MultiToolsManager

The MultiToolsManager class manages multiple tool registrations and lookups.

Description

MultiToolsManager provides a registry of tools keyed by function name, with support for enabling/disabling tools and conditional activation via enable_if. ToolsManager is the singleton subclass used by default.

Methods

• has_tool(name: str) -> bool: Whether a tool is registered and not disabled
• get_tool(name: str, default=None) -> ToolData | None: Get a tool by name; returns default if missing or disabled
• get_tool_meta(name: str, default=None) -> ToolFunctionSchema | None: Get a tool's metadata schema
• get_tool_func(name: str, default=None): Get a tool's implementation function
• get_tools() -> dict[str, ToolData]: All enabled tools
• tools_meta() -> dict[str, ToolFunctionSchema]: Metadata of all enabled tools
• tools_meta_dict(**kwargs) -> dict[str, dict]: Metadata dumped to dicts (kwargs forwarded to model_dump)
• register_tool(tool: ToolData) -> None: Register a tool; raises ValueError if the name already exists
• remove_tool(name: str) -> None: Unregister a tool
• enable_tool(name: str) -> None: Re-enable a disabled tool
• disable_tool(name: str) -> None: Disable a tool; raises ValueError if it does not exist
• get_disabled_tools() -> list[str]: Names of disabled tools

Example

from amrita_core.tools.manager import MultiToolsManager

manager = MultiToolsManager()
manager.register_tool(my_tool_data)
assert manager.has_tool("my_tool")
func = manager.get_tool_func("my_tool")