Error “[object Object]“ when using LangChain‘s AzureChatOpenAI in Node.js

**题意:**在Node.js中使用LangChain库的AzureChatOpenAI功能时遇到错误"Error 'object Object'"

问题背景:

I'm attempting to use LangChain's AzureChatOpenAI with the gpt-35-turbo-16k model in a Node.js application to create an OpenAI Function Agent. I've properly set up the required environment variables for Azure OpenAI setup. However, I'm encountering an error, which is not very descriptive, when trying to run the function.

在Node.js应用程序中尝试使用LangChain的AzureChatOpenAI与gpt-35-turbo-16k模型来创建一个OpenAI Function Agent时,已经正确地设置了Azure OpenAI所需的环境变量,但在尝试运行该函数时遇到了一个不太具体的错误

Here's the relevant code snippet: 下面是相关的代码片段:

cs 复制代码
import { pull } from "langchain/hub";
import type { ChatPromptTemplate } from "@langchain/core/prompts";
import { AzureChatOpenAI } from "@langchain/azure-openai";

const model = new AzureChatOpenAI({
  azureOpenAIApiDeploymentName: "gpt-35-turbo-16k",
  azureOpenAIApiVersion: "2024-03-01-preview",
  modelName: "gpt-35-turbo-16k",
  temperature: 0,
  maxRetries: 1,
});

export async function ChatWithFunctionAgent(question: string) {
  try {
    // Get the prompt to use - you can modify this!
    // If you want to see the prompt in full, you can at:
    // https://smith.langchain.com/hub/hwchase17/openai-functions-agent
    const prompt = await pull<ChatPromptTemplate>(
      "hwchase17/openai-functions-agent"
    );
    const agent = await createOpenAIFunctionsAgent({
      llm: model,
      tools: [],
      prompt,
    });
    const agentExecutor = new AgentExecutor({
      agent,
      tools: [],
    });
    const result = await agentExecutor.invoke({
      input: "what is LangChain?",
    });
    console.log(result);
    return result;
  } catch (error) {
    console.log(error);
  }
}

Error: 错误信息:

cs 复制代码
Http function processed request for url "http://localhost:7071/api/chat"
[1] [2024-04-04T05:27:52.461Z] Error: [object Object]
[1] [2024-04-04T05:27:52.461Z]     at /home/ujwal/code/roc-next/azure-func/node_modules/@langchain/core/dist/utils/async_caller.cjs:98:23
[1] [2024-04-04T05:27:52.461Z]     at process.processTicksAndRejections (node:internal/process/task_queues:95:5)
[2024-04-04T05:27:52.461Z]     at async RetryOperation._fn (/home/ujwal/code/roc-next/azure-func/node_modules/p-retry/index.js:50:12) {
[1] [2024-04-04T05:27:52.461Z]   attemptNumber: 2,
[1] [2024-04-04T05:27:52.461Z]   retriesLeft: 0
[1] [2024-04-04T05:27:52.461Z] }

What I've tried: 我已经尝试做的事情:

  • I've already double-checked to ensure all the required environment variables are set, including those needed for authentication with Azure.

我已经反复检查过了,确保所有必需的环境变量都已设置,包括与Azure进行身份验证所需的环境变量。

  • I've manually invoked the model, that is working fine, so no issues with Azure credentials or connection.

我已经手动调用了模型,它工作正常,所以Azure的凭据或连接没有问题。

Expected behavior: 期望结果:

  • AzureChatOpenAI to work with langchain OpenAIFunction Agent

让AzureChatOpenAI与langchain的OpenAIFunctionAgent协同工作

Any help would be greatly appreciated! 任何帮助都将不胜感激!

问题解决:

I figured out why this issue was occuring, if the tools array is empty, the AzureOpenAI API throws an error. To fix this, you need to have atleast one tool in the tools array, you can refer to Defining custom tools | 🦜️🔗 Langchain to create custom tools. You also need to make sure that the name of the tool does not contain any spaces.

我弄清楚了这个问题发生的原因,如果tools数组为空,AzureOpenAI API会抛出一个错误。为了解决这个问题,你需要在tools数组中至少包含一个工具。你可以参考"Defining custom tools | 🦜️🔗 Langchain"来创建自定义工具。此外,你还需要确保工具的名称中不包含任何空格。

cs 复制代码
import { DynamicTool } from "@langchain/core/tools";

const tools = [
  new DynamicTool({
    name: "FOO",
    description:
      "call this to get the value of foo. input should be an empty string.",
    func: async () => "baz",
  })
]
Hope this helps.
相关推荐
养肥胖虎2 小时前
CodeGraph学习笔记:给代码建索引,节省Token和时间
ai·codegraph·代码索引
10年前端老司机3 小时前
你的RAG检索正在“高效地重复废话”:一文彻底搞懂MMR算法
langchain·llm·agent
曹牧3 小时前
C#:未被引用的函数
c#
明月_清风4 小时前
Deno 终局来了:从挑战 Node 到被 Cloudflare 收编
前端·后端·node.js
Raas1005 小时前
MAI Gateway(魔芋企业级AI网关)能力解析:AI网关能做故障转移吗?AI网关核心功能详解
大数据·人工智能·网关·ai·gateway·mai gateway
全栈练习生5 小时前
AI Agent 沙箱
python·ai
麦客奥德彪8 小时前
Codex 最近只有约 25 token/s,我做了个插件,把 ChatGPT 接进 DSH
openai·ai编程·deepseek
guslegend8 小时前
脚手架原理与本地调试:从 bin 软链接到 npm link
前端·npm·node.js·脚手架·前端工程化
GlobalInfo8 小时前
2026年推理算力超越训练算力,市场调研该关注什么
大数据·人工智能·ai·芯片
JavaGuide8 小时前
最近爆火的 Muse 浙大开源版 nanoMuse,来了!
aigc·openai·ai编程