OpenAI Function Calling

Tutorial 1

Jan Kirenz

OpenAI Function Calling

Setup

import os
import openai
import json

from dotenv import load_dotenv, find_dotenv
_ = load_dotenv(find_dotenv()) 
openai.api_key = os.environ['OPENAI_API_KEY']

Functions

Weather example function

# Example dummy function hard coded to return the same weather
# In production, this could be your backend API or an external API
def get_current_weather(location, unit="celsius"):
    """Get the current weather in a given location"""
    weather_info = {
        "location": location,
        "temperature": "16",
        "unit": unit,
        "forecast": ["sunny", "windy"],
    }
    return json.dumps(weather_info)

Define a function

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. Stuttgart, BW",
                },
                "unit": {"type": "string", "enum": ["celsius", "fahrenheit"]},
            },
            "required": ["location"],
        },
    }
]

Messages

messages = [
    {
        "role": "user",
        "content": "What's the weather like in Stuttgart?"
    }
]
response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo-0613",
    messages=messages,
    functions=functions
)

Response message content

  • Content is empty
response_message["content"]
  • Function call is a dictionary
response_message["function_call"]
  • <OpenAIObject at 0x120f9dc10> JSON: { “name”: “get_current_weather”, “arguments”: “{"location": "Stuttgart"}” }

Inspect JSON

json.loads(response_message["function_call"]["arguments"])

{‘location’: ‘Stuttgart’}

args = json.loads(response_message["function_call"]["arguments"])
get_current_weather(args)
  • ‘{“location”: {“location”: “Stuttgart”}, “temperature”: “16”, “unit”: “celsius”, “forecast”: [“sunny”, “windy”]}’

New message

  • New message with no relation to weather
messages = [
    {
        "role": "user",
        "content": "hi!",
    }
]
response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo-0613",
    messages=messages,
    functions=functions,
)

Show response

print(response)

{ “id”: “chatcmpl-8GqJyabUr85mp4mGjFLZJuKGOoBu6”, “object”: “chat.completion”, “created”: 1699025062, “model”: “gpt-3.5-turbo-0613”, “choices”: [ { “index”: 0, “message”: { “role”: “assistant”, “content”: “Hello! How can I assist you today?” }, “finish_reason”: “stop” } ], “usage”: { “prompt_tokens”: 75, “completion_tokens”: 10, “total_tokens”: 85 } }

Use function call auto

messages = [
    {
        "role": "user",
        "content": "hi!",
    }
]

response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo-0613",
    messages=messages,
    functions=functions,
    function_call="auto",
)
print(response)

{ “id”: “chatcmpl-8Ezl8M209h6uqOXuFrGzVwAoSjXVI”, “object”: “chat.completion”, “created”: 1698584686, “model”: “gpt-3.5-turbo-0613”, “choices”: [ { “index”: 0, “message”: { “role”: “assistant”, “content”: “Hello! How can I assist you today?” }, “finish_reason”: “stop” } ], “usage”: { “prompt_tokens”: 75, “completion_tokens”: 10, “total_tokens”: 85 } }

Disable function usage

Don’t use the function

messages = [
    {
        "role": "user",
        "content": "hi!",
    }
]
response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo-0613",
    messages=messages,
    functions=functions,
    function_call="none",
)
print(response)

{ “id”: “chatcmpl-8GqKcx0cYK0MWb0ONpWqTTbU1hWlJ”, “object”: “chat.completion”, “created”: 1699025102, “model”: “gpt-3.5-turbo-0613”, “choices”: [ { “index”: 0, “message”: { “role”: “assistant”, “content”: “Hello! How can I assist you today?” }, “finish_reason”: “stop” } ], “usage”: { “prompt_tokens”: 76, “completion_tokens”: 9, “total_tokens”: 85 } }

Force function usage

  • force to use the function
messages = [
    {
        "role": "user",
        "content": "hi!",
    }
]
response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo-0613",
    messages=messages,
    functions=functions,
    function_call={"name": "get_current_weather"},
)
print(response)

{ “id”: “chatcmpl-8GqKwA2jwWb8PmgrXGZANDyxcsO0l”, “object”: “chat.completion”, “created”: 1699025122, “model”: “gpt-3.5-turbo-0613”, “choices”: [ { “index”: 0, “message”: { “role”: “assistant”, “content”: null, “function_call”: { “name”: “get_current_weather”, “arguments”: “{"location": "Stuttgart, BW"}” } }, “finish_reason”: “stop” } ], “usage”: { “prompt_tokens”: 82, “completion_tokens”: 12, “total_tokens”: 94 } }

Pass the reults back in the LLM

messages = [
    {
        "role": "user",
        "content": "What's the weather like in Boston!",
    }
]
response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo-0613",
    messages=messages,
    functions=functions,
    function_call={"name": "get_current_weather"},
)
print(response)

Response

{ “id”: “chatcmpl-8EzqG43iC6CWCOqCQiSWxngMHJNIQ”, “object”: “chat.completion”, “created”: 1698585004, “model”: “gpt-3.5-turbo-0613”, “choices”: [ { “index”: 0, “message”: { “role”: “assistant”, “content”: null, “function_call”: { “name”: “get_current_weather”, “arguments”: “{"location": "Boston, MA"}” } }, “finish_reason”: “stop” } ], “usage”: { “prompt_tokens”: 88, “completion_tokens”: 11, “total_tokens”: 99 } }

Append message

Append to list of messages

messages.append(response["choices"][0]["message"])
args = json.loads(response["choices"][0]["message"]['function_call']['arguments'])

observation = get_current_weather(args)
  • This is the response of calling a funtion
messages.append(
        {
            "role": "function",
            "name": "get_current_weather",
            "content": observation,
        }
)

Response

response = openai.ChatCompletion.create(
    model="gpt-3.5-turbo-0613",
    messages=messages,
)
print(response)

{ “id”: “chatcmpl-8EzsXbeQiRTRNHnGV9eCzlfCZrNU2”, “object”: “chat.completion”, “created”: 1698585145, “model”: “gpt-3.5-turbo-0613”, “choices”: [ { “index”: 0, “message”: { “role”: “assistant”, “content”: “The current temperature in Boston is 16 degrees Celsius. The weather is sunny and windy.” }, “finish_reason”: “stop” } ], “usage”: { “prompt_tokens”: 77, “completion_tokens”: 18, “total_tokens”: 95 } }

Acknowledgments

What’s next?

Congratulations! You have completed this tutorial 👍

Next, you may want to go back to the lab’s website

Jan Kirenz