Check, Please! Giving Agents Access to Data
At AgentsCafé, all drink orders are saved in a file called orders.csv, which includes the columns: table_id, drink_name, and size.
Instead of manually scrolling through the file every time a customer asks for their bill, you’ll build a tool that looks up all current orders for a specific table.
Note: tool and pandas have been imported for you.
This exercise is part of the course
AI Agents with Hugging Face smolagents
Exercise instructions
- Use
table_idas the function parameter so the agent knows which table's orders to retrieve. - Read the
orders.csvfile, which contains all drink orders. - Return the list of drink orders for the table.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Create a tool that receives the table_id as input
@tool
def lookup_orders(____: str) -> list[str]:
"""
Retrieves the current drink orders for a café table.
Args:
table_id (str): The table's identifier (e.g., "T5").
Returns:
list[str]: A list of drink orders, each formatted like "Latte (Large)".
"""
# Read the orders.csv file
df = pd.read_csv('____')
orders = df[df['table_id'] == table_id].apply(lambda row: f"{row['drink_name']} ({row['size']})", axis=1).tolist()
# Return the table orders
return ____