[Issue] ChatGPT reply function doesn't work with gpt-3.5-turbo model

[martin@rootjazz]

https://rootjazz.com/twitterdub/updatetesting.html

Read the changelog warning, as includes a major update to global processed logs

[mihail]

Thank you!
Going to update in case any further problems with chatgpt appear

[mihail]

Now I face the following
Probably the response from chat gpt also differs from the regular models.

Code: Select all

21:49:24: 200
https://api.openai.com/v1/chat/completions
{
  "id": "chatcmpl-7PHnOYenhX9YsfYuhmIcUwv863zF4",
  "object": "chat.completion",
  "created": 1686260962,
  "model": "gpt-3.5-turbo-0301",
  "usage": {
    "prompt_tokens": 65,
    "completion_tokens": 28,
    "total_tokens": 93
  },
  "choices": [
    {
      "message": {
        "role": "assistant",
        "content": "I'm sure the login issues are just a fun feature for players to enjoy. Keep up the good work! 😜 #Diablo4"
      },
      "finish_reason": "stop",
      "index": 0
    }
  ]
}
21:49:24: 
Exception: ExActionFailed
Exception of type 'LibProcessing.ExActionFailed' was thrown.
   at _rZefrgKOUbhsuojLJiifOImpxgd._4i7Hn8gV3ReK7BIkDDUVKxxAQ5O(String , DbAction , String , String , String , String )
   at _rZefrgKOUbhsuojLJiifOImpxgd._hwPPhDwPoKsCxsJnI8XSWkhut5I(String , DbAction , String )

[martin@rootjazz]

the response I got from that endpoint was handled, but perhaps it's like the other endpoint where multiple formats may be returned.

I will handle this format for todays update

[martin@rootjazz]

The next update will fix this. I shall let you know when it is ready.



Regards,
Martin

[martin@rootjazz]

https://rootjazz.com/twitterdub/updatetesting.html