cab2ebfd9d
The LLM arc completes at section 05 (agentic systems), with neural networks as a standalone ML deep-dive in section 06. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
65 lines
1.6 KiB
Python
65 lines
1.6 KiB
Python
# tool_demo.py
|
|
#
|
|
# Minimal example of Ollama tool calling.
|
|
# The model requests a function call; our code executes it.
|
|
#
|
|
# CHEG 667-013
|
|
# E. M. Furst
|
|
|
|
from ollama import chat
|
|
|
|
# Define a tool as a regular Python function.
|
|
# The type hints and docstring tell the model what the function does.
|
|
|
|
def add(a: int, b: int) -> int:
|
|
"""
|
|
Add two numbers together.
|
|
|
|
Args:
|
|
a: The first number
|
|
b: The second number
|
|
|
|
Returns:
|
|
The sum of the two numbers
|
|
"""
|
|
return a + b
|
|
|
|
# A lookup table so we can dispatch by name
|
|
tools = {
|
|
'add': add,
|
|
}
|
|
|
|
# Send a message with the tool available
|
|
messages = [{'role': 'user', 'content': 'What is 247 plus 863?'}]
|
|
|
|
response = chat(
|
|
'llama3.1:8b',
|
|
messages=messages,
|
|
tools=[add],
|
|
)
|
|
|
|
# The model responds with a tool call, not text
|
|
if response.message.tool_calls:
|
|
for tool_call in response.message.tool_calls:
|
|
name = tool_call.function.name
|
|
args = tool_call.function.arguments
|
|
print(f'Model wants to call: {name}({args})')
|
|
|
|
# Execute the function
|
|
result = tools[name](**args)
|
|
print(f'Result: {result}')
|
|
|
|
# Send the result back to the model
|
|
messages.append(response.message)
|
|
messages.append({
|
|
'role': 'tool',
|
|
'content': str(result),
|
|
'tool_name': name,
|
|
})
|
|
|
|
# Get the final natural-language response
|
|
final = chat('llama3.1:8b', messages=messages)
|
|
print(f'\nModel: {final.message.content}')
|
|
else:
|
|
# Model chose not to use a tool
|
|
print(f'Model: {response.message.content}')
|