ObservationEventSampler

class gammapy.datasets.ObservationEventSampler

Bases: MapDatasetEventSampler

Sample event lists for a given observation and signal models.

Signal events are sampled from the predicted counts distribution given by the product of the sky models and the expected exposure. They are then folded with the instrument response functions. To improve performance, IRFs are evaluated on a pre-defined binning, not at each individual event energy / coordinate.

Parameters:

random_state{int, ‘random-seed’, ‘global-rng’, RandomState}, optional

Defines random number generator initialisation via the get_random_state function.

oversample_energy_factorint, optional

Defines an oversampling factor for the energies; it is used only when sampling an energy-dependent time-varying source. Default is 10.

t_deltaQuantity, optional

Time interval used to sample the time-dependent source. Default is 0.5 s.

keep_mc_idbool, optional

Flag to tag sampled events from a given model with a Montecarlo identifier. Default is True. If set to False, no identifier will be assigned.

n_event_bunchint

Size of events bunches to sample. If None, sample all events in memory. Default is 10000.

dataset_kwargsdict, optional

Arguments passed to create_map_dataset_from_observation()

Methods Summary

run(observation[, models, dataset_name])

Sample events for given observation and signal models.

Methods Documentation

run(observation, models=None, dataset_name=None)

Sample events for given observation and signal models.

The signal distribution is sampled from the given models in true coordinates and energy. The true quantities are then folded with the IRFs to obtain the observable quantities.

Parameters:

observationObservation

Observation to be simulated.

modelsModels, optional

Models to simulate. Can be None to only sample background events. Default is None.

dataset_namestr, optional

If models contains one or multiple FoVBackgroundModel it should match the dataset_name of the background model to use. Default is None.

Returns:

observationObservation

A copy of the input observation with event list filled.

__init__(random_state='random-seed', oversample_energy_factor=10, t_delta=<Quantity 0.5 s>, keep_mc_id=True, n_event_bunch=10000, dataset_kwargs=None)

classmethod __new__(*args, **kwargs)