ractogateway.tools.registry

Tool Registry — convert Python functions and Pydantic models into tool schemas.

Users should never hand-write nested JSON dicts for function-calling. Instead, they decorate plain Python functions with @tool or register Pydantic models directly. Each LLM adapter then translates the registry’s canonical schema into the provider-specific format it needs.

class ractogateway.tools.registry.ParamSchema(**data)

Bases: BaseModel

Schema for a single function parameter.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

name: str

type: str

description: str

required: bool

enum: list[str] | None

default: Any

model_config: ClassVar[ConfigDict] = {'arbitrary_types_allowed': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

class ractogateway.tools.registry.ToolSchema(**data)

Bases: BaseModel

Provider-agnostic canonical representation of a callable tool.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

name: str

description: str

parameters: list[ParamSchema]

to_json_schema()

Return an OpenAI-compatible JSON Schema parameters object.

Return type:

dict[str, Any]

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

ractogateway.tools.registry.tool(fn=None, *, name=None, description=None, registry=None)

Decorator that marks a function as an LLM-callable tool.

Can be used bare (@tool) or with overrides (@tool(name="my_tool", description="…")).

You can also bind registration directly: @tool(registry) or @tool(registry=registry).

The decorated function gains a _tool_schema attribute containing the canonical ToolSchema.

Return type:

Callable[..., Any]

class ractogateway.tools.registry.ToolRegistry

Bases: object

A registry that collects tools and exposes them as canonical schemas.

Usage:

registry = ToolRegistry()

@registry.register
def get_weather(city: str, unit: str = "celsius") -> str:
    '''Get the current weather for a city.'''
    ...

# Or register a Pydantic model:
registry.register(WeatherRequest)

# Iterate schemas:
for schema in registry.schemas:
    print(schema.name)

register(target=None, *, name=None, description=None)

Register a function or Pydantic model as a tool.

Works as a decorator (@registry.register) or as a direct call (registry.register(MyModel)).

Return type:

Callable[..., Any] | type[BaseModel]

property schemas: list[ToolSchema]

Return all registered tool schemas.

property tools: dict[str, ToolSchema]

Backward-compatible mapping of tool name -> schema.

get_schema(name)

Look up a single tool schema by name.

Return type:

ToolSchema | None

get_callable(name)

Look up the original callable by tool name.

Return type:

Callable[..., Any] | None