Question 207955 in Javascript Question posted in
A very good W3school tutorial can be found here.

Javascript – Error in passing array data to createChatCompletion ChatGPT

I am exploring chatGPT API Documentation. I created a following array and sending it to generate response from the ChatGPT API. It works fine.

But whenever I am pushing an object inside that array and then sending it to the ChatGPT API it is showing an error.

const { Configuration, OpenAIApi } = require('openai');
const configuration = new Configuration({
  apiKey: 'Your OPENAI_API_KEY',
});

const chatGPTProvider = async () => {
  try {
    const url = 'https://api.openai.com/v1/chat/completions';

    let data = [
      { role: 'system', content: 'You are a helpful assistant.' },
      {
        role: 'user',
        content: 'Act as an intelligent AI.',
      },
    ];

    // data.push({ role: 'sytem', content: 'Hola GPT' });

    // console.log(data);

    const resp = await openai.createChatCompletion({
      model: 'gpt-3.5-turbo',
      messages: data,
    });

    return { status: 'SUCCESS', result: resp.data.choices[0].message };
  } catch (ex) {
    console.log(ex.message);

    return { status: 'ERROR', result: null };
  }
};

This code works fine. When you uncomment that push method it starts giving error.

Can you please tell why only pushing data into the array results in error?

Documentation: https://platform.openai.com/docs/api-reference/chat/create?lang=node.js

I am getting the following Error
Error: Request failed with status code 400
at createError (/home/sushant007/Interview-Project/GPT-chatbot/server/node_modules/openai/node_modules/axios/lib/core/createError.js:16:15)
at settle (/home/sushant007/Interview-Project/GPT-chatbot/server/node_modules/openai/node_modules/axios/lib/core/settle.js:17:12)
at IncomingMessage.handleStreamEnd (/home/sushant007/Interview-Project/GPT-chatbot/server/node_modules/openai/node_modules/axios/lib/adapters/http.js:322:11)
at IncomingMessage.emit (node:events:524:35)
at endReadableNT (node:internal/streams/readable:1378:12)
at process.processTicksAndRejections (node:internal/process/task_queues:82:21)

Tags:

2

Answers


  1. Chosen as BEST ANSWER
    0

    Only four type of role are possible in ChatGPT API. These are system, user, assistant, or function.

    Ref: https://platform.openai.com/docs/api-reference/chat/create#chat/create-role


  2. 0

    You need to call JSON.stringify() because the ChatGPT API is calling JSON.parse() on an already-parsed object, which should be throwing an error.

    const { Configuration, OpenAIApi } = require('openai');
const configuration = new Configuration({
  apiKey: 'Your OPENAI_API_KEY',
});

const chatGPTProvider = async () => {
  try {
    const url = 'https://api.openai.com/v1/chat/completions';

    let data = [
      { role: 'system', content: 'You are a helpful assistant.' },
      {
        role: 'user',
        content: 'Act as an intelligent AI.',
      },
    ];

    data.push(JSON.stringify({ role: 'sytem', content: 'Hola GPT' }));

    console.log(data);

    const resp = await openai.createChatCompletion({
      model: 'gpt-3.5-turbo',
      messages: data,
    });

    return { status: 'SUCCESS', result: resp.data.choices[0].message };
  } catch (ex) {
    console.log(ex.message);

    return { status: 'ERROR', result: null };
  }
};

    The 400 status code probably means malformed or strange syntax that the server can’t parse. So I’m assuming that’s what the problem is, so try the JSON.stringify() method.

    Login or Signup to reply.
Please signup or login to give your own answer.

Back To Top
Search