A Julia interface to the python nested sampling library dynesty
This is built on PythonCall and imports a lot of the functionality of dynesty. There are some differences in the interface to make the code more "Julian".
Here we will sample a 5-dimensional Gaussian restricted to the domain [-10,10]
using Distributions
using Dynesty
# define a distribution
ndim = 5
d = MvNormal(ones(ndim))
loglikelihood(x) = logpdf(d, x)
prior_transform(p) = -10.0 .+ 20.0.*p
smplr = NestedSampler()
res = dysample(loglikelihood, prior_transform, ndim, smplr; dlogz=0.5)
# plot the results
cornerplot(res)