{"id":3667,"date":"2026-08-17T12:03:47","date_gmt":"2026-08-17T04:03:47","guid":{"rendered":"https:\/\/ai.zhousir.top\/?p=3667"},"modified":"2026-08-17T12:03:47","modified_gmt":"2026-08-17T04:03:47","slug":"langchain-%e6%a1%86%e6%9e%b6%ef%bc%9a%e4%bb%8e%e7%bc%96%e6%8e%92%e5%b1%82%e5%88%b0-agent-engineering","status":"publish","type":"post","link":"https:\/\/ai.zhousir.top\/?p=3667","title":{"rendered":"LangChain \u6846\u67b6\uff1a\u4ece\u7f16\u6392\u5c42\u5230 Agent Engineering"},"content":{"rendered":"
\n
\n

\u4e0a\u7bc7\u4f60\u628a\u591a\u667a\u80fd\u4f53\u6d41\u6c34\u7ebf\u8dd1\u901a\u4e86\u2014\u2014\u8c03\u5ea6\u5458\u6d3e\u6d3b\uff0c\u7814\u7a76\u5458\u3001\u5199\u624b\u3001\u7f16\u8f91\u5404\u53f8\u5176\u804c\u3002\u4f46\u624b\u5199\u534f\u8bae\u3001\u624b\u5199\u91cd\u8bd5\u3001\u624b\u5199\u4e0a\u4e0b\u6587\u4f20\u9012\uff0c\u5de5\u7a0b\u8d1f\u62c5\u8d8a\u6765\u8d8a\u91cd\u3002\u8fd9\u7bc7\u6362\u4e00\u6761\u8def\uff1a\u8ba4\u8bc6 LangChain<\/strong>\u2014\u2014\u628a LLM\u3001\u5de5\u5177\u3001\u8bb0\u5fc6\u3001\u68c0\u7d22\u7ec4\u88c5\u6210\u5e94\u7528\u7684\u7f16\u6392\u5c42\uff0c\u4ee5\u53ca\u5b83\u773c\u91cc\u7684 Agent<\/strong>\u3002<\/p>\n<\/blockquote>\n

\n

\u5c0f\u660e\u60f3\u505a\u300cAI \u65c5\u6e38\u52a9\u624b\u300d<\/h2>\n

\u5c0f\u660e\u5df2\u7ecf\u4f1a Function Calling\u3001ReAct\u3001Plan-and-Execute\uff0c\u8fd8\u5199\u8fc7 Researcher \u2192 Writer \u2192 Editor\u3002\u540c\u4e8b\u4e22\u6765\u4e00\u4e2a\u65b0\u9700\u6c42\uff1a<\/p>\n

<\/section>\n
\u505a\u4e00\u4e2a\u00a0AI\u00a0\u65c5\u6e38\u52a9\u624b\uff1a\n-\u00a0\u80fd\u804a\u5929\uff08\u6a21\u578b\uff09\n-\u00a0\u80fd\u67e5\u5929\u6c14\u3001\u641c\u666f\u70b9\uff08\u5de5\u5177\uff09\n-\u00a0\u80fd\u8bb0\u4f4f\u7528\u6237\u504f\u597d\uff08\u8bb0\u5fc6\uff09\n-\u00a0\u80fd\u7ffb\u5185\u90e8\u653b\u7565\u5e93\uff08\u68c0\u7d22\uff09\n-\u00a0\u6700\u597d\u8fd8\u80fd\u770b\u4e00\u773c\u884c\u7a0b\u6709\u6ca1\u6709\u8dd1\u504f\uff08\u56de\u8c03\/\u65e5\u5fd7\uff09\n<\/code><\/pre>\n
\u4ed6\u6253\u5f00\u65e7\u9879\u76ee\uff0c\u6253\u7b97\u300c\u518d\u62fc\u4e00\u5957\u300d\u3002\u534a\u5c0f\u65f6\u540e\uff0c\u6587\u4ef6\u6811\u5df2\u7ecf\u4e71\u4e86\uff1a<\/p>\n
<\/section>\n
travel_bot\/\n\u00a0\u00a0openai_client.py\u00a0\u00a0\u00a0\u00a0#\u00a0\u8c03\u6a21\u578b\n\u00a0\u00a0tools_weather.py\u00a0\u00a0\u00a0\u00a0#\u00a0\u5de5\u5177\u00a0A\n\u00a0\u00a0tools_poi.py\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0#\u00a0\u5de5\u5177\u00a0B\n\u00a0\u00a0memory_list.py\u00a0\u00a0\u00a0\u00a0#\u00a0\u81ea\u5df1\u7ef4\u62a4\u7684\u00a0messages\n\u00a0\u00a0rag_hack.py\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0#\u00a0\u534a\u6210\u54c1\u68c0\u7d22\n\u00a0\u00a0agent_loop.py\u00a0\u00a0\u00a0\u00a0\u00a0#\u00a0\u53c8\u6284\u4e86\u4e00\u904d\u00a0ReAct\n\u00a0\u00a0logger.py\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0#\u00a0print\u00a0\u5f53\u4eea\u8868\u76d8\n<\/code><\/pre>\n
\u5c0f\u660e\u53bb\u627e\u8001\u5f20\uff1a”\u6982\u5ff5\u6211\u90fd\u61c2\uff0c\u4f46\u6bcf\u505a\u4e00\u4e2a\u52a9\u624b\uff0c\u5c31\u8981\u91cd\u5199\u4e00\u904d\u80f6\u6c34\u4ee3\u7801\u3002\u7f51\u4e0a\u90fd\u8bf4\u7528 LangChain\uff0c\u5b83\u5230\u5e95\u89e3\u51b3\u4ec0\u4e48\uff1f”<\/p>\n
\u8001\u5f20\u8bf4\uff1a”\u4f60\u73b0\u5728\u7f3a\u7684\u4e0d\u662f\u518d\u61c2\u4e00\u4e2a\u7b97\u6cd5\uff0c\u800c\u662f\u4e00\u4e2a\u7f16\u6392\u5c42<\/strong>\u2014\u2014\u628a\u6a21\u578b\u3001\u5de5\u5177\u3001\u8bb0\u5fc6\u3001\u68c0\u7d22\u7528\u540c\u4e00\u5957\u79ef\u6728\u62fc\u8d77\u6765\u3002”<\/p>\n
\n
\u4ec0\u4e48\u662f LangChain<\/h2>\n
\u8001\u5f20\u5728\u767d\u677f\u4e0a\u5199\u4e86\u4e09\u884c\uff1a<\/p>\n
<\/section>\n
LLM\u00a0\u00a0\u00a0\u00a0=\u00a0\u5927\u8111\uff08\u4f1a\u8bf4\u3001\u4f1a\u63a8\u7406\uff09\nAgent\u00a0\u00a0\u00a0\u00a0=\u00a0\u5e26\u624b\u811a\u4e0e\u601d\u7ef4\u903b\u8f91\u7684\u72ec\u7acb\u4e2a\u4f53\nLangChain\u00a0\u00a0=\u00a0\u5e2e\u4f60\u628a\u5927\u8111\u63a5\u5230\u6570\u636e\u3001API\u3001\u5de5\u5177\u4e0a\u7684\u7f16\u6392\u5e73\u53f0\n<\/code><\/pre>\n
“LangChain \u7531 Harrison Chase \u521b\u5efa\u4e8e 2022 \u5e74 10 \u6708\uff0c\u5b9a\u4f4d\u662f Agent Engineering<\/strong> \u5e73\u53f0\uff1a\u5e2e\u52a9\u5f00\u53d1\u8005\u628a\u5927\u8bed\u8a00\u6a21\u578b\u4e0e\u5916\u90e8\u6570\u636e\u6e90\u3001API\u3001\u5de5\u5177\u8fde\u63a5\u8d77\u6765\uff0c\u5feb\u901f\u6784\u5efa\u590d\u6742 AI \u5e94\u7528\u3002<\/p>\n
\n
\u5b98\u7f51\uff1ahttps:\/\/www.langchain.com\/<\/a>
\n\u4e2d\u6587\u6587\u6863\uff1ahttps:\/\/langchain-doc.cn\/<\/a><\/p>\n<\/blockquote>\n
\u4f60\u53ef\u4ee5\u628a\u5b83\u60f3\u6210\u667a\u80fd\u5e94\u7528\u7684\u4e50\u9ad8\u5e95\u677f<\/strong>\uff1a<\/p>\n