rpyc: difference between root.getmodule("module_name") and manually returning a module reference?

Question

I want to use a python module that is accessible on a remote rpyc server only. Is there a difference between the following two ways of accessing modules on a remote machine:


""" on the client side: """

1. my_local_mod_ref = my_rpyc_connection.root.getmodule("remote_module_name")
2. my_local_mod_ref = my_rpyc_connection.root.a_func_returning_the_module_ref()

""" on the server side: """

def exposed_a_func_returning_the_module_ref()
    import my_remote_module_name
    return my_remote_module_name

If there is a difference, which of the two alternatives is cleaner or preferable?

Answer 1

Here is the implementation of this getmodule:

def exposed_getmodule(self, name):
    """imports an arbitrary module"""
    return __import__(name, None, None, "*")

As you can see, if the module is not already loaded in the server, calling getmodule imports it, and (either way) a netref to the object module is returned.

If this matches the behavior of your a_func_returning_the_module_ref(), then there's no difference.

I'd guess getmodule is provided out-of-the-box, for being very useful, so you don't have to define it (or something similar) explicitly in order to achieve this goal.