OpenAI functions agent in LangChain - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ component_behavior

intermediate

What is the output of this LangChain OpenAI functions agent call?

Given the following code snippet using LangChain's OpenAI functions agent, what will be the printed output?

LangChain

from langchain.agents import create_openai_functions_agent
from langchain.chat_models import ChatOpenAI

def greet(name: str) -> str:
    return f"Hello, {name}!"

llm = ChatOpenAI(temperature=0)
agent = create_openai_functions_agent(llm, [greet])

response = agent.invoke({'name': 'Alice'})
print(response)

A KeyError: 'name'

B "Hello, {name}!"

C "Hello, Alice!"

D TypeError: greet() missing 1 required positional argument: 'name'

💡 Hint

Think about how the function greet is called with the argument 'name' from the input dictionary.

The agent calls the greet function with the argument 'name' set to 'Alice', so it returns 'Hello, Alice!'.

📝 Syntax

intermediate

Which option correctly defines a function for OpenAI functions agent in LangChain?

You want to define a function to be used by LangChain's OpenAI functions agent. Which of the following function definitions is syntactically correct and compatible?

A

def add_numbers(a: int, b: int) -> int:
    return a + b

B

def add_numbers(a, b):
return a + b

C

def add_numbers(a: int, b: int):
    print(a + b)

D

def add_numbers(a: int, b: int) -> int
    return a + b

💡 Hint

Check for proper indentation, type annotations, and function syntax.

Option A is correctly indented, has proper type annotations, and returns the sum as an int. Option A lacks indentation. Option A returns None because it prints instead of returning. Option A misses a colon after the function signature.

❓ state_output

advanced

What is the value of 'result' after invoking the agent with nested function calls?

Consider this code using LangChain's OpenAI functions agent with two functions. What is the value of 'result' after calling agent.invoke({'x': 3, 'y': 4})?

LangChain

from langchain.agents import create_openai_functions_agent
from langchain.chat_models import ChatOpenAI

def multiply(a: int, b: int) -> int:
    return a * b

def add_and_multiply(x: int, y: int) -> int:
    sum_val = x + y
    return multiply(sum_val, 2)

llm = ChatOpenAI(temperature=0)
agent = create_openai_functions_agent(llm, [add_and_multiply, multiply])

result = agent.invoke({'x': 3, 'y': 4})

A KeyError: 'x'

B 10

C TypeError due to missing argument in multiply

D 14

💡 Hint

Trace the calls: add_and_multiply adds x and y, then calls multiply with sum_val and 2.

add_and_multiply(3,4) computes sum_val=7, then calls multiply(7,2) which returns 14. So result is 14.

🔧 Debug

advanced

Which option causes a runtime error when invoking the OpenAI functions agent?

Given these function definitions for an OpenAI functions agent, which option will cause a runtime error when calling agent.invoke({'name': 'Bob'})?

LangChain

from langchain.agents import create_openai_functions_agent
from langchain.chat_models import ChatOpenAI

def say_hello(name: str) -> str:
    return f"Hello, {name}!"

def say_goodbye() -> str:
    return "Goodbye!"

llm = ChatOpenAI(temperature=0)
agent = create_openai_functions_agent(llm, [say_hello, say_goodbye])

A agent.invoke({})

B agent.invoke({'name': 'Bob'})

C agent.invoke({'name': 'Bob', 'extra': 123})

D agent.invoke({'name': None})

💡 Hint

Check which function requires arguments and what happens if they are missing.

Option A calls agent.invoke with an empty dict, so say_hello(name) is called without 'name', causing a TypeError.

🧠 Conceptual

expert

Which statement best describes how LangChain's OpenAI functions agent handles function calls?

Select the most accurate description of how LangChain's OpenAI functions agent processes function calls when given multiple functions.

A The agent calls all provided functions sequentially with the same input and returns the last function's output.

B The agent uses the LLM to decide which function to call based on the input and calls it with matching arguments automatically.

C The agent requires manual selection of which function to call before invocation.

D The agent only supports one function and ignores others if multiple are provided.

💡 Hint

Think about how the agent uses the language model to route calls.

LangChain's OpenAI functions agent uses the language model to interpret input and decide which function to call with appropriate arguments automatically.