gwkokab.models.constraints

Provides implementation of various constraints using Constraint.

Two families live here. The ordering constraints (increasing, decreasing, their strict and positive variants) apply to a whole vector and are what keep an ordered set of hyper-parameters – successive Gaussian means, the break points of a broken power law – from swapping places during sampling. The mass constraints encode the sandwich \(m_\min \leq m_2 \leq m_1 \leq m_\max\) in either mass coordinates.

all_constraint and any_constraint combine constraints, the former slice-wise across the event axis, which is how JointDistribution derives its support from those of its marginals. transform_constraint pushes a support through a chain of transforms, which is what ExtendedSupportTransformedDistribution uses.

Public names are lower-case aliases of the private classes: the parameterised ones (mass_sandwich, all_constraint, …) alias the class itself and must be called, while the parameter-free ones (increasing_vector, …) alias a singleton instance.

Attributes

all_constraint
any_constraint
decreasing_vector
increasing_vector
mass_ratio_mass_sandwich
mass_sandwich
positive_decreasing_vector
positive_increasing_vector
strictly_decreasing_vector
strictly_increasing_vector

Module Contents

gwkokab.models.constraints.all_constraint

gwkokab.models.constraints.any_constraint

gwkokab.models.constraints.decreasing_vector

gwkokab.models.constraints.increasing_vector

gwkokab.models.constraints.mass_ratio_mass_sandwich

gwkokab.models.constraints.mass_sandwich

gwkokab.models.constraints.positive_decreasing_vector

gwkokab.models.constraints.positive_increasing_vector

gwkokab.models.constraints.strictly_decreasing_vector

gwkokab.models.constraints.strictly_increasing_vector