`agentix.tools`¶

tools ¶

Defining tools.

The @tool decorator turns a plain, typed Python function into a registered :class:Tool whose JSON Schema is derived from the function's type hints and docstring. One decorated function is the single source of truth: the name, the description, the parameter schema the model sees, and the executable body — so the schema can't drift from the implementation.

@tool
def get_weather(city: str) -> str:
    '''Get the current weather for a city.

    Args:
        city: City name, e.g. 'Paris'.
    '''
    return f"{city}: 21C"

A :class:ToolRegistry collects tools, exposes their schemas for the model, and doubles as a :class:~agentix.executors.ToolExecutor for the loop.

Tool ¶

Tool(
    func: ToolFn,
    *,
    name: str,
    description: str,
    parameters: dict[str, Any],
)

A callable wrapped with the metadata the agent loop needs.

Stays callable, so a decorated function can still be invoked directly (handy in tests and for non-agent use).

schema `property` ¶

schema: ToolSchema

The provider-neutral schema dict (name, description, parameters).

ToolRegistry ¶

ToolRegistry(tools: Iterable[Tool | ToolFn] = ())

Collects tools; provides schemas for the model and acts as the executor for the loop (delegating to a :class:LocalToolExecutor).

tool ¶

tool(
    func: ToolFn | None = None,
    *,
    name: str | None = None,
    description: str | None = None,
) -> Any

Decorator that turns a function into a :class:Tool.

Usable bare (@tool) or with overrides (@tool(name="...", ...)). The schema is generated from type hints; the description defaults to the first paragraph of the docstring, and per-parameter descriptions are read from a Google-style Args: section.

agentix.tools¶