Nested Samplers

PolyChord

Supernest was historically designed around PolyChord and provided an OOP-like interface to take care of most of your housekeeping.

However it later became apparent that with a minimal modification to the API, one can extend supernest to work with other samplers. The issue is that for this to work, some of the housekeeping is delegated onto the user.

As per usual, you start by wrapping the prior functions:

from supernest import superimpose

super_n_dims, super_prior, super_like = superimpose(
	[(original_prior, original_log_like), (proposal_prior, proposal_loglike)],
	original_n_dims)

If any of this is confusing refer to the API documentation.

WARNING

This exact API is subject to change.

MultiNest

PyMultiNest has a very similar API to PolyChord, owing to it being a Python library.

from pymultinest import solve

solve(LogLikelihood=super_like, Prior=super_prior, n_dims=super_n_dims,
	  outputfiles=outputfiles)

WARNING

Integration with the FORTRAN framework MultiNest is planned.

WARNING

This doesn't exactly mean that everything shall work out of the box, PyMultiNest wasn't tested as extensively as PolyChord.

DyNesty

DANGER

DyNesty is not integrated.

DyPolyChord

DANGER

DyPolyChord is not integrated.

NestorFlow

DANGER

NestorFlow is not integrated.

Cosmological Inference Software

Cobaya

Historically the cobaya sampler framework has been integrated into the original supernest.

Unfortunately, that work was never upstreamed and now the work has to be done from scratch.

DANGER

Cobaya is not integrated.

CosmoChord

CosmoChord is a fork of CosmoMC that uses the PolyChord sampler in FORTRAN. As this is similar in scope to the integration with MultiNest, one should expect both to be handled at a similar time.

DANGER

CosmoChord is not integrated.