# Copyright (C) 2017 Christopher M. Biwer
# This program is free software; you can redistribute it and/or modify it
# under the terms of the GNU General Public License as published by the
# Free Software Foundation; either version 3 of the License, or (at your
# option) any later version.
#
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
# Public License for more details.
#
# You should have received a copy of the GNU General Public License along
# with this program; if not, write to the Free Software Foundation, Inc.,
# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
"""
This modules provides classes for evaluating multi-dimensional constraints.
"""
import logging
import re
import scipy.spatial
import numpy
import h5py
from pycbc import transforms
from pycbc.io import record
logger = logging.getLogger('pycbc.distributions.constraints')
[docs]class Constraint(object):
"""Creates a constraint that evaluates to True if parameters obey
the constraint and False if they do not.
"""
name = "custom"
def __init__(self, constraint_arg, static_args=None, transforms=None,
**kwargs):
static_args = (
{} if static_args is None
else dict(sorted(
static_args.items(), key=lambda x: len(x[0]), reverse=True))
)
for arg, val in static_args.items():
swp = f"'{val}'" if isinstance(val, str) else str(val)
# Substitute static arg name for value if it appears in the
# constraint_arg string at the beginning of a word and is not
# followed by an underscore or equals sign.
# This ensures that static_args that are also kwargs in function calls are
# handled correctly, i.e., the kwarg is not touched while its value is replaced
# with the static_arg value.
constraint_arg = re.sub(
r'\b{}(?!\_|\=)'.format(arg), swp, constraint_arg)
self.constraint_arg = constraint_arg
self.transforms = transforms
for kwarg in kwargs.keys():
setattr(self, kwarg, kwargs[kwarg])
def __call__(self, params):
"""Evaluates constraint.
"""
# cast to FieldArray
if isinstance(params, dict):
params = record.FieldArray.from_kwargs(**params)
elif not isinstance(params, record.FieldArray):
raise ValueError("params must be dict or FieldArray instance")
# try to evaluate; this will assume that all of the needed parameters
# for the constraint exists in params
try:
out = self._constraint(params)
except NameError:
# one or more needed parameters don't exist; try applying the
# transforms
params = transforms.apply_transforms(params, self.transforms) \
if self.transforms else params
out = self._constraint(params)
if isinstance(out, record.FieldArray):
out = out.item() if params.size == 1 else out
return out
def _constraint(self, params):
""" Evaluates constraint function.
"""
return params[self.constraint_arg]
[docs]class SupernovaeConvexHull(Constraint):
"""Pre defined constraint for core-collapse waveforms that checks
whether a given set of coefficients lie within the convex hull of
the coefficients of the principal component basis vectors.
"""
name = "supernovae_convex_hull"
required_parameters = ["coeff_0", "coeff_1"]
def __init__(self, constraint_arg, transforms=None, **kwargs):
super(SupernovaeConvexHull,
self).__init__(constraint_arg, transforms=transforms, **kwargs)
if 'principal_components_file' in kwargs:
pc_filename = kwargs['principal_components_file']
hull_dimention = numpy.array(kwargs['hull_dimention'])
self.hull_dimention = int(hull_dimention)
pc_file = h5py.File(pc_filename, 'r')
pc_coefficients = numpy.array(pc_file.get('coefficients'))
pc_file.close()
hull_points = []
for dim in range(self.hull_dimention):
hull_points.append(pc_coefficients[:, dim])
hull_points = numpy.array(hull_points).T
pc_coeffs_hull = scipy.spatial.Delaunay(hull_points)
self._hull = pc_coeffs_hull
def _constraint(self, params):
output_array = []
points = numpy.array([params["coeff_0"],
params["coeff_1"],
params["coeff_2"]])
for coeff_index in range(len(params["coeff_0"])):
point = points[:, coeff_index][:self.hull_dimention]
output_array.append(self._hull.find_simplex(point) >= 0)
return numpy.array(output_array)
# list of all constraints
constraints = {
Constraint.name : Constraint,
SupernovaeConvexHull.name : SupernovaeConvexHull,
}