No output: response sequence happens but output is empty

[fnune]

https://github.com/Robitx/gp.nvim/assets/16181067/dc4004bb-c225-4b13-b1ec-194289187915

I have set up my OpenAI API key and double-checked that it's correct. gp.nvim shows me no errors and I couldn't find log files from it in ~/.local/{share,state}/nvim.

Here's my config:

  {
    "robitx/gp.nvim",
    tag = "v1.3.8",
    config = function()
      require("gp").setup()
    end,
  },

(using Lazy).

Have I done something wrong?

[Robitx]

@fnune Hey, what do :checkhealth gp and :GpInspectPlugin (ideally run after :GpChatRespond) show?

[fnune]

:checkhealth gp:

gp.nvim checks ~
- OK require('gp') succeeded
- OK require('gp').setup() has been called
- OK config.openai_api_key is set
- OK curl is installed
- OK grep is installed
- OK ln is installed
- WARNING sox is not installed

sox is optional so I didn't add it.

Output of :GpInspectPlugin (:warning: to others doing this: this includes your OpenAI API key):

Plugin structure:
{
  Prompt = <function 1>,
  Target = {
    append = 1,
    enew = <function 2>,
    popup = 3,
    prepend = 2,
    rewrite = 0
  },
  Whisper = <function 3>,
  _H = {
    autocmd = <function 4>,
    create_popup = <function 5>,
    cursor_to_line = <function 6>,
    delete_buffer = <function 7>,
    feedkeys = <function 8>,
    get_buffer = <function 9>,
    get_filetype = <function 10>,
    grep_directory = <function 11>,
    last_content_line = <function 12>,
    once = <function 13>,
    process = <function 14>,
    set_keymap = <function 15>,
    template_render = <function 16>,
    template_replace = <function 17>
  },
  _Name = "Gp (GPT prompt)",
  _chat_popup = {},
  _chat_popup_close = <function 18>,
  _first_line = -1,
  _last_line = -1,
  _payload = {
    messages = { {
        content = "You are a general AI assistant.",
        role = "system"
      }, {
        content = "Hi there!",
        role = "user"
      }, {
        content = "",
        role = "assistant"
      }, {
        content = "Summarize the topic of our conversation above in two or three words. Respond only with those words.",
        role = "user"
      } },
    model = "gpt-3.5-turbo-16k",
    stream = true
  },
  _response = "",
  _setup_called = true,
  call_hook = <function 19>,
  chat_template = "# topic: ?\n\n- model: %s\n- file: %s\n- role: %s\n\nWrite your queries after %s. Use `%s` or :%sChatRespond to generate a response.\nChats are saved automatically. To delete this chat, use `%s`
 or :%sChatDelete.\nBe cautious of very long chats. Start a fresh chat by using `%s` or :%sChatNew.\n\n---\n\n%s",
  cmd = {
    Append = <function 20>,
    ChatDelete = <function 21>,
    ChatFinder = <function 22>,
    ChatNew = <function 23>,
    ChatRespond = <function 24>,
    ChatToggle = <function 25>,
    Enew = <function 26>,
    Popup = <function 27>,
    Prepend = <function 28>,
    Rewrite = <function 29>,
    Stop = <function 30>,
    Whisper = <function 31>,
    WhisperAppend = <function 32>,
    WhisperEnew = <function 33>,
    WhisperPopup = <function 34>,
    WhisperPrepend = <function 35>,
    WhisperRewrite = <function 36>
  },
  cmd_hooks = {
    Implement = <function 37>,
    InspectPlugin = <function 38>
  },
  config = {
    command_prompt_prefix = "🤖 ~ ",
    command_system_prompt = "You are an AI that strictly generates just the formated final code.",
    openai_api_endpoint = "https://api.openai.com/v1/chat/completions",
    openai_api_key = "<REDACTED-BUT-VALID>",
    template_command = "{{command}}",
    template_rewrite = "I have the following code from {{filename}}:\n\n```{{filetype}}\n{{selection}}\n```\n\n{{command}}\n\nRespond just with the snippet of code that should be inserted.",
    template_selection = "I have the following code from {{filename}}:\n\n```{{filetype}}\n{{selection}}\n```\n\n{{command}}",
    whisper_dir = "/tmp/gp_whisper",
    whisper_max_time = "05:00",
    whisper_silence = "1.75",
    whisper_tempo = "1.75"
  },
  create_handler = <function 39>,
  delete_chat = <function 40>,
  error = <function 41>,
  open_chat = <function 42>,
  prepare_commands = <function 43>,
  prepare_payload = <function 44>,
  query = <function 45>,
  repo_instructions = <function 46>,
  setup = <function 47>,
  template_render = <function 48>
}
Command params:
{
  args = "",
  bang = false,
  count = -1,
  fargs = {},
  line1 = 20,
  line2 = 20,
  mods = "",
  name = "GpInspectPlugin",
  range = 0,
  reg = "",
  smods = {
    browse = false,
    confirm = false,
    emsg_silent = false,
    hide = false,
    horizontal = false,
    keepalt = false,
    keepjumps = false,
    keepmarks = false,
    keeppatterns = false,
    lockmarks = false,
    noautocmd = false,
    noswapfile = false,
    sandbox = false,
    silent = false,
    split = "",
    tab = -1,
    unsilent = false,
    verbose = -1,
    vertical = false
  }
}

[fnune]

Huh, I added some logging:

  local on_exit = _H.once(vim.schedule_wrap(function(code, signal)
    stdout:read_stop()
    stderr:read_stop()
    stdout:close()
    stderr:close()
    if handle and not handle:is_closing() then
      handle:close()
    end
    vim.api.nvim_out_write(string.format("code: %s", code))
    vim.api.nvim_out_write(string.format("stdout_data: %s", stdout_data))
    vim.api.nvim_out_write(string.format("stderr_data: %s", stderr_data))
    if callback then
      callback(code, signal, stdout_data, stderr_data)
    end
    M._handle = nil
    M._pid = nil
  end))

And got this:

stderr_data: code: 0stdout_data: {
    "error": {
        "message": "You exceeded your current quota, please check your plan and billing details.",
        "type": "insufficient_quota",
        "param": null,
        "code": "insufficient_quota"
    }
}

That is strange, because I'm on the Plus plan.

[fnune]

My API key says "never used":

[fnune]

I tried another API key just to test for minor blunders, but the key is correct.

[fnune]

Oh... this is so silly. ChatGPT and the OpenAI API are two different services!?

[fnune]

:grimacing: https://help.openai.com/en/articles/7039783-how-can-i-access-the-chatgpt-api

Sorry to have wasted your time... although surfacing this error message to the user would be awesome :smile:

[Robitx]

@fnune You haven't wasted my time, you've helped to save time to people who may hit the same pitfall in the future. :slightly_smiling_face:

I'll keep this open so I don't forget to propagate these errors.

[EvanDaley]

fnune

Which file did you add the logging in? I'm facing a similar issue, but I don't think its the token because the same token works when I use it in another similar plugin.

[justrajdeep]

@EvanDaley did you find the issue?