# Licensed under a 3-clause BSD style license - see LICENSE.rst
import logging
import numpy as np
from astropy import units as u
from regions import CircleSkyRegion
from gammapy.maps import WcsGeom
from .core import CountsSpectrum
from .dataset import SpectrumDataset
__all__ = ["SpectrumDatasetMaker"]
log = logging.getLogger(__name__)
[docs]class SpectrumDatasetMaker:
"""Make spectrum for a single IACT observation.
The irfs and background are computed at a single fixed offset,
which is recommend only for point-sources.
Parameters
----------
containment_correction : bool
Apply containment correction for point sources and circular on regions.
selection : list
List of str, selecting which maps to make.
Available: 'counts', 'aeff', 'background', 'edisp'
By default, all spectra are made.
"""
available_selection = ["counts", "background", "aeff", "edisp"]
def __init__(self, containment_correction=False, selection=None):
self.containment_correction = containment_correction
if selection is None:
selection = self.available_selection
self.selection = selection
# TODO: move this to a RegionGeom class
[docs] @staticmethod
def geom_ref(region):
"""Reference geometry to project region"""
frame = region.center.frame.name
return WcsGeom.create(
skydir=region.center, npix=(1, 1), binsz=1, proj="TAN", frame=frame
)
[docs] def make_counts(self, region, energy_axis, observation):
"""Make counts.
Parameters
----------
region : `~regions.SkyRegion`
Region to compute counts spectrum for.
energy_axis : `~gammapy.maps.MapAxis`
Reconstructed energy axis.
observation: `~gammapy.data.DataStoreObservation`
Observation to compute effective area for.
Returns
-------
counts : `~gammapy.spectrum.CountsSpectrum`
Counts spectrum
"""
edges = energy_axis.edges
counts = CountsSpectrum(
energy_hi=edges[1:], energy_lo=edges[:-1], region=region
)
events_region = observation.events.select_region(
region, wcs=self.geom_ref(region).wcs
)
counts.fill_events(events_region)
return counts
[docs] @staticmethod
def make_background(region, energy_axis, observation):
"""Make background.
Parameters
----------
region : `~regions.SkyRegion`
Region to compute background spectrum for.
energy_axis : `~gammapy.maps.MapAxis`
Reconstructed energy axis.
observation: `~gammapy.data.DataStoreObservation`
Observation to compute effective area for.
Returns
-------
background : `~gammapy.spectrum.CountsSpectrum`
Background spectrum
"""
if not isinstance(region, CircleSkyRegion):
raise TypeError(
"Background computation only supported for circular regions."
)
offset = observation.pointing_radec.separation(region.center)
e_reco = energy_axis.edges
bkg = observation.bkg
data = bkg.evaluate_integrate(
fov_lon=0 * u.deg, fov_lat=offset, energy_reco=e_reco
)
solid_angle = 2 * np.pi * (1 - np.cos(region.radius)) * u.sr
data *= solid_angle
data *= observation.observation_time_duration
return CountsSpectrum(
energy_hi=e_reco[1:], energy_lo=e_reco[:-1], data=data.to_value(""), unit=""
)
[docs] def make_aeff(self, region, energy_axis_true, observation):
"""Make effective area.
Parameters
----------
region : `~regions.SkyRegion`
Region to compute background effective area.
energy_axis_true : `~gammapy.maps.MapAxis`
True energy axis.
observation: `~gammapy.data.DataStoreObservation`
Observation to compute effective area for.
Returns
-------
aeff : `~gammapy.irf.EffectiveAreaTable`
Effective area table.
"""
offset = observation.pointing_radec.separation(region.center)
aeff = observation.aeff.to_effective_area_table(
offset, energy=energy_axis_true.edges
)
if self.containment_correction:
if not isinstance(region, CircleSkyRegion):
raise TypeError(
"Containment correction only supported for circular regions."
)
psf = observation.psf.to_energy_dependent_table_psf(theta=offset)
containment = psf.containment(aeff.energy.center, region.radius)
aeff.data.data *= containment.squeeze()
return aeff
[docs] @staticmethod
def make_edisp(position, energy_axis, energy_axis_true, observation):
"""Make energy dispersion.
Parameters
----------
position : `~astropy.coordinates.SkyCoord`
Position to compute energy dispersion for.
energy_axis : `~gammapy.maps.MapAxis`
Reconstructed energy axis.
energy_axis_true : `~gammapy.maps.MapAxis`
True energy axis.
observation: `~gammapy.data.DataStoreObservation`
Observation to compute edisp for.
Returns
-------
edisp : `~gammapy.irf.EDispKernel`
Energy dispersion
"""
offset = observation.pointing_radec.separation(position)
return observation.edisp.to_energy_dispersion(
offset, e_reco=energy_axis.edges, e_true=energy_axis_true.edges
)
[docs] def run(self, dataset, observation):
"""Make spectrum dataset.
Parameters
----------
dataset : `~gammapy.spectrum.SpectrumDataset`
Spectrum dataset.
observation: `~gammapy.data.DataStoreObservation`
Observation to reduce.
Returns
-------
dataset : `~gammapy.spectrum.SpectrumDataset`
Spectrum dataset.
"""
kwargs = {
"gti": observation.gti,
"livetime": observation.observation_live_time_duration,
}
energy_axis = dataset.counts.energy
energy_axis_true = dataset.aeff.data.axis("energy")
region = dataset.counts.region
if "counts" in self.selection:
kwargs["counts"] = self.make_counts(region, energy_axis, observation)
if "background" in self.selection:
kwargs["background"] = self.make_background(
region, energy_axis, observation
)
if "aeff" in self.selection:
kwargs["aeff"] = self.make_aeff(region, energy_axis_true, observation)
if "edisp" in self.selection:
kwargs["edisp"] = self.make_edisp(
region.center, energy_axis, energy_axis_true, observation
)
return SpectrumDataset(name=dataset.name, **kwargs)