Authentication error: <empty message>

API
1

I am encountering a very strange behaviour using the API. It used to work perfectly fine but since a few days the API returns an empty message. I included a rstrip to remove any potential white spaces or \n at the end of the prompt.

Please help!!
This is the part of the script:
Remove any trailing whitespace or newline characters, this sometimes produces error for openai api

    content = content.rstrip()

    retry_limit = 5
    response = None

    for retry_count in range(retry_limit):
        try:
            response = self.call_openai_api(
                model="gpt-3.5-turbo",
                messages=[{"role": "user", "content": content}],
                max_tokens=2500,
                n=1,
                stop=None,
                temperature=0.0,
            )
            break  # If the call was successful, break the loop

        except RateLimitError as e:
            print(f"OpenAI API request exceeded rate limit: {e}. Attempt {retry_count + 1} of {retry_limit}")
            time.sleep(10)  # sleep for a while before retrying

        except APIError as e:
            print(f"OpenAI API returned an API Error: {e}")
            break

        except APIConnectionError as e:
            print(f"Failed to connect to OpenAI API: {e}")
            break

        except InvalidRequestError as e:
            print(f"Invalid request: {e}")
            break

        except AuthenticationError as e:
            print(f"Authentication error: {e}")
            break

        except ServiceUnavailableError as e:
            print(f"Service is currently unavailable: {e}")
            break

        except Exception as e:
            print(f"An unexpected error occurred: {e}")
            break

    else:  # This part will execute if the for loop completed all iterations without a break statement.
        print("Retry limit exceeded.")

And the response: Authentication error:
Using playground works perfectly fine. My API key is correct and working (when using a simple test prompt).

2

Welcome to the OpenAI community @cp1

Do you get this issue when using the openai python library?

3

Thank you @sps :slight_smile:

Yes, I am using the openai lib, next to the error handling packages:

import openai
from openai.error import APIError, APIConnectionError, RateLimitError, InvalidRequestError, AuthenticationError, ServiceUnavailableError

4

Interesting. I’m curious why aren’t you using the boilerplate code for calling the chat completion api?

I’m asking because the definition of call_openai_api() isn’t included in the code you shared, and the function looks just what the openai.chatCompletion.create() should be.

5

I defined prior a method to handle different parts of the API, but the call is basically the same as the standard call.

@retry(retry_on_exception=lambda e: isinstance(e, RateLimitError), stop_max_attempt_number=10, wait_exponential_multiplier=1000, wait_exponential_max=10000)
def call_openai_api(*args, **kwargs):
    return openai.ChatCompletion.create(*args, **kwargs)
6

Please use the boilerplate code with a newly generated api key. Let me know if your still run into the error.

7

That seemed to solve the issue - not sure exactly why though. The API key was not part of the issue.
Thanks for the help!!

1 Like