utils
functool_arg_to_jsonschematool_arg_to_jsonschema(spec, /, default_lang, fallback_lang) -> dict[str, ToolArgumentJSONSchema]Convert tool argument specification to JSON Schema format.
Transforms a tool argument specification into OpenAI-compatible JSON Schema format for LLM function calling.
Args: spec: Tool argument or object argument specification default_lang: Preferred language for description fallback_lang: Fallback language if default is unavailable
Returns: Dictionary mapping argument name to JSON Schema definition
Raises: InvalidSpecError: If description or type is invalid
paramspecToolArgument | ToolObjectArgumentparamdefault_langLanguageCodeparamfallback_langLanguageCodeReturns
dict[str, ToolArgumentJSONSchema]funcget_tool_arg_typeget_tool_arg_type(type_) -> str | list[str]Convert Python type names to JSON Schema type names.
Maps Python type annotations to their JSON Schema equivalents for use in OpenAI function calling schemas.
Args: type_: Python type name or list of type names
Returns: JSON Schema type name or list of type names
Raises: InvalidSpecError: If type is unsupported NotImplementedError: If array type is used (not yet supported)
paramtype_str | list[str]Returns
str | list[str]funcnoopnoop(*args, **kwargs) -> NoneNo-operation function used as fallback when strict mode is disabled.
Used when tool function loading fails but strict validation is off. Accepts any arguments and returns None.
Args: *args: Any positional arguments (ignored) **kwargs: Any keyword arguments (ignored)
paramargsAny= ()paramkwargsAny= {}Returns
NoneLast updated on