Instead of simply returning a NamedTuple containing NamedTuples, these methods now return NamedTuples (with keys result.mle and result.map) where the enclosed objects are custom structs that support random sampling from the posterior / likelihood approximation as rand(result.map, 10) for example.
Instead of simply returning a
NamedTuplecontainingNamedTuples, these methods now returnNamedTuples (with keysresult.mleandresult.map) where the enclosed objects are custom structs that support random sampling from the posterior / likelihood approximation asrand(result.map, 10)for example.