smolagents_example.py

# pytest: ollama, e2e
"""Example showing how to use pre-built Hugging Face smolagents tools with Mellea.

This demonstrates loading existing tools from the smolagents ecosystem,
similar to how you can use langchain tools with MelleaTool.from_langchain().

The smolagents library provides various pre-built tools like:
- PythonInterpreterTool for code execution
- DuckDuckGoSearchTool for web search (requires ddgs package)
- WikipediaSearchTool for Wikipedia queries
- And many others from the Hugging Face ecosystem

Prerequisites:
    uv sync --extra tools
    ollama pull granite4.2:3b
    ollama serve

Usage:
    uv run python docs/examples/tools/smolagents_example.py
"""

from mellea import start_session
from mellea.backends import ModelOption
from mellea.backends.tools import MelleaTool

try:
    # Import a pre-built tool from smolagents
    from smolagents import PythonInterpreterTool

    # Create the smolagents tool instance
    python_tool_hf = PythonInterpreterTool()

    # Convert to Mellea tool - now you can use it with Mellea!
    python_tool: MelleaTool = MelleaTool.from_smolagents(python_tool_hf)

    # Use with Mellea session
    m = start_session()
    result = m.instruct(
        "Calculate the sum of numbers from 1 to 10 using Python",
        model_options={ModelOption.TOOLS: [python_tool]},
        tool_calls=True,
    )

    print(f"Response: {result}")

    if result.tool_calls:
        try:
            tool_call = next(
                (tc for tc in result.tool_calls if tc.name == python_tool.name), None
            )
            if tool_call:
                calc_result = tool_call.call_func()
                print(f"\nCalculation result: {calc_result}")
        except Exception as e:
            print(f"\nTool execution failed: {e}")

except ImportError as e:
    print("Please install smolagents: uv pip install 'mellea[tools]'")
    print(f"Error: {e}")