def convert_points_mev(self, energies_mev, dnde_mev):
# (a) convert to acutal units. gtlike spectra take energy in MeV, return flux in ph/cm^2/s/MeV
return self.convert_points(
units.tosympy(energies_mev, units.MeV),
units.tosympy(dnde_mev,
units.ph / units.cm**2 / units.s / units.MeV))
def get_dnde_error(spectrum, covariance_matrix, energies):
""" assume energies has energy units and return flux in flux units. """
energies = units.tonumpy(energies, units.MeV)
dnde_error = SpectrumPlotter.get_dnde_error_mev(
spectrum, covariance_matrix, energies)
return units.tosympy(dnde_error,
units.ph / units.cm**2 / units.s / units.MeV)
def plot(self, filename=None, axes=None, title=None,
fignum=None, figsize=(4,4), **kwargs):
""" Plot the upper limit. """
spectral_kwargs=dict(color='red',zorder=1.9)
spectral_kwargs.update(kwargs)
file_energy_units = units.fromstring(self.results['energy_units'])
emin = self.results['emin']*file_energy_units
emax = self.results['emax']*file_energy_units
if axes is None:
fig = P.figure(fignum,figsize)
axes = SpectralAxes(fig=fig,
rect=(0.22,0.15,0.75,0.8),
flux_units=self.flux_units,
energy_units=self.energy_units)
fig.add_axes(axes)
axes.set_xlim_units(emin,emax)
# plot the spectral model
spectrum = self.results['spectrum']
sp=SpectrumPlotter(axes=axes)
sp.plot(spectrum, emin=emin, emax=emax, **spectral_kwargs)
# plot in the middle an arrow pointing down
e_middle = units.tosympy([np.sqrt(self.results['emin']*self.results['emax'])],file_energy_units)
dnde = sp.get_dnde(spectrum, e_middle)
energies, e2_dnde = axes.convert_points(e_middle, dnde)
if 'autoscale' in spectral_kwargs: spectral_kwargs.pop('autoscale')
if 'label' in spectral_kwargs: spectral_kwargs.pop('label')
plot_points(x=energies, y=e2_dnde,
xlo=None, xhi=None,
y_lower_err=None, y_upper_err=None,
y_ul=e2_dnde,
significant=False,
axes=axes, **spectral_kwargs)
if title is not None: axes.set_title(title)
if filename is not None:
P.savefig(expandvars(filename))
return axes
def get_dnde_error(spectrum,covariance_matrix,energies):
""" assume energies has energy units and return flux in flux units. """
energies=units.tonumpy(energies,units.MeV)
dnde_error=SpectrumPlotter.get_dnde_error_mev(spectrum,covariance_matrix,energies)
return units.tosympy(dnde_error,units.ph/units.cm**2/units.s/units.MeV)
def get_dnde(spectrum,energies):
""" assume energies has energy units and return flux in flux units. """
energies=units.tonumpy(energies,units.MeV)
dnde=SpectrumPlotter.get_dnde_mev(spectrum,energies)
return units.tosympy(dnde,units.ph/units.cm**2/units.s/units.MeV)
def convert_points_mev(self, energies_mev, dnde_mev):
# (a) convert to acutal units. gtlike spectra take energy in MeV, return flux in ph/cm^2/s/MeV
return self.convert_points(units.tosympy(energies_mev,units.MeV),
units.tosympy(dnde_mev,units.ph/units.cm**2/units.s/units.MeV))
def __load__(self, sed_file):
""" Parse a HESS-style plain text SED file.
Note, there is no real consistency to
these files, so just try many
examples. """
lines = open(expandvars(sed_file)).readlines()
# strip whitespace + newline characters
lines = [line.strip() for line in lines]
# remove blank lines + comments
lines = [line for line in lines if line is not '' and line[0] != '#' ]
# easier to parse '/TeV/cm^2/s' than '/TeV cm^2 s'
lines = [line.replace('[/TeV cm^2 s]','[/TeV/cm^2/s]') for line in lines]
# split on empty spaces
lines = [ line.split() for line in lines ]
energy_units='[TeV]'
flux_units='[/TeV/cm^2/s]'
header = lines[0]
units_line = lines[1]
for u in units_line:
if u not in [energy_units, flux_units]:
raise Exception("%s is an unrecognized unit" % u)
# get all the data lines
data = lines[2:]
# convert to float + transpose rows + convert to numpy
data = [ map(float,i) for i in data]
data = zip(*data)
data = [np.asarray(i,dtype=float) for i in data]
if header == ['Energy','Flux','Flux','Error_low','Flux','Error_high']:
# This is like SEDs from the galactic plane survey
# * http://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/ApJ_636.html
self.energy, self.dnde, dnde_lower, dnde_upper = data
self.dnde_err = (dnde_lower + dnde_upper)/2
# lines are energy, flux, flux_lower, flux_upper
elif header == ['Energy','interval','Mean','energy','Flux','Flux','Error']:
# This is like the SED for HESS J1303-631
# * http://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/AA439_1013.html
print 'data',data
self.lower_energy, self.upper_energy, self.energy, self.dnde, self.dnde_err = data
elif header == ['Mean','energy','Flux','Flux','Error']:
# This is like the SED for Vela X
# * http://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/VelaX_auxinfo.html
self.energy, self.dnde, self.dnde_err = data
elif header == ['Energy','interval','Mean','energy','Flux','Flux','Error']:
# This is like the SED for MSH 15-52
# * http://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/AA435_L17.html
self.lower_energy, self.upper_energy, self.dnde, self.dnde_err = data
elif header == ['Mean','energy','Flux','Flux','Error','(','-','/','+',')']:
# This is like the SED for the Kookaburra
# * http://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/kookaburra_auxinfo.html
raise Exception("not implemented")
if self.energy is None:
self.energy = np.sqrt(self.lower_energy*self.upper_energy)
if self.lower_energy is None or self.upper_energy is None:
# if only energy is given, assume no range in energy
self.lower_energy = self.energy
self.upper_energy = self.energy
self.significant = np.ones_like(self.energy).astype(bool)
def estimate_flux(dnde,energy,emin,emax, e_weight):
""" estimate the emin to emax flux for a source with prefactor
dnde at the given energy. assuming the source has a
spectral index of 2.
Note, for our situation dnde has units [ph/cm^2/s/TeV]
and energy has units [TeV], but the ouptut of the
i_flux function is correct. If e_weight=0,
the return has units [ph/cm^2/s]. If e_weight=1,
the return has units [TeV/cm^2/s]. """
model = PowerLaw(index=2, norm=dnde, e0=energy)
return model.i_flux(emin=emin, emax=emax, e_weight=e_weight)
estimate_flux = np.vectorize(estimate_flux)
self.flux = estimate_flux(self.dnde, self.energy, self.lower_energy, self.upper_energy, 0)
self.eflux = estimate_flux(self.dnde, self.energy, self.lower_energy, self.upper_energy, 1)
# Note, we can use this shortcut to comptue the errors,
# But ONLY because the function flux(dnde) is linear in dnde and dnde is the only
# parameter with an error: flux=dnde*f(gamma,e0) so flux_err=dnde_err*f(gamma,e0)
self.flux_err = estimate_flux(self.dnde_err, self.energy, self.lower_energy, self.upper_energy, 0)
self.eflux_err = estimate_flux(self.dnde_err, self.energy, self.lower_energy, self.upper_energy, 1)
# no upper limits
self.dnde_ul=np.nan*self.dnde
self.flux_ul=np.nan*self.flux
self.eflux_ul=np.nan*self.eflux
for values, u in [
[['lower_energy', 'upper_energy', 'energy'], units.TeV],
[['dnde', 'dnde_err', 'dnde_ul'], units.ph/units.cm**2/units.s/units.TeV],
[['flux', 'flux_err', 'flux_ul'], units.ph/units.cm**2/units.s],
[['eflux','eflux_err', 'eflux_ul'], units.TeV/units.cm**2/units.s]]:
for v in values:
self.__dict__[v] = units.tosympy(self.__dict__[v], u)
# all hess points are significant (as far as I can tell)
self.significant = np.ones(len(self.energy),dtype=bool)
print 'en',self.energy
print 'sig',self.significant
def get_dnde(spectrum, energies):
""" assume energies has energy units and return flux in flux units. """
energies = units.tonumpy(energies, units.MeV)
dnde = SpectrumPlotter.get_dnde_mev(spectrum, energies)
return units.tosympy(dnde,
units.ph / units.cm**2 / units.s / units.MeV)
def __load__(self, sed_file):
""" Parse a HESS-style plain text SED file.
Note, there is no real consistency to
these files, so just try many
examples. """
lines = open(expandvars(sed_file)).readlines()
# strip whitespace + newline characters
lines = [line.strip() for line in lines]
# remove blank lines + comments
lines = [line for line in lines if line is not '' and line[0] != '#']
# easier to parse '/TeV/cm^2/s' than '/TeV cm^2 s'
lines = [
line.replace('[/TeV cm^2 s]', '[/TeV/cm^2/s]') for line in lines
]
# split on empty spaces
lines = [line.split() for line in lines]
energy_units = '[TeV]'
flux_units = '[/TeV/cm^2/s]'
header = lines[0]
units_line = lines[1]
for u in units_line:
if u not in [energy_units, flux_units]:
raise Exception("%s is an unrecognized unit" % u)
# get all the data lines
data = lines[2:]
# convert to float + transpose rows + convert to numpy
data = [map(float, i) for i in data]
data = zip(*data)
data = [np.asarray(i, dtype=float) for i in data]
if header == [
'Energy', 'Flux', 'Flux', 'Error_low', 'Flux', 'Error_high'
]:
# This is like SEDs from the galactic plane survey
# * http://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/ApJ_636.html
self.energy, self.dnde, dnde_lower, dnde_upper = data
self.dnde_err = (dnde_lower + dnde_upper) / 2
# lines are energy, flux, flux_lower, flux_upper
elif header == [
'Energy', 'interval', 'Mean', 'energy', 'Flux', 'Flux', 'Error'
]:
# This is like the SED for HESS J1303-631
# * http://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/AA439_1013.html
print 'data', data
self.lower_energy, self.upper_energy, self.energy, self.dnde, self.dnde_err = data
elif header == ['Mean', 'energy', 'Flux', 'Flux', 'Error']:
# This is like the SED for Vela X
# * http://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/VelaX_auxinfo.html
self.energy, self.dnde, self.dnde_err = data
elif header == [
'Energy', 'interval', 'Mean', 'energy', 'Flux', 'Flux', 'Error'
]:
# This is like the SED for MSH 15-52
# * http://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/AA435_L17.html
self.lower_energy, self.upper_energy, self.dnde, self.dnde_err = data
elif header == [
'Mean', 'energy', 'Flux', 'Flux', 'Error', '(', '-', '/', '+',
')'
]:
# This is like the SED for the Kookaburra
# * http://www.mpi-hd.mpg.de/hfm/HESS/pages/publications/auxiliary/kookaburra_auxinfo.html
raise Exception("not implemented")
if self.energy is None:
self.energy = np.sqrt(self.lower_energy * self.upper_energy)
if self.lower_energy is None or self.upper_energy is None:
# if only energy is given, assume no range in energy
self.lower_energy = self.energy
self.upper_energy = self.energy
self.significant = np.ones_like(self.energy).astype(bool)
def estimate_flux(dnde, energy, emin, emax, e_weight):
""" estimate the emin to emax flux for a source with prefactor
dnde at the given energy. assuming the source has a
spectral index of 2.
Note, for our situation dnde has units [ph/cm^2/s/TeV]
and energy has units [TeV], but the ouptut of the
i_flux function is correct. If e_weight=0,
the return has units [ph/cm^2/s]. If e_weight=1,
the return has units [TeV/cm^2/s]. """
model = PowerLaw(index=2, norm=dnde, e0=energy)
return model.i_flux(emin=emin, emax=emax, e_weight=e_weight)
estimate_flux = np.vectorize(estimate_flux)
self.flux = estimate_flux(self.dnde, self.energy, self.lower_energy,
self.upper_energy, 0)
self.eflux = estimate_flux(self.dnde, self.energy, self.lower_energy,
self.upper_energy, 1)
# Note, we can use this shortcut to comptue the errors,
# But ONLY because the function flux(dnde) is linear in dnde and dnde is the only
# parameter with an error: flux=dnde*f(gamma,e0) so flux_err=dnde_err*f(gamma,e0)
self.flux_err = estimate_flux(self.dnde_err, self.energy,
self.lower_energy, self.upper_energy, 0)
self.eflux_err = estimate_flux(self.dnde_err, self.energy,
self.lower_energy, self.upper_energy, 1)
# no upper limits
self.dnde_ul = np.nan * self.dnde
self.flux_ul = np.nan * self.flux
self.eflux_ul = np.nan * self.eflux
for values, u in [[['lower_energy', 'upper_energy', 'energy'],
units.TeV],
[['dnde', 'dnde_err', 'dnde_ul'],
units.ph / units.cm**2 / units.s / units.TeV],
[['flux', 'flux_err', 'flux_ul'],
units.ph / units.cm**2 / units.s],
[['eflux', 'eflux_err', 'eflux_ul'],
units.TeV / units.cm**2 / units.s]]:
for v in values:
self.__dict__[v] = units.tosympy(self.__dict__[v], u)
# all hess points are significant (as far as I can tell)
self.significant = np.ones(len(self.energy), dtype=bool)
print 'en', self.energy
print 'sig', self.significant
