhouse
from langchain.embeddings import HuggingFaceEmbeddings

model_kwargs = {"device": "cuda"}
embeddings = HuggingFaceEmbeddings(
    model_name="/root/models/moka-ai-m3e-large", model_kwargs=model_kwargs)

settings = clickhouse.ClickhouseSettings(
    table="jd_docs_m3e_with_url", username="default", password="xxxxxx", host="10.0.1.94")

docsearch = clickhouse.Clickhouse.from_documents(
    docs, embeddings, config=settings)

入庫成功后，進(jìn)行一下檢驗(yàn)

import langchain_community.vectorstores.clickhouse as clickhouse
from langchain.embeddings import HuggingFaceEmbeddings

model_kwargs = {"device": "cuda"}~~~~
embeddings = HuggingFaceEmbeddings(
    model_name="/root/models/moka-ai-m3e-large", model_kwargs=model_kwargs)

settings = clickhouse.ClickhouseSettings(
    table="jd_docs_m3e_with_url_splited", username="default", password="xxxx", host="10.0.1.94")
ck_db = clickhouse.Clickhouse(embeddings, config=settings)
ck_retriever = ck_db.as_retriever(
    search_type="similarity_score_threshold", search_kwargs={'score_threshold': 0.9})
ck_retriever.get_relevant_documents("如何創(chuàng)建mysql rds")

有了知識庫以后，可以構(gòu)建一個簡單的restful 服務(wù)，我們這里使用fastapi做這個事兒

from fastapi import FastAPI
from pydantic import BaseModel
from singleton_decorator import singleton
from langchain_community.embeddings import HuggingFaceEmbeddings
import langchain_community.vectorstores.clickhouse as clickhouse
import uvicorn
import json

app = FastAPI()
app = FastAPI(docs_url=None)
app.host = "0.0.0.0"

model_kwargs = {"device": "cuda"}
embeddings = HuggingFaceEmbeddings(
    model_name="/root/models/moka-ai-m3e-large", model_kwargs=model_kwargs)
settings = clickhouse.ClickhouseSettings(
    table="jd_docs_m3e_with_url_splited", username="default", password="xxxx", host="10.0.1.94")
ck_db = clickhouse.Clickhouse(embeddings, config=settings)
ck_retriever = ck_db.as_retriever(
    search_type="similarity", search_kwargs={"k": 3})


class question(BaseModel):
    content: str


@app.get("/")
async def root():
    return {"ok"}


@app.post("/retriever")
async def retriver(question: question):
    global ck_retriever
    result = ck_retriever.invoke(question.content)
    return result


if __name__ == '__main__':
    uvicorn.run(app='retriever_api:app', host="0.0.0.0",
                port=8000, reload=True)

返回結(jié)構(gòu)大概這樣：

[
  {
    "page_content": "云緩存 Redis--Redis遷移解決方案n###RedisSyncer 操作步驟n####數(shù)據(jù)校驗(yàn)n```nwget   https://github.com/TraceNature/rediscompare/releases/download/v1.0.0/rediscompare-1.0.0-linux-amd64.tar.gznrediscompare compare single2single --saddr "10.0.1.101:6479" --spassword "redistest0102" --taddr "10.0.1.102:6479" --tpassword  "redistest0102" --comparetimes 3nn```  n**Github 地址：** [https://github.com/TraceNature/redissyncer-server]("https://github.com/TraceNature/redissyncer-server")",
    "metadata": {
      "product": "云緩存 Redis",
      "source": "https://docs.jdcloud.com/cn/jcs-for-redis/doc-2",
      "title": "Redis遷移解決方案"
    },
    "type": "Document"
  },
  {
    "page_content": "云緩存 Redis--Redis遷移解決方案n###RedisSyncer 操作步驟n####數(shù)據(jù)校驗(yàn)n```nwget   https://github.com/TraceNature/rediscompare/releases/download/v1.0.0/rediscompare-1.0.0-linux-amd64.tar.gznrediscompare compare single2single --saddr "10.0.1.101:6479" --spassword "redistest0102" --taddr "10.0.1.102:6479" --tpassword  "redistest0102" --comparetimes 3nn```  n**Github 地址：** [https://github.com/TraceNature/redissyncer-server]("https://github.com/TraceNature/redissyncer-server")",
    "metadata": {
      "product": "云緩存 Redis",
      "source": "https://docs.jdcloud.com/cn/jcs-for-redis/doc-2",
      "title": "Redis遷移解決方案"
    },
    "type": "Document"
  },
  {
    "page_content": "云緩存 Redis--Redis遷移解決方案n###RedisSyncer 操作步驟n####數(shù)據(jù)校驗(yàn)n```nwget   https://github.com/TraceNature/rediscompare/releases/download/v1.0.0/rediscompare-1.0.0-linux-amd64.tar.gznrediscompare compare single2single --saddr "10.0.1.101:6479" --spassword "redistest0102" --taddr "10.0.1.102:6479" --tpassword  "redistest0102" --comparetimes 3nn```  n**Github 地址：** [https://github.com/TraceNature/redissyncer-server]("https://github.com/TraceNature/redissyncer-server")",
    "metadata": {
      "product": "云緩存 Redis",
      "source": "https://docs.jdcloud.com/cn/jcs-for-redis/doc-2",
      "title": "Redis遷移解決方案"
    },
    "type": "Document"
  }
]

返回一個向量距離最小的list

結(jié)合模型和prompt，回答問題

為了節(jié)約算力資源，我們選擇qwen 1.8B模型，一張v100卡剛好可以容納一個qwen模型和一個m3e-large embedding 模型

answer 服務(wù)

from fastapi import FastAPI
from pydantic import BaseModel
from langchain_community.llms import VLLM
from transformers import AutoTokenizer
from langchain.prompts import PromptTemplate
import requests
import uvicorn
import json
import logging

app = FastAPI()
app = FastAPI(docs_url=None)
app.host = "0.0.0.0"

logger = logging.getLogger()
logger.setLevel(logging.INFO)
to_console = logging.StreamHandler()
logger.addHandler(to_console)


# load model
# model_name = "/root/models/Llama3-Chinese-8B-Instruct"
model_name = "/root/models/Qwen1.5-1.8B-Chat"
tokenizer = AutoTokenizer.from_pretrained(model_name)
llm_llama3 = VLLM(
    model=model_name,
    tokenizer=tokenizer,
    task="text-generation",
    temperature=0.2,
    do_sample=True,
    repetition_penalty=1.1,
    return_full_text=False,
    max_new_tokens=900,
)

# prompt
prompt_template = """
你是一個云技術(shù)專家
使用以下檢索到的Context回答問題。
如果不知道答案，就說不知道。
用中文回答問題。
Question: {question}
Context: {context}
Answer: 
"""

prompt = PromptTemplate(
    input_variables=["context", "question"],
    template=prompt_template,
)


def get_context_list(q: str):
    url = "http://10.0.0.7:8000/retriever"
    payload = {"content": q}
    res = requests.post(url, json=payload)
    return res.text


class question(BaseModel):
    content: str


@app.get("/")
async def root():
    return {"ok"}


@app.post("/answer")
async def answer(q: question):
    logger.info("invoke!!!")
    global prompt
    global llm_llama3
    context_list_str = get_context_list(q.content)

    context_list = json.loads(context_list_str)
    context = ""
    source_list = []

    for context_json in context_list:
        context = context+context_json["page_content"]
        source_list.append(context_json["metadata"]["source"])
    p = prompt.format(context=context, question=q.content)
    answer = llm_llama3(p)
    result = {
        "answer": answer,
        "sources": source_list
    }
    return result


if __name__ == '__main__':
    uvicorn.run(app='retriever_api:app', host="0.0.0.0",
                port=8888, reload=True)

代碼通過使用Retriever接口查找與問題相似的文檔，作為context組合prompt推送給模型生成答案。
主要服務(wù)就緒后可以開始畫一張臉了，使用gradio做個簡易對話界面

gradio 服務(wù)

import json
import gradio as gr
import requests


def greet(name, intensity):
    return "Hello, " + name + "!" * int(intensity)


def answer(question):
    url = "http://127.0.0.1:8888/answer"
    payload = {"content": question}
    res = requests.post(url, json=payload)
    res_json = json.loads(res.text)
    return [res_json["answer"], res_json["sources"]]


demo = gr.Interface(
    fn=answer,
    # inputs=["text", "slider"],
    inputs=[gr.Textbox(label="question", lines=5)],
    # outputs=[gr.TextArea(label="answer", lines=5),
    #          gr.JSON(label="urls", value=list)]
    outputs=[gr.Markdown(label="answer"),
             gr.JSON(label="urls", value=list)]
)


demo.launch(server_name="0.0.0.0")


審核編輯 黃宇