Hello!
I was wondering, when using function calls, we have to essentially make two requests to chat-gpt. First to pass in the list of available tools, second to pass the output text from the tools back to the api. Example from the docs:
from openai import OpenAI
import json
client = OpenAI()
# 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="fahrenheit"):
"""Get the current weather in a given location"""
if "tokyo" in location.lower():
return json.dumps({"location": "Tokyo", "temperature": "10", "unit": unit})
elif "san francisco" in location.lower():
return json.dumps({"location": "San Francisco", "temperature": "72", "unit": unit})
elif "paris" in location.lower():
return json.dumps({"location": "Paris", "temperature": "22", "unit": unit})
else:
return json.dumps({"location": location, "temperature": "unknown"})
def run_conversation():
# Step 1: send the conversation and available functions to the model
messages = [{"role": "user", "content": "What's the weather like in San Francisco, Tokyo, and Paris?"}]
tools = [
{
"type": "function",
"function": {
"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"],
},
},
}
]
response = client.chat.completions.create(
model="gpt-3.5-turbo-1106",
messages=messages,
tools=tools,
tool_choice="auto", # auto is default, but we'll be explicit
)
response_message = response.choices[0].message
tool_calls = response_message.tool_calls
# Step 2: check if the model wanted to call a function
if tool_calls:
# Step 3: call the function
# Note: the JSON response may not always be valid; be sure to handle errors
available_functions = {
"get_current_weather": get_current_weather,
} # only one function in this example, but you can have multiple
messages.append(response_message) # extend conversation with assistant's reply
# Step 4: send the info for each function call and function response to the model
for tool_call in tool_calls:
function_name = tool_call.function.name
function_to_call = available_functions[function_name]
function_args = json.loads(tool_call.function.arguments)
function_response = function_to_call(
location=function_args.get("location"),
unit=function_args.get("unit"),
)
messages.append(
{
"tool_call_id": tool_call.id,
"role": "tool",
"name": function_name,
"content": function_response,
}
) # extend conversation with function response
second_response = client.chat.completions.create(
model="gpt-3.5-turbo-1106",
messages=messages,
) # get a new response from the model where it can see the function response
return second_response
print(run_conversation())
What actually happens under the hood in step 4? How is the text from both function calls actually formatted and sent to the model?
I’m asking this because I wonder if, for the second request, what is the difference between:
-
Sending responses from multiple tool calls and separately appending each output to messages, like the above example.
-
Sending responses from multiple tool calls, but concatenating all the text together so we only need to append once to messages?