How to handle libraries with type information in the docstrings?

[astronuttt]

I'm fairly new to static type checking and pyrefly. I've been using pyright in vscode with basic/standart type checking and there was very little issue with untyped libraries.
For example I'm using the vmware-vcenter library in my django application and the com.vmware package does not have any type hints or type stubs, but it has type information in the docstrings:

# (Some of the code unrelated to my question was removed)
class VM(VapiInterface):
    """
    The ``VM`` class provides methods for managing the lifecycle of a virtual
    machine.
    """
    ...
    def get(self,
            vm,
            ):
        """
        Returns information about a virtual machine.

        :type  vm: :class:`str`
        :param vm: Virtual machine identifier.
            The parameter must be an identifier for the resource type:
            ``VirtualMachine``.
        :rtype: :class:`VM.Info`
        :return: Information about the specified virtual machine.
        :raise: :class:`com.vmware.vapi.std.errors_client.Error` 
            if the system reports an error while responding to the request.
        ...
        """
        return self._invoke('get',
                            {
                            'vm': vm,
                            })

When using the library pyrefly can't infer the types and throws all kind of errors.

what is the recommended approach for handling this kind of issues in general? does pyrefly have any tools for doing something about this?

[yangdanny97]

If you're dealing with a lot of un-annotated code, try setting untyped-def-behavior to skip-and-infer-return-any in your config: https://pyrefly.org/en/docs/configuration/#untyped-def-behavior