{"id":1303,"date":"2024-05-02T07:36:53","date_gmt":"2024-05-02T07:36:53","guid":{"rendered":"https:\/\/www.nicekj.com\/?p=1303"},"modified":"2024-05-02T07:38:28","modified_gmt":"2024-05-02T07:38:28","slug":"jinghuabijiwuendaxlangchainshiyonglangchaingoujianyushujuduihuadeliaotianjiqirenxia","status":"publish","type":"post","link":"https:\/\/www.nicekj.com\/jinghuabijiwuendaxlangchainshiyonglangchaingoujianyushujuduihuadeliaotianjiqirenxia.html","title":{"rendered":"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe x LangChain \u300a\u4f7f\u7528LangChain\u6784\u5efa\u4e0e\u6570\u636e\u5bf9\u8bdd\u7684\u804a\u5929\u673a\u5668\u4eba\u300b\uff08\u4e0b\uff09"},"content":{"rendered":"

\u95ee\u7b54<\/h1>\n

\n

\"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe<\/figure>\n<\/p>\n

\u5728\u8fd9\u8282\u8bfe\u4e2d\uff0c\u6211\u4eec\u5c06\u5b66\u4e60\u5982\u4f55\u5229\u7528\u68c0\u7d22\u5230\u7684\u6587\u6863\u6765\u56de\u7b54\u7528\u6237\u7684\u95ee\u9898\u3002<\/p>\n

\n

\"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe<\/figure>\n<\/p>\n

\u6574\u4e2a\u8fc7\u7a0b\u53ef\u4ee5\u62c6\u5206\u4e3a\u4ee5\u4e0b\u51e0\u4e2a\u6b65\u9aa4\uff1a<\/p>\n

    \n
  1. \u7528\u6237\u8f93\u5165\u4e00\u4e2a\u95ee\u9898\uff08Question\uff09<\/li>\n
  2. \u4ece\u5411\u91cf\u5b58\u50a8\uff08Store\uff09\u4e2d\u68c0\u7d22\u51fa\u4e0e\u95ee\u9898\u76f8\u5173\u7684\u6587\u6863\u5206\u5757\uff08Relavant splits\uff09<\/li>\n
  3. \u5c06\u8fd9\u4e9b\u5206\u5757\u8fde\u540c\u7cfb\u7edf\u63d0\u793a\uff08System:Prompt\uff09\u548c\u7528\u6237\u95ee\u9898\uff08Human:Question\uff09\u4e00\u8d77\u4f5c\u4e3a\u8f93\u5165\u4f20\u7ed9\u8bed\u8a00\u6a21\u578b\uff08LLM\uff09<\/li>\n
  4. \u8bed\u8a00\u6a21\u578b\u6839\u636e\u8f93\u5165\u751f\u6210\u7b54\u6848\uff08Answer\uff09<\/li>\n<\/ol>\n

    \u9ed8\u8ba4\u4f7f\u7528\u7684\u662fstuff<\/code>\u65b9\u6cd5\uff0c\u5176\u7279\u70b9\u5982\u4e0b\uff1a<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
    \u7279\u70b9<\/th>\u4f18\u70b9<\/th>\u7f3a\u70b9<\/th><\/tr><\/thead>
    \u5c06\u6240\u6709\u68c0\u7d22\u5230\u7684\u5206\u5757\u653e\u5165\u540c\u4e00\u4e2a\u4e0a\u4e0b\u6587\u7a97\u53e3\u4e2d\uff0c\u53ea\u9700\u8981\u5bf9\u8bed\u8a00\u6a21\u578b\u8fdb\u884c\u4e00\u6b21\u8c03\u7528\u3002<\/td>\u7b80\u5355\u3001\u5ec9\u4ef7\u4e14\u6548\u679c\u4e0d\u9519\u3002<\/td>\u5f53\u68c0\u7d22\u5230\u7684\u6587\u6863\u8fc7\u591a\u65f6\uff0c\u7531\u4e8e\u4e0a\u4e0b\u6587\u7a97\u53e3\u957f\u5ea6\u6709\u9650\uff0c\u53ef\u80fd\u65e0\u6cd5\u5c06\u6240\u6709\u5206\u5757\u90fd\u4f20\u5165\u3002<\/td><\/tr><\/tbody><\/table>\n

    \n

    \"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe<\/figure>\n<\/p>\n

    \u4e3a\u4e86\u89e3\u51b3\u4e0a\u4e0b\u6587\u7a97\u53e3\u957f\u5ea6\u9650\u5236\u7684\u95ee\u9898\uff0c\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528Map-reduce<\/code>\u3001Refine<\/code>\u548cMap-rerank<\/code>\u4e09\u79cd\u65b9\u6cd5\uff0c\u8fd9\u4e9b\u65b9\u6cd5\u6211\u4eec\u5728\u4e4b\u524d\u7684\u8bfe\u7a0b\u4e2d\u5df2\u7ecf\u7b80\u8981\u4ecb\u7ecd\u8fc7\u4e86\uff0c\u4eca\u5929\u6211\u4eec\u5c06\u8fdb\u4e00\u6b65\u6df1\u5165\u4e86\u89e3\u3002<\/p>\n

    stuff<\/h2>\n

    \u6b65\u9aa41\uff1a\u52a0\u8f7d\u4e4b\u524d\u4fdd\u5b58\u7684\u5411\u91cf\u6570\u636e\u5e93<\/h3>\n
    <\/div>
    ini<\/span>
    \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>from langchain.vectorstores import Chroma<\/span>\nfrom langchain.embeddings.openai import OpenAIEmbeddings<\/span>\npersist_directory<\/span> = 'docs\/chroma\/'<\/span><\/span>\nembedding<\/span> = OpenAIEmbeddings()<\/span>\nvectordb<\/span> = Chroma(persist_directory=persist_directory, embedding_function=embedding)<\/span>\n<\/code><\/pre>\n
    \u6b65\u9aa42\uff1a\u521d\u59cb\u5316\u5c06\u7528\u4e8e\u56de\u7b54\u95ee\u9898\u7684\u8bed\u8a00\u6a21\u578b<\/h3>\n
    <\/div>
    ini<\/span>
    \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>llm_name<\/span> = \"gpt-3.5-turbo\"<\/span><\/span>\nfrom langchain.chat_models import ChatOpenAI<\/span>\nllm<\/span> = ChatOpenAI(model_name=llm_name, temperature=0<\/span>)<\/span>\n<\/code><\/pre>\n
    temperature\u53c2\u6570\u8bbe\u7f6e\u4e3a0\uff0c\u53ef\u4ee5\u5e2e\u52a9\u6211\u4eec\u5f97\u5230\u66f4\u51c6\u786e\u7684\u7b54\u6848\uff0c\u56e0\u4e3a\u5b83\u964d\u4f4e\u4e86\u8bed\u8a00\u6a21\u578b\u7684\u53ef\u53d8\u6027\uff0c\u901a\u5e38\u80fd\u7ed9\u6211\u4eec\u6700\u9ad8\u7f6e\u4fe1\u5ea6\u3001\u6700\u53ef\u9760\u7684\u7b54\u6848\u3002<\/p>\n
    \u6b65\u9aa43\uff1a\u5bfc\u5165\u3001\u521b\u5efa\u3001\u8c03\u7528\u68c0\u7d22\u95ee\u7b54\u94fe\uff0c\u8f93\u5165\u95ee\u9898\uff0c\u5e76\u83b7\u53d6\u7b54\u6848<\/h3>\n
    <\/div>
    ini<\/span>
    \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>question<\/span> = \"What are major topics for this class?\"<\/span><\/span>\n<\/span>\nfrom langchain.chains import RetrievalQA<\/span>\nqa_chain<\/span> = RetrievalQA.from_chain_type(<\/span>\n    llm,<\/span>\n    retriever<\/span>=vectordb.as_retriever()<\/span>\n)<\/span>\nresult<\/span> = qa_chain({\"query\"<\/span>: question})<\/span>\nresult[\"result\"]<\/span><\/span>\n<\/code><\/pre>\n
    \n
    ‘The major topic for this class is machine learning. Additionally, the class may cover statistics and algebra as refreshers in the discussion sections. Later in the quarter, the discussion sections will also cover extensions for the material taught in the main lectures.’<\/p>\n<\/blockquote>\n
    \u6b65\u9aa44\uff1a\u4f7f\u7528\u63d0\u793a\u6a21\u677f\u4f18\u5316\u8f93\u51fa\u7ed3\u679c<\/h3>\n
    \u63d0\u793a\u6a21\u677f\u662f\u4e00\u79cd\u53ef\u4ee5\u5e2e\u52a9\u8bed\u8a00\u6a21\u578b\u751f\u6210\u66f4\u7b26\u5408\u8981\u6c42\u7684\u8f93\u51fa\u7ed3\u679c\u7684\u6280\u5de7\uff0c\u6211\u4eec\u5728\u4e0a\u4e00\u95e8\u8bfe<\/a>\u4e2d\u5df2\u7ecf\u4ecb\u7ecd\u8fc7\u4e86\u3002\u8fd9\u91cc\u6211\u4eec\u4f7f\u7528\u7684\u63d0\u793a\u6a21\u677f\u4e3b\u8981\u662f\u4e3a\u4e86\u8ba9\u8f93\u51fa\u7ed3\u679c\u66f4\u7b80\u6d01\u3001\u66f4\u5c11\u7f16\u9020\u3001\u66f4\u793c\u8c8c\u3002<\/p>\n
    <\/div>
    ini<\/span>
    \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>from langchain.prompts import PromptTemplate<\/span>\n<\/span>\n# \u6784\u5efa\u63d0\u793a\u8bcd<\/span><\/span>\n# {context}\uff1a\u4e0a\u4e0b\u6587\u5360\u4f4d\u7b26\uff0c\u7528\u4e8e\u653e\u7f6e\u6587\u6863\u5185\u5bb9<\/span><\/span>\n# {question}\uff1a\u95ee\u9898\u5360\u4f4d\u7b26\uff0c\u653e\u7f6e\u8981\u67e5\u8be2\u7684\u95ee\u9898<\/span><\/span>\ntemplate<\/span> = \"\"\"Use the following pieces of context to answer the question at the end. If you don't know the answer, just say that you don't know, don't try to make up an answer. Use three sentences maximum. Keep the answer as concise as possible. Always say \"thanks for asking!\" at the end of the answer. <\/span><\/span>\n{context}<\/span>\nQuestion: {question}<\/span>\nHelpful Answer:\"\"\"<\/span>\nQA_CHAIN_PROMPT<\/span> = PromptTemplate.from_template(template)<\/span>\n<\/span>\n# \u8fd0\u884c\u94fe<\/span><\/span>\n# return_source_documents=True\u7528\u4e8e\u652f\u6301\u67e5\u770b\u68c0\u7d22\u5230\u7684\u6587\u6863<\/span><\/span>\nqa_chain<\/span> = RetrievalQA.from_chain_type(<\/span>\n    llm,<\/span>\n    retriever<\/span>=vectordb.as_retriever(),<\/span>\n    return_source_documents<\/span>=True<\/span>,<\/span>\n    chain_type_kwargs<\/span>={\"prompt\"<\/span>: QA_CHAIN_PROMPT}<\/span>\n)<\/span>\n<\/span>\nquestion<\/span> = \"Is probability a class topic?\"<\/span><\/span>\nresult<\/span> = qa_chain({\"query\"<\/span>: question})<\/span>\nresult[\"result\"]<\/span><\/span>\n<\/code><\/pre>\n
    \n
    ‘Yes, probability is assumed to be a prerequisite for this class. The instructor assumes familiarity with basic probability and statistics, and will go over some of the prerequisites in the discussion sections as a refresher course. Thanks for asking!’<\/p>\n<\/blockquote>\n
    \u6b65\u9aa45\uff1a\u67e5\u770b\u8fd4\u56de\u7684\u6e90\u6587\u6863\uff0c\u7406\u89e3\u5176\u4ece\u54ea\u91cc\u83b7\u53d6\u6570\u636e<\/h3>\n
    <\/div>
    css<\/span>
    \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>result[\"source_documents\"<\/span>]<\/span>[0]<\/span><\/span>\n<\/code><\/pre>\n
    \n
    Document(page_content=”of this class will not be very program ming intensive, although we will do some nprogramming, mostly in either MATLAB or Octa ve. I’ll say a bit more about that later.  nI also assume familiarity with basic proba bility and statistics. So most undergraduate nstatistics class, like Stat 116 taught here at Stanford, will be more than enough. I’m gonna nassume all of you know what ra ndom variables are, that all of you know what expectation nis, what a variance or a random variable is. And in case of some of you, it’s been a while nsince you’ve seen some of this material. At some of the discussion sections, we’ll actually ngo over some of the prerequisites, sort of as  a refresher course under prerequisite class. nI’ll say a bit more about that later as well.  nLastly, I also assume familiarity with basi c linear algebra. And again, most undergraduate nlinear algebra courses are more than enough. So if you’ve taken courses like Math 51, n103, Math 113 or CS205 at Stanford, that would be more than enough. Basically, I’m ngonna assume that all of you know what matrix es and vectors are, that you know how to nmultiply matrices and vectors and multiply matrix and matrices, that you know what a matrix inverse is. If you know what an eigenvect or of a matrix is, that’d be even better. nBut if you don’t quite know or if you’re not qu ite sure, that’s fine, too. We’ll go over it in nthe review sections.”, metadata={‘source’: ‘docs\/cs229_lectures\/MachineLearning-Lecture01.pdf’, ‘page’: 4})<\/p>\n<\/blockquote>\n
    Map-reduce<\/h2>\n
    \n
    \"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe<\/figure>\n<\/p>\n
    Map-reduce\u65b9\u6cd5\u7684\u7279\u70b9\u5982\u4e0b\uff1a<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
    \u7279\u70b9<\/th>\u4f18\u70b9<\/th>\u7f3a\u70b9<\/th><\/tr><\/thead>
    1.\u5c06\u6bcf\u4e2a\u6587\u6863\u5355\u72ec\u53d1\u9001\u5230\u8bed\u8a00\u6a21\u578b\u4e2d\uff0c\u6839\u636e\u5355\u4e2a\u6587\u6863\u751f\u6210\u7b54\u6848\uff1b<\/td>\u53ef\u4ee5\u5904\u7406\u4efb\u610f\u6570\u91cf\u7684\u6587\u6863\u3002<\/td>1.\u6d89\u53ca\u5230\u5bf9\u8bed\u8a00\u6a21\u578b\u7684\u591a\u6b21\u8c03\u7528\uff0c\u901f\u5ea6\u8f83\u6162\uff1b<\/td><\/tr>
    2.\u5c06\u6240\u6709\u8fd9\u4e9b\u7b54\u6848\u7ec4\u5408\u5728\u4e00\u8d77\uff0c\u518d\u8c03\u7528\u8bed\u8a00\u6a21\u578b\u751f\u6210\u6700\u7ec8\u7b54\u6848\u3002<\/td><\/td>2.\u4fe1\u606f\u53ef\u80fd\u5206\u6563\u5728\u4e0d\u540c\u7684\u6587\u6863\u4e2d\uff0c\u65e0\u6cd5\u57fa\u4e8e\u540c\u4e00\u4e2a\u4e0a\u4e0b\u6587\u83b7\u53d6\u4fe1\u606f\uff0c\u7ed3\u679c\u53ef\u80fd\u4e0d\u51c6\u786e\u3002<\/td><\/tr><\/tbody><\/table>\n
    <\/div>
    ini<\/span>
    \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>qa_chain_mr<\/span> = RetrievalQA.from_chain_type(<\/span>\n    llm,<\/span>\n    retriever<\/span>=vectordb.as_retriever(),<\/span>\n    chain_type<\/span>=\"map_reduce\"<\/span><\/span>\n)<\/span>\n<\/span>\nresult<\/span> = qa_chain_mr({\"query\"<\/span>: question})<\/span>\nresult[\"result\"]<\/span><\/span>\n<\/code><\/pre>\n
    \n
    ‘There is no clear answer to this question based on the given portion of the document. The document mentions familiarity with basic probability and statistics as a prerequisite for the class, and there is a brief mention of probability in the text, but it is not clear if it is a main topic of the class.’<\/p>\n<\/blockquote>\n
    \u4f7f\u7528LangSmith\u5e73\u53f0\u4e86\u89e3\u8fd9\u4e9b\u94fe\u5185\u90e8\u7684\u8c03\u7528\u60c5\u51b5<\/h3>\n
    LangSmith \u662f\u4e00\u4e2a\u7528\u4e8e\u6784\u5efa\u751f\u4ea7\u7ea7 LLM \u5e94\u7528\u7a0b\u5e8f\u7684\u5e73\u53f0\u3002<\/p>\n
    \u5b83\u53ef\u4ee5\u8ba9\u60a8\u8f7b\u677e\u5730\u8c03\u8bd5\u3001\u6d4b\u8bd5\u3001\u8bc4\u4f30\u548c\u76d1\u63a7\u57fa\u4e8e\u4efb\u4f55 LLM \u6846\u67b6\u6784\u5efa\u7684\u94fe\u548c\u667a\u80fd\u4ee3\u7406\uff0c\u5e76\u4e0e\u4f7f\u7528 LLM \u6784\u5efa\u7684\u5f00\u6e90\u6846\u67b6 LangChain \u5b8c\u7f8e\u96c6\u6210\u3002<\/p>\n
    \u8981\u4f53\u9a8c\u8fd9\u4e2a\u5e73\u53f0\u7684\u529f\u80fd\uff0c\u4f60\u9700\u8981\uff1a<\/p>\n
      \n
    1. \u524d\u5f80LangSmith\u5e73\u53f0<\/a>\u6ce8\u518c\uff08\u53ef\u80fd\u9700\u8981\u6392\u961f\uff09<\/li>\n
    2. \u521b\u5efa API \u5bc6\u94a5<\/li>\n
    3. \u5728\u4ee5\u4e0b\u4ee3\u7801\u4e2d\u4f7f\u7528\u8fd9\u4e2a API \u5bc6\u94a5<\/li>\n
    4. \u53d6\u6d88\u6ce8\u91ca\u4ee5\u4e0b\u4ee3\u7801\uff0c\u5e76\u91cd\u65b0\u8fd0\u884cMapReduce\u94fe<\/li>\n<\/ol>\n
      <\/div>
      lua<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>import os<\/span><\/span>\nos<\/span>.environ[\"LANGCHAIN_TRACING_V2\"<\/span>] = \"true\"<\/span><\/span>\nos<\/span>.environ[\"LANGCHAIN_ENDPOINT\"<\/span>] = \"https:\/\/api.langchain.plus\"<\/span><\/span>\nos<\/span>.environ[\"LANGCHAIN_API_KEY\"<\/span>] = \"...\"<\/span> # \u66ff\u6362...\u4e3a\u4f60\u7684 API \u5bc6\u94a5<\/span>\n<\/code><\/pre>\n
      \u4e4b\u540e\uff0c\u5c31\u53ef\u4ee5\u5207\u6362\u5230LangSmith\u5e73\u53f0\uff0c\u627e\u5230\u521a\u521a\u8fd0\u884c\u7684RetrievalQA\uff0c\u67e5\u770b\u8f93\u5165\u3001\u8f93\u51fa\u4ee5\u53ca\u8c03\u7528\u94fe\u4e86\uff1a<\/p>\n
      \n
      \"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe<\/figure>\n<\/p>\n
      \u53ef\u4ee5\u770b\u5230\uff0cMapReduceDocumentChain\u4e2d\u6d89\u53ca\u5230\u4e86\u5bf9\u8bed\u8a00\u6a21\u578b\u7684\u56db\u6b21\u72ec\u7acb\u8c03\u7528\uff0c\u70b9\u51fb\u5176\u4e2d\u4e00\u4e2a\uff0c\u5c31\u53ef\u4ee5\u770b\u5230\u6bcf\u4e2a\u6587\u6863\u7684\u5177\u4f53\u8f93\u5165\u548c\u8f93\u51fa\uff1a<\/p>\n
      \n
      \"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe<\/figure>\n<\/p>\n
      \u5e76\u4e14\uff0c\u53ef\u4ee5\u770b\u5230\uff0c\u5b83\u4eec\u5728\u6700\u540e\u7684\u94fe\u4e2d\u88ab\u5408\u5e76\u4e3a\u4e86StuffDocumentChain\uff0c\u4e5f\u5373\u628a\u6240\u6709\u7ed3\u679c\u653e\u5230\u4e86\u6700\u7ec8\u8c03\u7528\u4e2d\u3002\u70b9\u51fb\u8fdb\u5165\u5c31\u53ef\u4ee5\u770b\u5230\uff0c\u7cfb\u7edf\u6d88\u606f\u4e2d\u5305\u542b\u4e86\u6765\u81ea\u524d\u9762\u6587\u6863\u7684\u56db\u4e2a\u6458\u8981\uff1a<\/p>\n
      \n
      \"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe<\/figure>\n<\/p>\n
      Refine<\/h2>\n
      \n
      \"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe<\/figure>\n<\/p>\n
      Refine\u65b9\u6cd5\u7684\u7279\u70b9\u5982\u4e0b\uff1a<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
      \u7279\u70b9<\/th>\u4f18\u70b9<\/th>\u7f3a\u70b9<\/th><\/tr><\/thead>
      \u8fed\u4ee3\u5730\u5904\u7406\u591a\u4e2a\u6587\u6863\uff0c\u57fa\u4e8e\u524d\u4e00\u4e2a\u6587\u6863\u7684\u7b54\u6848\u6765\u6784\u5efa\u4e00\u4e2a\u66f4\u597d\u7684\u7b54\u6848\u3002<\/td>\u5141\u8bb8\u7ec4\u5408\u4fe1\u606f\uff0c\u66f4\u9f13\u52b1\u4fe1\u606f\u7684\u4f20\u9012<\/td>\u901f\u5ea6\u8f83\u6162<\/td><\/tr><\/tbody><\/table>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>qa_chain_mr<\/span> = RetrievalQA.from_chain_type(<\/span>\n    llm,<\/span>\n    retriever<\/span>=vectordb.as_retriever(),<\/span>\n    chain_type<\/span>=\"refine\"<\/span><\/span>\n)<\/span>\nresult<\/span> = qa_chain_mr({\"query\"<\/span>: question})<\/span>\nresult[\"result\"]<\/span><\/span>\n<\/code><\/pre>\n
      \n
      “The main topic of the class is machine learning algorithms, including linear regression and classification. Basic probability and statistics, as well as linear algebra, are prerequisites for the class, but the instructor will provide a refresher course on these topics in some of the discussion sections. Later in the quarter, the discussion sections will also cover extensions for the material taught in the main lectures. The instructor will focus on a few important extensions that there wasn’t enough time to cover in the main lectures.”<\/p>\n<\/blockquote>\n
      \u73b0\u5728\u8fd8\u6709\u4e00\u4e2a\u95ee\u9898\uff0c\u6211\u4eec\u76ee\u524d\u4f7f\u7528\u7684\u94fe\u662f\u6ca1\u6709\u201c\u8bb0\u5fc6\u201d\u8fd9\u4e2a\u6982\u5ff5\u7684\uff0c\u8fd9\u5c31\u5bfc\u81f4\u4e86\u5b83\u4e0d\u4f1a\u8bb0\u4f4f\u4e4b\u524d\u7684\u95ee\u9898\u6216\u7b54\u6848\u3002\u4e3a\u4e86\u89e3\u51b3\u8fd9\u4e2a\u95ee\u9898\uff0c\u6211\u4eec\u9700\u8981\u5f15\u5165\u201c\u8bb0\u5fc6\u201d\u529f\u80fd\uff0c\u8fd9\u5c31\u662f\u6211\u4eec\u4e0b\u4e00\u8282\u8981\u8bb2\u7684\u5185\u5bb9\u3002<\/p>\n
      Chat<\/h1>\n
      \n
      \"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe<\/figure>\n<\/p>\n
      \u5728\u8fd9\u8282\u8bfe\u4e2d\uff0c\u6211\u4eec\u5c06\u6784\u5efa\u4e00\u4e2a\u5b8c\u6574\u7684\u95ee\u7b54\u804a\u5929\u673a\u5668\u4eba\uff0c\u5b83\u5c06\u7ed3\u5408\u6211\u4eec\u4e4b\u524d\u8bb2\u8fc7\u7684\u6240\u6709\u7ec4\u4ef6\uff0c\u5e76\u5f15\u5165\u201c\u804a\u5929\u5386\u53f2\u201d\u8fd9\u4e2a\u6982\u5ff5\uff0c\u8ba9\u5b83\u5728\u56de\u7b54\u95ee\u9898\u65f6\u80fd\u591f\u8003\u8651\u5230\u4e4b\u524d\u7684\u5bf9\u8bdd\u6216\u4fe1\u606f\uff0c\u4e5f\u5c31\u662f\u8bf4\uff0c\u5b83\u80fd\u8bb0\u4f4f\u4f60\u4e4b\u524d\u8bf4\u8fc7\u4ec0\u4e48\u3002<\/p>\n
      \u6b65\u9aa41\uff1a\u521d\u59cb\u5316\u7528\u4e8e\u4fdd\u5b58\u5927\u91cf\u6587\u6863\u5185\u5bb9\u7684\u5411\u91cf\u6570\u636e\u5e93<\/h3>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>from langchain.vectorstores import Chroma<\/span>\nfrom langchain.embeddings.openai import OpenAIEmbeddings<\/span>\npersist_directory<\/span> = 'docs\/chroma\/'<\/span><\/span>\nembedding<\/span> = OpenAIEmbeddings()<\/span>\nvectordb<\/span> = Chroma(persist_directory=persist_directory, embedding_function=embedding)<\/span>\n<\/code><\/pre>\n
      \u6b65\u9aa42\uff1a\u521d\u59cb\u5316\u5c06\u4f5c\u4e3a\u804a\u5929\u673a\u5668\u4eba\u4f7f\u7528\u7684\u8bed\u8a00\u6a21\u578b<\/h3>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>llm_name<\/span> = \"gpt-3.5-turbo\"<\/span><\/span>\nfrom langchain.chat_models import ChatOpenAI<\/span>\nllm<\/span> = ChatOpenAI(model_name=llm_name, temperature=0<\/span>)<\/span>\n<\/code><\/pre>\n
      \u6b65\u9aa43\uff1a\u521d\u59cb\u5316\u63d0\u793a\u6a21\u677f\uff0c\u8ba9\u8f93\u51fa\u7ed3\u679c\u66f4\u7b80\u4ecb\u3001\u66f4\u5c11\u7f16\u9020\u3001\u66f4\u793c\u8c8c<\/h3>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div># \u6784\u5efa\u63d0\u793a<\/span><\/span>\nfrom langchain.prompts import PromptTemplate<\/span>\ntemplate<\/span> = \"\"\"Use the following pieces of context to answer the question at the end. If you don't know the answer, just say that you don't know, don't try to make up an answer. Use three sentences maximum. Keep the answer as concise as possible. Always say \"thanks for asking!\" at the end of the answer. <\/span><\/span>\n{context}<\/span>\nQuestion: {question}<\/span>\nHelpful Answer:\"\"\"<\/span>\nQA_CHAIN_PROMPT<\/span> = PromptTemplate(input_variables=[\"context\"<\/span>, \"question\"<\/span>],template=template,)<\/span>\n<\/span>\n<\/code><\/pre>\n
      \u6b65\u9aa44\uff1a\u521b\u5efa\u68c0\u7d22QA\u94fe\uff0c\u7528\u4e8e\u5408\u5e76\u68c0\u7d22\u5230\u7684\u6587\u672c\u7247\u6bb5\u5e76\u8c03\u7528\u8bed\u8a00\u6a21\u578b<\/h3>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div># \u8fd0\u884c\u94fe<\/span><\/span>\nfrom langchain.chains import RetrievalQA<\/span>\nqa_chain<\/span> = RetrievalQA.from_chain_type(llm,<\/span>\n                                       retriever<\/span>=vectordb.as_retriever(),<\/span>\n                                       return_source_documents<\/span>=True<\/span>,<\/span>\n                                       chain_type_kwargs<\/span>={\"prompt\"<\/span>: QA_CHAIN_PROMPT})<\/span>\n<\/code><\/pre>\n
      \u6b65\u9aa45\uff1a\u4f7f\u7528ConversationBufferMemory\u589e\u52a0\u804a\u5929\u673a\u5668\u4eba\u7684\u8bb0\u5fc6\u529f\u80fd<\/h3>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>from langchain.memory import ConversationBufferMemory<\/span>\nmemory<\/span> = ConversationBufferMemory(<\/span>\n    memory_key<\/span>=\"chat_history\"<\/span>,<\/span>\n    return_messages<\/span>=True<\/span><\/span>\n)<\/span>\n<\/code><\/pre>\n
      ConversationBufferMemory\u63d0\u4f9b\u4e86\u4e00\u4e2a\u804a\u5929\u6d88\u606f\u5386\u53f2\u7684\u8bb0\u5fc6\u7f13\u51b2\u533a\uff0c\u5e76\u4e14\u6bcf\u6b21\u90fd\u4f1a\u5c06\u8fd9\u90e8\u5206\u5386\u53f2\u6d88\u606f\u4f20\u5165\u804a\u5929\u673a\u5668\u4eba\u3002<\/p>\n
      return_messages=True\u8868\u793a\u5c06\u8fd4\u56de\u4e00\u4e2a\u5217\u8868\u7c7b\u578b\u7684\u804a\u5929\u5386\u53f2\u8bb0\u5f55\uff0c\u800c\u4e0d\u662f\u4e00\u4e2a\u5b57\u7b26\u4e32\u3002<\/p>\n
      \u6b65\u9aa46\uff1a\u521b\u5efaConversationalRetrievalChain\uff08\u5bf9\u8bdd\u68c0\u7d22\u94fe\uff09\uff0c\u4f20\u5165\u8bed\u8a00\u6a21\u578b\u3001\u68c0\u7d22\u5668\u548c\u8bb0\u5fc6\u7cfb\u7edf<\/h3>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>from langchain.chains import ConversationalRetrievalChain<\/span>\nretriever<\/span>=vectordb.as_retriever()<\/span>\nqa<\/span> = ConversationalRetrievalChain.from_llm(<\/span>\n    llm,<\/span>\n    retriever<\/span>=retriever,<\/span>\n    memory<\/span>=memory<\/span>\n)<\/span>\n<\/code><\/pre>\n
      ConversationalRetrievalChain\u4f1a\u5728RetrievalQAChain\u7684\u57fa\u7840\u4e0a\uff0c\u5c06\u804a\u5929\u5386\u53f2\u548c\u65b0\u63d0\u7684\u95ee\u9898\u6574\u5408\u6210\u4e00\u4e2a\u65b0\u7684\u72ec\u7acb\u95ee\u9898\uff0c\u4ee5\u4f20\u9012\u7ed9\u5411\u91cf\u5b58\u50a8\u5e93\uff0c\u67e5\u627e\u76f8\u5173\u6587\u6863\u3002<\/p>\n
      \u6b65\u9aa47\uff1a\u4f7f\u7528PyPDFLoader\u52a0\u8f7d\u6240\u8981\u53c2\u8003\u7684\u6587\u6863<\/h3>\n
      <\/div>
      python<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>from<\/span> langchain.embeddings.openai import<\/span> OpenAIEmbeddings<\/span>\nfrom<\/span> langchain.text_splitter import<\/span> CharacterTextSplitter, RecursiveCharacterTextSplitter<\/span>\nfrom<\/span> langchain.vectorstores import<\/span> DocArrayInMemorySearch<\/span>\nfrom<\/span> langchain.document_loaders import<\/span> TextLoader<\/span>\nfrom<\/span> langchain.chains import<\/span> RetrievalQA,  ConversationalRetrievalChain<\/span>\nfrom<\/span> langchain.memory import<\/span> ConversationBufferMemory<\/span>\nfrom<\/span> langchain.chat_models import<\/span> ChatOpenAI<\/span>\nfrom<\/span> langchain.document_loaders import<\/span> TextLoader<\/span>\nfrom<\/span> langchain.document_loaders import<\/span> PyPDFLoader<\/span>\n<\/span>\ndef<\/span> load_db<\/span>(file, chain_type, k<\/span>):<\/span>\n    # \u52a0\u8f7d\u6587\u6863<\/span><\/span>\n    loader = PyPDFLoader(file)<\/span>\n    documents = loader.load()<\/span>\n    ...<\/span>\n<\/code><\/pre>\n
      \u6b65\u9aa48\uff1a\u5206\u5272\u6587\u6863\uff0c\u4e3a\u6bcf\u4e2a\u5206\u5757\u521b\u5efa\u5d4c\u5165\uff0c\u5e76\u5b58\u50a8\u5230\u5411\u91cf\u5b58\u50a8\u5e93\u4e2d\u3002<\/h3>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>def load_db(file, chain_type, k):<\/span>\n    ...<\/span>\n    # \u5206\u9694\u6587\u6863<\/span><\/span>\n    text_splitter<\/span> = RecursiveCharacterTextSplitter(chunk_size=1000<\/span>, chunk_overlap=150<\/span>)<\/span>\n    docs<\/span> = text_splitter.split_documents(documents)<\/span>\n    # \u5b9a\u4e49\u5d4c\u5165<\/span><\/span>\n    embeddings<\/span> = OpenAIEmbeddings()<\/span>\n    # \u57fa\u4e8e\u6587\u6863\u6570\u636e\u521b\u5efa\u5411\u91cf\u6570\u636e\u5e93<\/span><\/span>\n    db<\/span> = DocArrayInMemorySearch.from_documents(docs, embeddings)<\/span>\n    ...<\/span>\n<\/code><\/pre>\n
      \u6b65\u9aa49\uff1a\u4ece\u5411\u91cf\u6570\u636e\u5e93\u521b\u5efa\u4e00\u4e2a\u57fa\u4e8e\u201c\u76f8\u4f3c\u5ea6\u201d\u7684\u68c0\u7d22\u5668\u3002<\/h3>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>def load_db(file, chain_type, k):<\/span>\n    ...<\/span>\n    # \u5b9a\u4e49\u68c0\u7d22\u5668<\/span><\/span>\n    retriever<\/span> = db.as_retriever(search_type=\"similarity\"<\/span>, search_kwargs={\"k\"<\/span>: k})<\/span>\n    ...<\/span>\n<\/code><\/pre>\n
      \u6b65\u9aa410\uff1a\u521b\u5efa\u5bf9\u8bdd\u68c0\u7d22\u94fe\uff0c\u7528\u4e8e\u5c06\u804a\u5929\u5386\u53f2\u548c\u65b0\u63d0\u7684\u95ee\u9898\u6574\u5408\u6210\u4e00\u4e2a\u65b0\u7684\u72ec\u7acb\u95ee\u9898<\/h3>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>def load_db(file, chain_type, k):<\/span>\n    ...<\/span>\n    # create a chatbot chain. Memory is managed externally.<\/span><\/span>\n    qa<\/span> = ConversationalRetrievalChain.from_llm(<\/span>\n        llm<\/span>=ChatOpenAI(model_name=llm_name, temperature=0<\/span>), <\/span>\n        chain_type<\/span>=chain_type, <\/span>\n        retriever<\/span>=retriever, <\/span>\n        return_source_documents<\/span>=True<\/span>,<\/span>\n        return_generated_question<\/span>=True<\/span>,<\/span>\n    )<\/span>\n    return qa <\/span>\n<\/code><\/pre>\n
      \u9700\u8981\u6ce8\u610f\u7684\u662f\uff0c\u8fd9\u91cc\u6211\u4eec\u5e76\u6ca1\u6709\u4f20\u5165\u8bb0\u5fc6\u7cfb\u7edf\uff0c\u800c\u662f\u5c06\u8bb0\u5fc6\u7ba1\u7406\u4ea4\u7ed9\u4e86GUI\uff0c\u8fd9\u610f\u5473\u7740\u804a\u5929\u5386\u53f2\u9700\u8981\u5728\u94fe\u4e4b\u5916\u8fdb\u884c\u7ef4\u62a4\u3002<\/p>\n
      \u6b65\u9aa411\uff1a\u63d0\u4f9b\u4e00\u4e2a\u4e0e\u804a\u5929\u673a\u5668\u4eba\u4ea4\u4e92\u7684\u7528\u6237\u754c\u9762<\/h3>\n
      <\/div>
      less<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>import<\/span> panel<\/span> as<\/span> pn<\/span><\/span>\nimport<\/span> param<\/span><\/span>\n<\/span>\nclass<\/span> cbfs<\/span>(param.Parameterized):<\/span>\n    chat_history<\/span> = param<\/span>.List<\/span>([])<\/span>\n    answer<\/span> = param<\/span>.String<\/span>(\"\"<\/span>)<\/span>\n    db_query<\/span>  = param<\/span>.String<\/span>(\"\"<\/span>)<\/span>\n    db_response<\/span> = param<\/span>.List<\/span>([])<\/span>\n    <\/span>\n    def<\/span> __init__<\/span>(self,  **params):<\/span>\n        super<\/span>(cbfs, self).__init__<\/span>( **params)<\/span>\n        self<\/span>.panels<\/span> = []<\/span><\/span>\n        self<\/span>.loaded_file<\/span> = \"docs<\/span>\/cs229_lectures<\/span>\/MachineLearning-Lecture01<\/span>.pdf<\/span>\"<\/span>\n        self<\/span>.qa<\/span> = load_db<\/span>(self.loaded_file,\"stuff\"<\/span>, 4<\/span>)<\/span>\n    <\/span>\n    def<\/span> call_load_db<\/span>(self, count):<\/span>\n        if<\/span> count<\/span> == 0<\/span> or<\/span> file_input<\/span>.value<\/span> is<\/span> None<\/span>:  # init<\/span> or<\/span> no<\/span> file<\/span> specified<\/span> :<\/span>\n            return<\/span> pn<\/span>.pane<\/span>.Markdown<\/span>(f\"Loaded File: {self.loaded_file}\"<\/span>)<\/span>\n        else<\/span>:<\/span>\n            file_input<\/span>.save<\/span>(\"temp.pdf\"<\/span>)  # local<\/span> copy<\/span><\/span>\n            self<\/span>.loaded_file<\/span> = file_input<\/span>.filename<\/span><\/span>\n            button_load<\/span>.button_style<\/span>=\"outline<\/span>\"<\/span>\n            self<\/span>.qa<\/span> = load_db<\/span>(\"temp.pdf\"<\/span>, \"stuff\"<\/span>, 4<\/span>)<\/span>\n            button_load<\/span>.button_style<\/span>=\"solid<\/span>\"<\/span>\n        self<\/span>.clr_history<\/span>()<\/span>\n        return<\/span> pn<\/span>.pane<\/span>.Markdown<\/span>(f\"Loaded File: {self.loaded_file}\"<\/span>)<\/span>\n<\/span>\n    def<\/span> convchain<\/span>(self, query):<\/span>\n        if<\/span> not<\/span> query<\/span>:<\/span>\n            return<\/span> pn<\/span>.WidgetBox<\/span>(pn.Row<\/span>('User:'<\/span>, pn.pane.Markdown<\/span>(\"\"<\/span>, width<\/span>=600<\/span>)), scroll=True)<\/span>\n        result<\/span> = self<\/span>.qa<\/span>({\"question<\/span>\": query<\/span>, \"chat_history<\/span>\": self<\/span>.chat_history<\/span>})<\/span>\n        self<\/span>.chat_history<\/span>.extend<\/span>([(query, result[\"answer\"<\/span>])])<\/span>\n        self<\/span>.db_query<\/span> = result<\/span>[\"generated_question\"<\/span>]<\/span><\/span>\n        self<\/span>.db_response<\/span> = result<\/span>[\"source_documents\"<\/span>]<\/span><\/span>\n        self<\/span>.answer<\/span> = result<\/span>['answer'<\/span>]<\/span> <\/span>\n        self<\/span>.panels<\/span>.extend<\/span>([<\/span>\n            pn.Row<\/span>('User:'<\/span>, pn.pane.Markdown<\/span>(query, width<\/span>=600<\/span>)),<\/span>\n            pn.Row<\/span>('ChatBot:'<\/span>, pn.pane.Markdown<\/span>(self.answer, width<\/span>=600<\/span>, style={'background-color'<\/span>: '#F6F6F6'<\/span>}))<\/span>\n        ])<\/span>\n        inp<\/span>.value<\/span> = ''  #clears<\/span> loading<\/span> indicator<\/span> when<\/span> cleared<\/span>\n        return pn.WidgetBox<\/span>(*self.panels,scroll=True)<\/span>\n<\/span>\n    @param<\/span>.depends<\/span>('db_query '<\/span>, )<\/span>\n    def get_lquest<\/span>(self):<\/span>\n        if not<\/span> self.db_query <\/span>:<\/span>\n            return pn.Column<\/span>(<\/span>\n                pn.Row<\/span>(pn.pane.Markdown<\/span>(f\"Last question to DB:\"<\/span>, styles={'background-color'<\/span>: '#F6F6F6'<\/span>})),<\/span>\n                pn.Row<\/span>(pn.pane.Str<\/span>(\"no DB accesses so far\"<\/span>))<\/span>\n            )<\/span>\n        return pn.Column<\/span>(<\/span>\n            pn.Row<\/span>(pn.pane.Markdown<\/span>(f\"DB query:\"<\/span>, styles={'background-color'<\/span>: '#F6F6F6'<\/span>})),<\/span>\n            pn.pane.Str<\/span>(self.db_query )<\/span>\n        )<\/span>\n<\/span>\n    @param<\/span>.depends<\/span>('db_response'<\/span>, )<\/span>\n    def get_sources<\/span>(self):<\/span>\n        if not<\/span> self.db_response<\/span>:<\/span>\n            return <\/span>\n        rlist=[pn.Row<\/span>(pn.pane.Markdown<\/span>(f\"Result of DB lookup:\"<\/span>, styles={'background-color'<\/span>: '#F6F6F6'<\/span>}))]<\/span>\n        for doc in self.db_response<\/span>:<\/span>\n            rlist.append<\/span>(pn.Row<\/span>(pn.pane.Str<\/span>(doc)))<\/span>\n        return pn.WidgetBox<\/span>(*rlist, width<\/span>=600<\/span>, scroll=True)<\/span>\n<\/span>\n    @param<\/span>.depends<\/span>('convchain'<\/span>, 'clr_history'<\/span>) <\/span>\n    def get_chats<\/span>(self):<\/span>\n        if not<\/span> self.chat_history<\/span>:<\/span>\n            return pn.WidgetBox<\/span>(pn.Row<\/span>(pn.pane.Str<\/span>(\"No History Yet\"<\/span>)), width<\/span>=600<\/span>, scroll=True)<\/span>\n        rlist=[pn.Row<\/span>(pn.pane.Markdown<\/span>(f\"Current Chat History variable\"<\/span>, styles={'background-color'<\/span>: '#F6F6F6'<\/span>}))]<\/span>\n        for exchange in self.chat_history<\/span>:<\/span>\n            rlist.append<\/span>(pn.Row<\/span>(pn.pane.Str<\/span>(exchange)))<\/span>\n        return pn.WidgetBox<\/span>(*rlist, width<\/span>=600<\/span>, scroll=True)<\/span>\n<\/span>\n    def clr_history<\/span>(self,count=0<\/span>):<\/span>\n        self.chat_history = []<\/span>\n        return <\/span>\n<\/code><\/pre>\n
      <\/div>
      ini<\/span>
      \u590d\u5236\u4ee3\u7801<\/div><\/div><\/div>cb<\/span> = cbfs()<\/span>\n<\/span>\nfile_input<\/span> = pn.widgets.FileInput(accept='.pdf'<\/span>)<\/span>\nbutton_load<\/span> = pn.widgets.Button(name=\"Load DB\"<\/span>, button_type='primary'<\/span>)<\/span>\nbutton_clearhistory<\/span> = pn.widgets.Button(name=\"Clear History\"<\/span>, button_type='warning'<\/span>)<\/span>\nbutton_clearhistory.on_click(cb.clr_history)<\/span>\ninp<\/span> = pn.widgets.TextInput( placeholder='Enter text here\u2026'<\/span>)<\/span>\n<\/span>\nbound_button_load<\/span> = pn.bind(cb.call_load_db, button_load.param.clicks)<\/span>\nconversation<\/span> = pn.bind(cb.convchain, inp) <\/span>\n<\/span>\njpg_pane<\/span> = pn.pane.Image( '.\/img\/convchain.jpg'<\/span>)<\/span>\n<\/span>\ntab1<\/span> = pn.Column(<\/span>\n    pn.Row(inp),<\/span>\n    pn.layout.Divider(),<\/span>\n    pn.panel(conversation,  loading_indicator<\/span>=True<\/span>, height=300<\/span>),<\/span>\n    pn.layout.Divider(),<\/span>\n)<\/span>\ntab2<\/span>= pn.Column(<\/span>\n    pn.panel(cb.get_lquest),<\/span>\n    pn.layout.Divider(),<\/span>\n    pn.panel(cb.get_sources ),<\/span>\n)<\/span>\ntab3<\/span>= pn.Column(<\/span>\n    pn.panel(cb.get_chats),<\/span>\n    pn.layout.Divider(),<\/span>\n)<\/span>\ntab4<\/span>=pn.Column(<\/span>\n    pn.Row( file_input, button_load, bound_button_load),<\/span>\n    pn.Row( button_clearhistory, pn.pane.Markdown(\"Clears chat history. Can use to start a new topic\" )),<\/span>\n    pn.layout.Divider(),<\/span>\n    pn.Row(jpg_pane.clone(width<\/span>=400<\/span>))<\/span>\n)<\/span>\ndashboard<\/span> = pn.Column(<\/span>\n    pn.Row(pn.pane.Markdown('# ChatWithYourData_Bot')),<\/span><\/span>\n    pn.Tabs(('Conversation', tab1), ('Database', tab2), ('Chat History', tab3),('Configure', tab4))<\/span>\n)<\/span>\ndashboard<\/span>\n<\/code><\/pre>\n
      \u6b65\u9aa412\uff1a\u5728\u8fd0\u884c\u8d77\u6765\u7684\u7528\u6237\u754c\u9762\u4e0a\u8fdb\u884c\u5b9e\u9645\u7684\u95ee\u7b54\u5bf9\u8bdd\u3002<\/h3>\n
      \n
      \"\u7cbe\u534e\u7b14\u8bb0\uff1a\u5434\u6069\u8fbe<\/figure>\n<\/p>","protected":false},"excerpt":{"rendered":"
      \u6211\u4eec\u5c06\u5b66\u4e60\u5982\u4f55\u5229\u7528\u68c0\u7d22\u5230\u7684\u6587\u6863\u6765\u56de\u7b54\u7528\u6237\u7684\u95ee\u9898\uff0c\u4ee5\u53ca\u6784\u5efa\u4e00\u4e2a\u5b8c\u6574\u7684\u95ee\u7b54\u804a\u5929\u673a\u5668\u4eba\uff0c\u5b83\u5c06\u7ed3\u5408\u6211\u4eec\u4e4b\u524d\u8bb2\u8fc7\u7684\u6240\u6709\u7ec4\u4ef6\uff0c\u5e76\u5f15\u5165\u201c\u804a\u5929\u5386\u53f2\u201d\u8fd9\u4e2a\u6982\u5ff5\uff0c\u8ba9\u5b83\u5728\u56de\u7b54\u95ee\u9898\u65f6\u80fd\u591f\u8003\u8651\u5230\u4e4b\u524d\u7684\u5bf9\u8bdd\u6216\u4fe1\u606f\u3002<\/p>\n","protected":false},"author":1,"featured_media":5104,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"rank_math_title":"import | \u590d\u5236 | \u4ee3\u7801 - \u5948\u65af\u79d1\u6280\u793e\u533a","rank_math_description":"\u5948\u65af\u79d1\u6280\u793e\u533a\u4e13\u6ce8\u4e8eimport\u3001\u590d\u5236\u3001\u4ee3\u7801\u3001question\u3001\u6b65\u9aa4\uff0c\u63d0\u4f9b\u4e13\u4e1a\u7684\u5206\u4eab\u548c\u8ba8\u8bba\u5e73\u53f0\u3002","rank_math_focus_keyword":"import,\u590d\u5236,\u4ee3\u7801,question,\u6b65\u9aa4","views":"28","footnotes":""},"categories":[3],"tags":[126,127,128,129,136],"collection":[],"class_list":["post-1303","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-fenlei2","tag-gpt","tag-ai","tag-128","tag-129","tag-136"],"acf":[],"_links":{"self":[{"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/posts\/1303","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/comments?post=1303"}],"version-history":[{"count":0,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/posts\/1303\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/media\/5104"}],"wp:attachment":[{"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/media?parent=1303"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/categories?post=1303"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/tags?post=1303"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.nicekj.com\/nicekj2024\/wp\/v2\/collection?post=1303"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}