Integrating custom tools and queries
You now have a custom math tool for calculating roof length. You can integrate it with an agentic workflow by creating a variable called query
that accepts the user's natural language question as a string. Your tool is already loaded as hypotenuse_length
, as well as your model
.
This exercise is part of the course
Designing Agentic Systems with LangChain
Exercise instructions
- Create a list variable called
tools
and include your tool,hypotenuse_length
, inside the list. - Create a variable called
query
that accepts questions as natural language strings. - Use the
create_react_agent()
function to create the agent, passing in themodel
andtools
. - Invoke the agent
app
, passing in yourquery
labeled"human"
, before storing and printing the agent'sresponse
.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Create a list variable and pass in your tool
____ = [____]
# Create a query using natural language
____ = "What is the hypotenuse length of a triangle with side lengths of 10 and 12?"
# Pass in the hypotenuse length tool and create the agent
app = ____(____, ____)
# Invoke the agent and print the response
response = ____.____({"messages": [("____", ____)]})
print(____['messages'][-1].content)