timkmecl / codegpt Goto Github PK

My request to inspect the code fails with the error below on Intellij Idea:

"PKIX path building failed: sun.security.provider.certpath.SunCertPathBuilderException: unable to find valid certification path to requested target"

I have added Open API certificate after downloading it to Intellij Idea via Tools-> Server Certificates.

Plugin Settings below:

Please advise.

Sometimes I work at a site where Codegpt extension does not work because access to the internet is restricted. The way around this is to forward traffic for a website to a proxy running on 192.168.1.123.

Is there a way to use this amazing VScode extension via a proxy server? Thanks!

Whenever I try to make a query to ChatGPT I get an error outputting [ERROR] 404 undefined.

Query I'm trying to send: "what's up".

W11, VSC Version 1.76.0,
Max Tokens: 1024, Model: ChatGPT, Temperature: 0,5

API Key last used is updated on OpenGPT's site but no response on VSC

After pressing shortcut for "Ask CodeGPT" or "Focus CodeGPT" user will immediately be able to start typing his prompt, without having to use his mouse to click on "Ask a question..." to switch focus on it.

How to use with Pycharm?

Error loading webview: Error: Could not register service worker: InvalidStateError: Failed to register a ServiceWorker: The document is in an invalid state..

I am not able load the sidebar because of the above error.

I am using Ubuntu 22 and vscode

Extension sometime erases template literals while inserting answer of GPT3:

Full code with that issue:

// const SipUser = require('../../models/SipUser.model');
const fs = require('fs');
const path = require('path');
const ccQueue = require('../../models/ccQueue.model');
// const {
//   handleResponse,
//   logger,
// } = require('../../../../utils/');
// TODO fix logger?

// This module handles the requests for configuration files 
module.exports = async (req, res) => {
  const params = req.body; 

  // Logging the params
  console.log(params.key_name, params.key_value);

  // Checking if we need to send the full directory or a file
  if (params.key_name === 'name' && params.key_value === 'callcenter.conf') {
    // Getting the queues
    let queues = await ccQueue.find({});
    queues = queues.map((queue) => queue.toXML()); 

    // Setting the response headers
    res.setHeader('Content-Type', 'freeswitch/xml');
    res.status(200);

    // Sending the response
    return res.send(`
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
<document type="freeswitch/xml">
  <section name="configuration">
    <configuration name="callcenter.conf" description="CallCenter (generated by **********)">                       
      <settings>

        <param name="odbc-dsn" value="mariadb://Server=secret :)"/>
        <param name="truncate-agents-on-load" value="false"/>
        <param name="truncate-tiers-on-load" value="false"/>

      </settings>

      <queues>

        ${queues.join('\n\n')}

      </queues>

    </configuration> 
  </section>
</document>
  `);
  }

  // Generating the file path
  const fileName = path.join(__dirname, '../../static/autoload_configs', `${params.key_value}.xml`);

  // Checking if that file exists
  if (fs.existsSync(fileName)) {
    // Reading the file
    const config = fs.readFileSync(fileName, 'utf8'); 

    // Setting the response headers
    res.setHeader('Content-Type', 'freeswitch/xml');
    res.status(200);

    // Sending the response
    return res.send(`<?xml version="1.0" encoding="UTF-8" standalone="no"?>
    <document type="freeswitch/xml">
      <section name="configuration">
        ${config}
      </section>
    </document>`);
  }

  // Sending a not found response
  res.setHeader('Content-Type', 'freeswitch/xml');
  res.status(200);
  res.send(`<?xml version="1.0" encoding="UTF-8" standalone="no"?>
  <document type="freeswitch/xml">
    <section name="result">
      <result status="not found" />
    </section>
  </document>`);
};

First of all, thank you for this wonderful extension!
It would be very helpful to provide a workflow that supports the insertion of generated documentations into the code editor.

I used codegpt extension in vscode for somedays without any problem. but today it doesn't work. codegpt chat doesn't answer. when I send a msg, it just shows a notif "stop responding" for some moments. When I write a comment and press ctrl+shift+i the window opens at right side but it doesn't work and just says "OpenAI API could not process the query, try selecting the code and using Ask CodeGPT to write your own query"

I have tried reopening vs, reinstalling codegpt, reentering openai key and didn't succeed in solving the problem.

I uninstalled copilot and gitlenz and some other extensions and then this problem occured. I have no idea if it would be related.

When I tried to use chatGPT it showed me this message

i installed in my laptop and getting the error
[ERROR] 429 [object Object]
but it works well in other laptops.
i tried it many times for a number of days . tried different emails but it didnt work

First of all response is very slow it may takes about 1 min.

All of Models didn't work expect "code-davinci-002", which give me wrong answers.

Type: Feature Request

its showing erroe eveerytime after my input

Extension version: 1.1.2
VS Code version: Code 1.76.2 (ee2b180d582a7f601fa6ecfdad8d9fd269ab1884, 2023-03-14T17:55:54.936Z)
OS version: Windows_NT x64 10.0.22621
Modes:
Sandboxed: No

If I want to use open-source model such as Llama2 or codellama. How can I change the code

How can I fix this?

Tried switching models from gpt3.5 to gpt 4 or gpt 4 32k, codegpt stops giving responses .
I had to uninstall and install it again , to make it work.
Please note that I am still in GPT-4 waitlist.

This error is coming after running codegpt.

System info: Apple M1 Pro.
using -- code-davinci-002

Please help.

Does anyone know if its possible to set api_version like "2024-01-02" In the Azure Openai settings ?

Since updating to new version it cant find the azure api key, which is still set and worked before:

Phenomenon

I entered my API key in CodeGPT: Set API KEY, and ask something. The answers are all `OpenAI API Response was: Error 401``

My Trial

• I tried to use Genie ChatGPT addon before, and have the same question.
• I tried to search the fucking website with keyword like Genie Azure, CodeGPT Azure, but got nothing useful.
• I tried to read the fucking document of Genie and CodeGPT, but none of the two mentioned Azure OpenAI API usage or guidance.
  Maybe because of my poor English reading ability :(

Azure OpenAI API Setting

GPT 35 turbo (the same as setted in CodeGPT)

My Wants

Please Help me with HOW TO FIND the guidance or usage of Azure OpenAI API
If you could solve my problem directly, words fails me to express my gratitude.

I 've installed the extenstion to vs-code from the marketplace and I 'm trying to set my api key and it gets me this error

Version: 1.72.1 (Universal)
OS : Macos Ventura version 13.0 (M1 Chip)

Would be nice to have word wrap option of code preview

I was excited to try this extension, but after installing and adding my API key if I ask a question in the panel i get [ERROR] 429 [object Object]

For me the extension is not working and I have API key set.

This is what I get instead of output: [ERROR] 429 [object Object]

Type: Bug

when typing a query for codeGPT I get the error: 404 undefined

Extension version: 1.1.2
VS Code version: Code 1.85.2 (8b3775030ed1a69b13e4f4c628c612102e30a681, 2024-01-18T06:40:10.514Z)
OS version: Windows_NT x64 10.0.22621
Modes:

Hello,

after I updated my API key (was deleted after plugin update again), I did the following configuration:

But the model responses with the following:

It works via the browser, though:

Actually when you write a text and try to use Ctrl + Shift + I (or by using Ask CodeGPT) it is returning this:

sk-<> API could not process the query, try selecting the code and using Ask CodeGPT to write your own query

Type: Feature Request

its showing error everytime after my input

Extension version: 1.1.2
VS Code version: Code 1.76.2 (ee2b180d582a7f601fa6ecfdad8d9fd269ab1884, 2023-03-14T17:55:54.936Z)
OS version: Windows_NT x64 10.0.22621
Modes:
Sandboxed: No

I tried to install this extension at Micro$oft Visual Studio but can not. I assume both are not compatible and this extension only can be installed in VSCode ?

@ALL

First of all - thank you for a tool that normally gives very good answers but not today

After a long useless conversation today I received this message from CODEGPT in VsCode:

I suggest you inform the coders responsible for CODEGPT of this failed attempt to help.
You can do this by submitting an issue to the CODEGPT repository on Github.
In the issue, you can include the code block you provided, along with a description of the issue and any other relevant information.
You can also tag the coders responsible for CODEGPT in the issue to ensure they see it.

The code I provided was a preliminary code used to investigate how to use EJS in my file webshop.ejs
I also provide the console output and resulting HTML-code

CODEGPT advised me to use exactly the same code

I tryied multiple times to explain that its advices did not work but it only resulted in an endless loop of questions and replies that was already given and which I had already informed that did not work

I also asked if this could be caused by a bug in EJS but received no relevant answer on that question

It seems that CODEGPT was just as puzzled as me and that's not what is expected, is it?

<%#
  No LI's in UL tag
    https://github.com/timkmecl/codegpt/issues/23  
%>
<%
const dbQuery = function(mysql)  {
  return new Promise((resolve, reject) => {
    con.query(mysql, (err, result) => {
      if (err) {
        reject(err);
      }
      else {
        resolve(result);
      }
    }); 
  });  
}
%>
<h2>Products</h2>
<ul>
  <% 
  dbQuery("SELECT * FROM products")
  .then(products => { 
    console.log("Data received from database: ", products);
    products.forEach(product => {
      console.log("Product name: ", product.name);
      %>
      <li> <%= product.name %> </li>
    <%  } ) } )
  .catch(err => console.log("Error fetching data from database: ", err));
  %>
</ul>

Console output:

Data received from database:  [
  RowDataPacket { id: 1, name: 'Product 1', price: 101 },
  RowDataPacket { id: 2, name: 'Product 2', price: 51 },
  RowDataPacket { id: 3, name: 'Product 3', price: 12 }
]
Product name:  Product 1
Product name:  Product 2
Product name:  Product 3

Resulting html:

<ul>
    
  </ul>

Hello,

loving this plugin. But the code generated by GPT is often very ugly. That's not really a big problem as I normally directly run the 'refactor' action of CodeGPT over it. I wonder if it might be possible to do that in one step:

• Insert the generated code at the position of the cursor in the editor
• Afer that, pass the code through GPT again to refactor it

Thank you! :)

trying to install and i see the above error:

Can't install 'timekmecl.codegpt3' extension because 
it is not compatible with the current version of 
Visual Studio Code (version 1.72.1).

Selecting ChatGPT as GPT3 model to use - results in [ERROR] 404 undefined

Selecting other models, including the default text-davinci-003, works.

When I change the VSCode theme to Light, the code is displayed with dark colors on a dark background.

API key is working, Text-davinci 3 is working fine.

It looks like one of the requests did get through based on the the usage logging but either it was too slow to respond or I missed that one. Tried with big chunks of code and a simple if-statement.
Language is Powershell (7.3).

From OpenAI:
Usage report:
20:15
Local time: 22 mrt 2023 21:15
text-davinci, 4 requests
1,270 prompt + 624 completion = 1,894 tokens
20:20
Local time: 22 mrt 2023 21:20
code-davinci, 1 request
417 prompt + 86 completion = 503 tokens
20:30
Local time: 22 mrt 2023 21:30
text-davinci, 1 request
672 prompt + 248 completion = 920 tokens