pisa.stages.cont_sys package

Submodules

pisa.stages.cont_sys.snowstorm_hist module

PISA stage to apply detector systematics to snowstorm simulation by splitting and histogramming the simulation set. The method is based on this paper: https://arxiv.org/pdf/1909.01530

pisa.stages.cont_sys.snowstorm_hist.init_test(**param_kwargs)[source]

Instantiation example

class pisa.stages.cont_sys.snowstorm_hist.snowstorm_hist(systematics, simulation_dists, simulation_dists_params, additional_params=None, tolerances=None, **std_kwargs)[source]

Bases: Stage

Service to apply detector systematics through splitting and histogramming of snowstorm simulation.

Expected container keys are:

"weights" + all detector systematics that should be used

The stage also expects that “regularized_output_binning” is part of the global auxilary variables of the data container set. This key is added by the utils.hist stage.

Parameters:

  • systematics (list of str) – List of the systematic parameters.

  • simulation_dists (list of str) – The distribution of the systematic parameter in the snowstorm simulation. Has to be either ‘gauss’ or ‘uniform’.

  • simulation_dists_params (list of tuples of floats) – Parameters of the simulation distributions. (mean, std) for ‘gauss’ and (min, max) for ‘uniform’.

  • additional_params (list of str (default None)) – Parameters that are no detector systematics but if changed require a re-calculation of the gradients (e.g. osc params).

  • tolerances (list of float (default None)) – Numerical tolerances for additional_params, to be given in the same order. If none are given (default), all gradients will be updated each time at least one parameter value changes (most accurate but slowest option). If tolerances are given, though, and all parameters move away from their initial values by less than the specified tolerances over the course of a single fit, gradients will never be re-calculated. Only once at least one parameter’s overall delta exceeds its corresponding tolerance the gradients will be updated (and a new reference point set). Subsequent modfications of the parameter vector will then be compared to this updated referenced point.

  • params (ParamSet) – Note that the params required to be in params are those listed in systematics plus those listed in additional_params.

apply_function()[source]

Implement in services (subclasses of Stage)

central_values

Central values of the systematic parameters in the snowstorm set.

compute_function()[source]

Implement in services (subclasses of Stage)

grads

Place to store gradients to save computing time.

setup_function()[source]

Implement in services (subclasses of Stage)

Module contents