Working with Prompt IDs in client.responses.create

API
1

Hi,

Sorry in advance for the newbie question but I am a little confused.

I have this code segment

response = self.client.responses.create(
    prompt={
        "id": self.prompt_id,
        "version": self.prompt_version,
    },
    input=message,
)

The code works but ‘prompt’ is complaining with the following message

Expected type ‘ResponsePromptParam | None | Omit’, got ‘dict[str, Any]’ instead

Is there a correction I should make?

Thanks Chris.

As an aside I’m so confused by the various options

  • chat.completions.create
  • openai chat prompts
  • openai assistants

I started with chat.completions.create passing system messages but when I could no longer add notes to projects in ChatGPT I discovered chat prompts and assistants. Assistants seems to be dead or on its way.

Is chat prompts a way of not sending system messages with every request like you would in completions.create?

My full method is like this

    def get_response(self, message):

        if self.response_id is None:
            response = self.client.responses.create(
                prompt={
                    "id": self.prompt_id,
                    "version": self.prompt_version,
                },
                input=message,
            )
            self.response_id = response.id
        else:
            response = self.client.responses.create(
                prompt={
                    "id": self.prompt_id,
                    "version": self.prompt_version,
                },
                input=message,
            )

        return response.output_text