Lots of issues using openAI in Python only on Ubuntu

I have a project utilising openAI

One of the changes required me to run openai migrate, it says

"You can run openai migrate to automatically upgrade your codebase to use the 1.0.0 interface. "

I did that on windows and had no issues. I can't remember how I did it. I think I just opened up powershell or terminal and wrote openai migrate (which doesn't really make sense to me, is it scanning my entire system and makes changes to update the code? Whatever, I digress).

Anyway, I'm moving to linux and ran my program (Django).

It says the exact same thing as before - that you need to run openai migrate.

So I open a terminal on my desktop, 'openai migrate' and viola..nothing. OpenAI not found.

I know that openai is installed in ./local after trying to follow another debugging step.

It seems I have 2 options

1 - manually refactor my code to comply with the new API or

2 - run the openAI migrate command which I believe does that already for you??

Anyway, I'm going with option #2 but I cannot even run open ai migrate.

I'm not a linux expert (yet..) and don't know about file systems, permissions, path. Defininitely do want to learn more about that in the future but not in the current time.

How can I solve this issue?

From a code perspective, this is the main areas of interest:

import openai
openai.api_key = 'xx'
def process_text(text):
    # Define the conversation as a list of messages
    conversation = [
        {
            "role": "system",
            "content": text
        }
    ]

    # Send the conversation to GPT-3.5 Turbo for completion
    response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo",
    messages=conversation,
    temperature=0.5,  # Slightly higher temperature for some response variation
    max_tokens=2000,  # Adjust for shorter responses, if needed
    top_p=0.9,  # Allow some controlled randomness for creativity
    frequency_penalty=0.2,  # Slightly penalize frequent words for variety
    presence_penalty=0.0  # No penalty for specific words
    )

    # Extract the chat response from GPT-3.5 Turbo
    chat_response = response.choices[0].message["content"]
    text = chat_response
    return text

Back to Top