This is how I should set the analyze images function calling with Assistants.. right?

API
1

Hi,

  1. Define a function that returns a prompt and a file ID (the ID of the image), I have to make it return a file ID because I can’t pass a URL with the message to the thread, I only can pass file IDs.

  2. make an ID for the image with a “purpose” as “assistant”.

  3. add a message with the ID of the image to a thread using a body like this:

"role": "user", 
"content": "What in this image", 
"file_ids": [“file-YN31uCRnUkpM8AsUqJRzlR30”]

  1. Checking the run will return the prompt + the file ID.

  2. make a URL for the image.

  3. pass the prompt+ URL to a vision model.

etc.

Right?