Function Calling

The Prompt Registry supports the OpenAI functions format. You can create a prompt using functions both visually and programmatically.

Be sure to read this OpenAI guide for more context.

Creating Visually

Functions can be defined, called, and set up visually through the Prompt Registry.

Publishing Programmatically

To publish a prompt template with functions programatically is very similar to how you would publish a regular prompt template. To add functions you can add the arguments functions and function_call to your prompt_template object.

If you’re already calling OpenAI API, you can bundle the functions and publish it into the Prompt Registry, like this:

Python SDK

messages = [{"role": "user", "content": "What's the weather like in Boston?"}]
functions = [
  {
    "name": "get_current_weather",
    "description": "Get the current weather in a given location",
    "parameters": {
      "type": "object",
      "properties": {
        "location": {
          "type": "string",
          "description": "The city and state, e.g. San Francisco, CA",
        },
        "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
      },
      "required": ["location"],
    },
  }
]

promptlayer.templates.publish({
    "prompt_name": "test_chat_functions",
    "prompt_template": {
        "type": "chat",
        "function_call": "auto",
        "messages": [
            {
                "role": role,
                "content": [
                    {
                        "type": "text",
                        "text": content,
                    }
                ],
            }
            for (role, content) in map(lambda x: x.values(), messages)
        ],
        "functions": functions,
    }
  })

response = openai.ChatCompletion.create(
  model="gpt-3.5-turbo",
  messages=messages,
  functions=functions,
  function_call="auto",
)