rxmc.constraint.Constraint

class rxmc.constraint.Constraint(observations: list[Observation], physical_model: PhysicalModel, likelihood_model: LikelihoodModel)

Bases: object

Pair observations with a physical model and a likelihood model.

A Constraint is the composition of one or more Observation objects with a PhysicalModel and a LikelihoodModel. It acts as a box that accepts model parameters and returns the log likelihood or other statistics, acting as a constraint on those parameters.

Parameters:

• observations (list of Observation) – The observed data that the model will attempt to reproduce.

• physical_model (PhysicalModel) – Model that predicts the observed data.

• likelihood_model (LikelihoodModel) – Model that defines the likelihood of the observations given the physical-model predictions.

__init__(observations: list[Observation], physical_model: PhysicalModel, likelihood_model: LikelihoodModel)

Methods

__init__(observations, physical_model, ...)
chi2(model_params[, likelihood_params])	Generalised chi-squared (Mahalanobis distance) summed over observations.
empirical_coverage(ylow, yhigh[, xlim])	Fraction of data points within a predictive interval.
log_likelihood(model_params[, likelihood_params])	Total log likelihood over all observations.
marginal_log_likelihood(ym, *likelihood_params)	Log likelihood given pre-computed model predictions.
model(model_params)	Compute the physical model output for each observation.
num_pts_within_interval(ylow, yhigh[, xlim])	Count data points that fall within a predictive interval.
predict(*model_params)	Generate predictions for each observation.

model(model_params)

Compute the physical model output for each observation.

Parameters:

model_params (tuple) – Parameters of the physical model.

Returns:

list of np.ndarray – Model predictions, one array per observation.

log_likelihood(model_params, likelihood_params=())

Total log likelihood over all observations.

Parameters:

• model_params (tuple) – Parameters of the physical model.

• likelihood_params (tuple, optional) – Additional parameters for the likelihood model.

Returns:

float – Sum of log likelihoods across all observations.

marginal_log_likelihood(ym: list, *likelihood_params)

Log likelihood given pre-computed model predictions.

Parameters:

• ym (list of np.ndarray) – Pre-computed model predictions for each observation.

• *likelihood_params (float) – Additional parameters for the likelihood model.

Returns:

float – Sum of log likelihoods across all observations.

chi2(model_params, likelihood_params=())

Generalised chi-squared (Mahalanobis distance) summed over observations.

Parameters:

• model_params (tuple) – Parameters of the physical model.

• likelihood_params (tuple, optional) – Additional parameters for the likelihood model.

Returns:

float – Total chi-squared statistic.

predict(*model_params)

Generate predictions for each observation.

Parameters:

*model_params (float) – Parameters of the physical model.

Returns:

list of np.ndarray – Predicted values for each observation.

num_pts_within_interval(ylow: list[ndarray], yhigh: list[ndarray], xlim=None)

Count data points that fall within a predictive interval.

Parameters:

• ylow (list of np.ndarray) – Lower bounds of the interval for each observation.

• yhigh (list of np.ndarray) – Upper bounds of the interval for each observation.

• xlim (tuple, optional) – (x_min, x_max) range to restrict the count.

Returns:

int – Total number of points within the interval across all observations.

empirical_coverage(ylow: list[ndarray], yhigh: list[ndarray], xlim=None)

Fraction of data points within a predictive interval.

Parameters:

• ylow (list of np.ndarray) – Lower bounds of the interval for each observation.

• yhigh (list of np.ndarray) – Upper bounds of the interval for each observation.

• xlim (tuple, optional) – (x_min, x_max) range to restrict the count.

Returns:

float – Empirical coverage fraction in [0, 1].