Source code for bxa.xspec.priors

#!/usr/bin/env python
# -*- coding: utf-8 -*- 

"""
BXA (Bayesian X-ray Analysis) for Xspec

Copyright: Johannes Buchner (C) 2013-2020

Priors
"""

from __future__ import print_function
from math import log10

# priors


def create_uniform_prior_for(model, par):
	"""
	Use for location variables (position)
	The uniform prior gives equal weight in non-logarithmic scale.
	"""
	pval, pdelta, pmin, pbottom, ptop, pmax = par.values
	print('  uniform prior for %s between %f and %f ' % (par.name, pmin, pmax))
	# TODO: should we use min/max or bottom/top?
	low = float(pmin)
	spread = float(pmax - pmin)
	if pmin > 0 and pmax / pmin > 100:
		print('   note: this parameter spans several dex. Should it be log-uniform (create_jeffreys_prior_for)?')
	def uniform_transform(x): return x * spread + low
	return dict(model=model, index=par._Parameter__index, name=par.name, 
		transform=uniform_transform, aftertransform=lambda x: x)





def create_jeffreys_prior_for(model, par):
	"""deprecated, use create_loguniform_prior_for instead. """
	return create_loguniform_prior_for(model, par)





def create_loguniform_prior_for(model, par):
	"""
	Use for scale variables (order of magnitude)
	The Jeffreys prior gives equal weight to each order of magnitude between the
	minimum and maximum value. Flat in logarithmic scale
	"""
	pval, pdelta, pmin, pbottom, ptop, pmax = par.values
	# TODO: should we use min/max or bottom/top?
	#print '  ', par.values
	print('  jeffreys prior for %s between %e and %e ' % (par.name, pmin, pmax))
	if pmin == 0:
		raise Exception('You forgot to set reasonable parameter limits on %s' % par.name)
	low = log10(pmin)
	spread = log10(pmax) - log10(pmin)
	if spread > 10:
		print('   note: this parameter spans *many* dex. Double-check the limits are reasonable.')
	def log_transform(x): return x * spread + low
	def log_after_transform(x): return 10**x
	return dict(model=model, index=par._Parameter__index, name='log(%s)' % par.name, 
		transform=log_transform, aftertransform=log_after_transform)





def create_gaussian_prior_for(model, par, mean, std):
	"""
	Use for informed variables.
	The Gaussian prior weights by a Gaussian in the parameter.
	"""
	import scipy.stats
	pval, pdelta, pmin, pbottom, ptop, pmax = par.values
	rv = scipy.stats.norm(mean, std)
	def gauss_transform(x): 
		return max(pmin, min(pmax, rv.ppf(x)))
	print('  gaussian prior for %s of %f +- %f' % (par.name, mean, std))
	return dict(model=model, index=par._Parameter__index, name=par.name, 
		transform=gauss_transform, aftertransform=lambda x: x)





def create_custom_prior_for(model, par, transform, aftertransform = lambda x: x):
	"""
	Pass your own prior weighting transformation
	"""
	print('  custom prior for %s' % (par.name))
	return dict(model=model, index=par._Parameter__index, name=par.name, 
		transform=transform, aftertransform=aftertransform)





def create_prior_function(transformations):
	"""
	Creates a single prior transformation function from parameter transformations
	"""

	def prior(cube, ndim, nparams):
		for i, t in enumerate(transformations):
			transform = t['transform']
			cube[i] = transform(cube[i])

	return prior