# Licensed under a 3-clause BSD style license - see LICENSE.rst
import logging
import numpy as np
from astropy.units import Quantity
from .effective_area import EffectiveAreaTable
from .energy_dispersion import EDispKernel
__all__ = ["IRFStacker"]
log = logging.getLogger(__name__)
[docs]class IRFStacker:
r"""
Stack instrument response functions.
Compute mean effective area and the mean energy dispersion for a given for a
given list of instrument response functions. Results are stored as
attributes.
The stacking of :math:`j` elements is implemented as follows. :math:`k`
and :math:`l` denote a bin in reconstructed and true energy, respectively.
.. math::
\epsilon_{jk} =\left\{\begin{array}{cl} 1, & \mbox{if
bin k is inside the energy thresholds}\\ 0, & \mbox{otherwise} \end{array}\right.
\overline{t} = \sum_{j} t_i
\overline{\mathrm{aeff}}_l = \frac{\sum_{j}\mathrm{aeff}_{jl}
\cdot t_j}{\overline{t}}
\overline{\mathrm{edisp}}_{kl} = \frac{\sum_{j} \mathrm{edisp}_{jkl}
\cdot \mathrm{aeff}_{jl} \cdot t_j \cdot \epsilon_{jk}}{\sum_{j} \mathrm{aeff}_{jl}
\cdot t_j}
Parameters
----------
list_aeff : list
list of `~gammapy.irf.EffectiveAreaTable`
list_livetime : list
list of `~astropy.units.Quantity` (livetime)
list_edisp : list
list of `~gammapy.irf.EDispKernel`
list_low_threshold : list
list of low energy threshold, optional for effective area mean computation
list_high_threshold : list
list of high energy threshold, optional for effective area mean computation
"""
def __init__(
self,
list_aeff,
list_livetime,
list_edisp=None,
list_low_threshold=None,
list_high_threshold=None,
):
self.list_aeff = list_aeff
self.list_livetime = Quantity(list_livetime)
self.list_edisp = list_edisp
self.list_low_threshold = list_low_threshold
self.list_high_threshold = list_high_threshold
self.stacked_aeff = None
self.stacked_edisp = None
[docs] def stack_aeff(self):
"""
Compute mean effective area (`~gammapy.irf.EffectiveAreaTable`).
"""
nbins = self.list_aeff[0].energy.nbin
aefft = Quantity(np.zeros(nbins), "cm2 s")
livetime_tot = np.sum(self.list_livetime)
for i, aeff in enumerate(self.list_aeff):
aeff_data = aeff.evaluate_fill_nan()
aefft_current = aeff_data * self.list_livetime[i]
aefft += aefft_current
stacked_data = aefft / livetime_tot
energy = self.list_aeff[0].energy.edges
self.stacked_aeff = EffectiveAreaTable(
energy_lo=energy[:-1], energy_hi=energy[1:], data=stacked_data.to("cm2")
)
[docs] def stack_edisp(self):
"""
Compute mean energy dispersion (`~gammapy.irf.EDispKernel`).
"""
reco_bins = self.list_edisp[0].e_reco.nbin
true_bins = self.list_edisp[0].e_true.nbin
aefft = Quantity(np.zeros(true_bins), "cm2 s")
temp = np.zeros(shape=(reco_bins, true_bins))
aefftedisp = Quantity(temp, "cm2 s")
for i, edisp in enumerate(self.list_edisp):
aeff_data = self.list_aeff[i].evaluate_fill_nan()
aefft_current = aeff_data * self.list_livetime[i]
aefft += aefft_current
edisp_data = edisp.pdf_in_safe_range(
self.list_low_threshold[i], self.list_high_threshold[i]
)
aefftedisp += edisp_data.transpose() * aefft_current
with np.errstate(divide="ignore", invalid="ignore"):
stacked_edisp = np.nan_to_num(aefftedisp / aefft)
e_true = self.list_edisp[0].e_true.edges
e_reco = self.list_edisp[0].e_reco.edges
self.stacked_edisp = EDispKernel(
e_true_lo=e_true[:-1],
e_true_hi=e_true[1:],
e_reco_lo=e_reco[:-1],
e_reco_hi=e_reco[1:],
data=stacked_edisp.transpose(),
)