Turning an arbitrary-precision float from this library back into a double

Near as I can tell, there is no function that does this. Is this intentional? Knowing the risks of truncation, how would I go about doing this kind of transformation?