def test_wcsndmap_read_write(tmp_path, npix, binsz, frame, proj, skydir, axes):
geom = WcsGeom.create(npix=npix,
binsz=binsz,
proj=proj,
frame=frame,
axes=axes)
path = tmp_path / "tmp.fits"
m0 = WcsNDMap(geom)
fill_poisson(m0, mu=0.5)
m0.write(path, overwrite=True)
m1 = WcsNDMap.read(path)
m2 = Map.read(path)
m3 = Map.read(path, map_type="wcs")
assert_allclose(m0.data, m1.data)
assert_allclose(m0.data, m2.data)
assert_allclose(m0.data, m3.data)
m0.write(path, sparse=True, overwrite=True)
m1 = WcsNDMap.read(path)
m2 = Map.read(path)
m3 = Map.read(path, map_type="wcs")
assert_allclose(m0.data, m1.data)
assert_allclose(m0.data, m2.data)
assert_allclose(m0.data, m3.data)
# Specify alternate HDU name for IMAGE and BANDS table
m0.write(path, hdu="IMAGE", hdu_bands="TEST", overwrite=True)
m1 = WcsNDMap.read(path)
m2 = Map.read(path)
m3 = Map.read(path, map_type="wcs")
def test_wcsndmap_read_write(tmpdir, npix, binsz, coordsys, proj, skydir,
axes):
geom = WcsGeom.create(npix=npix,
binsz=binsz,
proj=proj,
coordsys=coordsys,
axes=axes)
filename = str(tmpdir / "map.fits")
m0 = WcsNDMap(geom)
fill_poisson(m0, mu=0.5)
m0.write(filename, overwrite=True)
m1 = WcsNDMap.read(filename)
m2 = Map.read(filename)
m3 = Map.read(filename, map_type="wcs")
assert_allclose(m0.data, m1.data)
assert_allclose(m0.data, m2.data)
assert_allclose(m0.data, m3.data)
m0.write(filename, sparse=True, overwrite=True)
m1 = WcsNDMap.read(filename)
m2 = Map.read(filename)
m3 = Map.read(filename, map_type="wcs")
assert_allclose(m0.data, m1.data)
assert_allclose(m0.data, m2.data)
assert_allclose(m0.data, m3.data)
# Specify alternate HDU name for IMAGE and BANDS table
m0.write(filename, hdu="IMAGE", hdu_bands="TEST", overwrite=True)
m1 = WcsNDMap.read(filename)
m2 = Map.read(filename)
m3 = Map.read(filename, map_type="wcs")
def images():
"""Load some `counts`, `counts_off`, `acceptance_on`, `acceptance_off" images"""
filename = "$GAMMAPY_DATA/tests/unbundled/hess/survey/hess_survey_snippet.fits.gz"
return {
"counts": WcsNDMap.read(filename, hdu="ON"),
"counts_off": WcsNDMap.read(filename, hdu="OFF"),
"acceptance": WcsNDMap.read(filename, hdu="ONEXPOSURE"),
"acceptance_off": WcsNDMap.read(filename, hdu="OFFEXPOSURE"),
"exposure": WcsNDMap.read(filename, hdu="EXPGAMMAMAP"),
}
def sky_image_plot(which):
filename = f"results/maps/{which}.fits.gz"
map = WcsNDMap.read(filename)
fig, ax, im = map.plot(stretch="sqrt")
filename = f"results/maps/{which}.png"
log.info(f"Writing {filename}")
fig.savefig(filename)
def sky_image_plot(which):
import matplotlib.pyplot as plt
# TODO: smooth and plot counts image
# TODO: overplot on and off regions
filename = f"results/skyimage/{which}.fits.gz"
map = WcsNDMap.read(filename)
fig, ax, im = map.plot(stretch="sqrt")
filename = f"results/skyimage/{which}.png"
log.info(f"Writing {filename}")
fig.savefig(filename)
def main():
usage = "usage: %(prog)s -c config.yaml"
description = "Run the analysis"
parser = argparse.ArgumentParser(usage=usage,description=description)
parser.add_argument('-c', '--conf', required = True)
parser.add_argument('--overwrite', required=False, default = 0,
help='Overwrite existing files', type=int)
parser.add_argument('--state', default = ['setup'],
choices = ['setup','avgspec','avgspec_ebl','lcbin', 'lcmonthly'],
help='Analysis state')
parser.add_argument('-i', required=False, default = 0,
help='Set local or scratch calculation', type=int)
parser.add_argument('--specin', required=False, default = -2.1,
help='Spectral index used for gtexposure in lieu of target in xml file',
type=float)
parser.add_argument('--addnewsrcs', default = 0,
help='Search for and add new sources and create residual map',
type=int)
parser.add_argument('--reloadfit', default = 0,
help='Reload ROI from avgspec xml file',
type=int)
parser.add_argument('--relocalize', default = 0,
help='Relocalize central source',
type=int)
parser.add_argument('--createsed', default = 0,
help='Create SED from best fit model',
type=int)
parser.add_argument('--forcespec', default = 0,
help='Recompute model parameters',
type=int)
parser.add_argument('--freezesupexp', default = 0,
help='freeze super exponential index parameters',
type=int)
parser.add_argument('--restorecatspec', default = 0,
help='Restore intitial catalog spectrum',
type=int)
parser.add_argument('--sethardexpcutoff', default = 0,
help='Manually change parameters of PL with SuperExpCutoff',
type=int)
parser.add_argument('--pivotE_free', default = 0,
help='let the pivot energy free during fit if spectrum is changed',
type=int)
parser.add_argument('--forcepl', default = 0,
help='Force the target source to have power-law shape',
type=int)
parser.add_argument('--profile2d', default = 0,
help='Compute 2D likelihood surface for PL index and normalization',
type=int)
parser.add_argument('--srcprob', default = 0,
help='Calculate the source probability for the photons,' \
' only works when no sub orbit time scales are used',
type=int)
parser.add_argument('--psf', default = 0,
help='Calculate the psf',
type=int)
parser.add_argument('--make_plots', default = 0, type=int,
help='Create plots',
)
parser.add_argument('--drm', default = 0,
help='Calculate the detector response matrix',
type=int)
args = parser.parse_args()
gta, config, fit_config, job_id = setup.init_gta(args.conf, i = args.i, logging_level = "INFO")
logging.info('Running fermipy setup')
if args.make_plots:
init_matplotlib_backend()
if args.reloadfit:
files = [fn for fn in glob(fit_config['avgspec']) if fn.find('.xml') > 0 or fn.find('.npy') > 0]
if len(files):
utils.copy2scratch(files, gta.workdir)
else:
logging.error("No files found in {0:s}".format(fit_config['avgspec']))
args.reloadfit = False
if args.state == 'lcbin':
pps = PreparePointSource(config,logging={'verbosity' : 3})
pps.setup()
pps._bin_data_lc(overwrite = args.overwrite)
pps._compute_lc_exp(overwrite = args.overwrite,
specin = args.specin)
return pps
elif args.state == 'setup':
try:
gta.setup()
except RuntimeError as e:
logging.error("setup ended with runtime error:\n{0}.".format(e))
if args.psf:
logging.info("Running psf")
gta.compute_psf(overwrite = True)
if args.drm:
logging.info("Running drm")
gta.compute_drm(overwrite = True)
return None, gta, fit_config
elif args.state.find('avgspec') >= 0:
gta.setup()
if args.psf:
logging.info("Running psf")
gta.compute_psf(overwrite = True)
if args.drm:
logging.info("Running drm")
gta.compute_drm(overwrite = True)
if not type(config['selection']['target']) == str:
# target name not given
# take closest source to ROI center if separation
# is less then 0.1 degree
logging.warning("Target name is {0}".format(config['selection']['target']))
sep = gta.roi.sources[0]['offset']
logging.warning("Source closets to ROI center is {0:.3f} degree away".format(sep))
if sep < 0.1:
config['selection']['target'] = gta.roi.sources[0]['name']
logging.info("Set target to {0:s}".format(config['selection']['target']))
else: # add source at center of ROI
csrc = SkyCoord(ra = config['selection']['ra'],
dec = config['selection']['dec'], frame = 'fk5', unit = 'degree')
if csrc.dec.value < 0.:
sign = '-'
else:
sign = '+'
newname = 'j{0:02.0f}{1:02.0f}{2:s}{3:02.0f}{4:02.0f}'.format(
csrc.ra.hms.h, csrc.ra.hms.m, sign,
np.abs(csrc.dec.dms.d), np.abs(csrc.dec.dms.m))
gta.add_source(newname,{
'ra' : config['selection']['ra'], 'dec' : config['selection']['dec'],
'SpectrumType' : 'PowerLaw', 'Index' : fit_config['new_src_pl_index'],
'Scale' : fit_config['pivotE'] if 'pivotE' in fit_config.keys() else 1000.,
'Prefactor' : 1e-11,
'SpatialModel' : 'PointSource' })
config['selection']['target'] = newname
logging.info("Set target to {0:s}".format(config['selection']['target']))
if args.reloadfit:
# save old spectrum
spec_cat = gta.roi.get_source_by_name(config['selection']['target'])
try:
gta.load_roi(args.state) # reload the average spectral fit
except:
logging.error("Could not reload fit. Continuing anyway.")
logging.info('Running fermipy optimize and fit')
# gives "failed to create spline" in get_parameter_limits function
if 'source_spec' in fit_config.keys():
m = gta.roi.get_source_by_name(config['selection']['target'])
if not m['SpectrumType'] == fit_config['source_spec'] or args.forcespec:
if fit_config['source_spec'] == 'PowerLaw':
gta, _ = set_src_spec_pl(gta, gta.get_source_name(config['selection']['target']),
fit_config['pivotE'] if 'pivotE' in fit_config.keys() else None,
e0_free = args.pivotE_free)
elif fit_config['source_spec'] == 'PLSuperExpCutoff':
gta, _ = set_src_spec_plexpcut(gta, gta.get_source_name(config['selection']['target']),
fit_config['pivotE'] if 'pivotE' in fit_config.keys() else None,
e0_free = args.pivotE_free)
#elif fit_config['source_spec'] == 'LogParabola':
#gta = set_src_spec_lp(gta, gta.get_source_name(config['selection']['target']))
else:
logging.warning("Spectrum {0:s} not supported, spectrum not changed".format(fit_config['source_spec']))
# restore spectrum from catalog
if args.restorecatspec:
#for k in ['alpha','Index','Index1']:
#if k in spec_cat.spectral_pars.keys():
#spec_cat.spectral_pars[k]['value'] -= 0.5
#spec_cat.spectral_pars['Eb']['free'] = True
#spec_cat.spectral_pars['Eb']['value'] = 2000.
#spec_cat.spectral_pars['alpha']['value'] = 1.8
#print spec_cat.spectral_pars
gta.set_source_spectrum(config['selection']['target'],
spectrum_type = spec_cat['SpectrumType'],
spectrum_pars= spec_cat.spectral_pars)
logging.info("restored catalog spectrum")
# for some sources modeled with PL with super exponential cutoff I
# have to do this to get a nice SED, but not for 3C454.3!
if gta.roi.get_source_by_name(config['selection']['target'])['SpectrumType'] ==\
'PLSuperExpCutoff' and args.sethardexpcutoff :
pars = {}
old_spec_pars = copy.deepcopy(gta.roi.get_source_by_name(config['selection']['target']))
for k in ['Prefactor','Scale','Index1','Index2','Cutoff']:
pars[k] = old_spec_pars.spectral_pars[k]
if config['selection']['target'] == '3C454.3':
# values from Romoli et al 2017
pars['Prefactor']['value'] = 4.7
pars['Index1']['value'] = 1.87
pars['Index2']['value'] = 0.4
pars['Cutoff']['value'] = 1100.
pars['Cutoff']['scale'] = 1.
pars['Cutoff']['min'] = 100.
pars['Cutoff']['max'] = 10000.
else:
pars['Index1']['value'] = 1.8
pars['Index2']['value'] = 1.
pars['Cutoff']['value'] = 5e4
gta.set_source_spectrum(config['selection']['target'],
# spectrum_type = 'PLSuperExpCutoff2',
spectrum_type = 'PLSuperExpCutoff',
spectrum_pars= pars)
logging.info("changed spectral parameters to {0}".format(
gta.roi.get_source_by_name(config['selection']['target']).spectral_pars))
else:
old_spec_pars = None
if args.forcepl and not \
gta.roi.get_source_by_name(config['selection']['target'])['SpectrumType'] == 'PowerLaw':
gta, _ = set_src_spec_pl(gta, gta.get_source_name(config['selection']['target']))
args.state += "_pl"
if args.state.find('ebl') >= 0:
gta = fa.utils.add_ebl_atten(gta,config['selection']['target'],fit_config['z'])
gta = set_free_pars_avg(gta, fit_config, freezesupexp = args.freezesupexp)
f,gta = fit_with_retries(gta, fit_config, config['selection']['target'],
alt_spec_pars = old_spec_pars)
logging.debug(f)
try:
get_best_fit_covar(gta, config['selection']['target'], prefix = args.state)
except IndexError:
logging.error("Covariance matrix calculation failed")
#relocalize central source and refit
if args.relocalize and type(config['selection']['target']) == str:
loc = gta.localize(config['selection']['target'], make_plots=args.make_plots,
free_background = fit_config['reloc_bkg'],
free_radius = fit_config['reloc_rad'],
update=True)
logging.info('new position is {0[pos_offset]:.3f} degrees from old position'.format(loc))
logging.info('Pos uncertainty is {0[pos_r68]:.3f} (68%); {0[pos_r95]:.3f} (95%) degrees'.format(loc))
logging.info('Refitting with new source position ...')
logging.info('free source parameters:')
for s in gta.get_sources() :
for k in s.spectral_pars.keys():
if s.spectral_pars[k]['free']:
logging.info('{0:s}: {1:s}'.format(s.name, k))
f = gta.fit()
#f,gta = fit_with_retries(gta, fit_config, config['selection']['target'])
gta.print_roi()
gta.write_roi(args.state)
elif args.state == 'lcmonthly':
gta.setup()
gta.load_roi('avgspec') # reload the average spectral fit
logging.info('Running the 30-day bin' + \
'light curve for {0[target]:s}'.format(config['selection']))
lc = gta.lightcurve(config['selection']['target'],
binsz = 30. * 24. * 60. * 60.)
model = {'Scale': 1000., 'Index' : fit_config['new_src_pl_index'], 'SpatialModel' : 'PointSource'}
if args.addnewsrcs:
gta.load_roi(args.state) # reload the average spectral fit
max_sqrt_ts = 1000.
irun = 0
# define the test source
#model = {'Index' : 2.0, 'SpatialModel' : 'PointSource'}
# run ts map and add new sources with sqrt(ts) > 5
# reoptimize iteratively for each new source
# this is only done for outer RoI
while max_sqrt_ts >= fit_config['max_sqrt_ts']:
# run ts and residual maps
ts_maps = gta.tsmap(args.state,model=model,
write_fits = True, write_npy = True, make_plots = args.make_plots)
# get the skydirs
#coords = ts_maps['sqrt_ts'].get_pixel_skydirs()
coords = ts_maps['sqrt_ts'].geom.get_coord()
if ts_maps['sqrt_ts'].geom.coordsys == 'CEL':
frame = 'fk5'
elif ts_maps['sqrt_ts'].geom.coordsys == 'GAL':
frame = 'galactic'
c = SkyCoord(coords[0], coords[1], unit = 'deg',
frame = frame)
#sqrt_ts = ts_maps['sqrt_ts'].get_map_values(coords.ra, coords.dec) # these are all nans. workaround: load fits file
#sqrt_ts_map = Map.create_from_fits(path.join(gta.workdir,ts_maps['file']),hdu = 'SQRT_TS_MAP')
#n_map = Map.create_from_fits(path.join(gta.workdir,ts_maps['file']),hdu = 'N_MAP')
sqrt_ts_map = WcsNDMap.read(path.join(gta.workdir,ts_maps['file']),hdu = 'SQRT_TS_MAP')
n_map = WcsNDMap.read(path.join(gta.workdir,ts_maps['file']),hdu = 'N_MAP')
sqrt_ts = sqrt_ts_map.data
amplitudes = n_map.data
# get the angular separation from RoI center
sep = gta.roi.skydir.separation(c)
# mask nan values and pixels close to central source
m = np.isfinite(sqrt_ts) & (sep.value > fit_config['new_src_search_rad'])
if not np.sum(m):
logging.warning('No pixels that are finite at distance > {0[new_src_search_rad]:.2f}'.format(fit_config))
raise RuntimeError
# get max ts value
max_sqrt_ts = np.max(sqrt_ts[m])
if max_sqrt_ts < fit_config['max_sqrt_ts']: break
# get the coords of max ts
idx = np.argmax(sqrt_ts[m])
logging.info('Found new source with sqrt(ts) = {0:.2f} at ra,dec = {1:.2f}, {2:.2f}'.format(
sqrt_ts[m][idx], c.ra[m][idx].value, c.dec[m][idx].value))
# add a new source
csrc = SkyCoord(ra = c.ra[m][idx], dec = c.dec[m][idx], frame = 'fk5')
if csrc.dec.value < 0.:
sign = '-'
else:
sign = '+'
newname = 'j{0:02.0f}{1:02.0f}{2:s}{3:02.0f}{4:02.0f}'.format(
csrc.ra.hms.h, csrc.ra.hms.m, sign,
np.abs(csrc.dec.dms.d), np.abs(csrc.dec.dms.m))
gta.add_source(newname,{
'ra' : c.ra[m][idx].value, 'dec' : c.dec[m][idx].value,
'SpectrumType' : 'PowerLaw', 'Index' : fit_config['new_src_pl_index'],
'Scale' : 1000, 'Prefactor' : amplitudes[m][idx],
'SpatialModel' : 'PointSource' })
logging.debug('Amplitude of source: {0}'.format(amplitudes[m][idx]))
gta.free_source(newname, pars=['norm','index'], free = True)
f = gta.fit()
gta.print_roi()
irun += 1
# if new sources where added, save output
if irun > 0:
gta.print_roi()
# gta = reset_diff_filenames(gta)
# refit the model with new sources present
gta = set_free_pars_avg(gta, fit_config)
f,gta = fit_with_retries(gta, fit_config, config['selection']['target'])
gta.write_roi(args.state)
else:
ts_maps = gta.tsmap(args.state,model=model,
write_fits = True, write_npy = True, make_plots = args.make_plots)
try:
resid_maps = gta.residmap(args.state,model=model, make_plots=args.make_plots, write_fits = True, write_npy = True)
except:
logging.error("Residual map computation and plotting failed")
if args.profile2d:
compute_profile2d(gta, config['selection']['target'], prefix = args.state,
sigma = 5., xsteps = 30, ysteps = 31)
if args.createsed:
if fit_config.get('force_free_index', False):
gta.free_index(config['selection']['target'], free = False)
gta.free_index(config['selection']['target'], free = True)
gta.load_roi(args.state) # reload the average spectral fit
logging.info('Running sed for {0[target]:s}'.format(config['selection']))
sed = gta.sed(config['selection']['target'],
prefix = args.state,
#outfile = 'sed.fits',
#free_radius = #sed_config['free_radius'],
#free_background= #sed_config['free_background'],
#free_pars = fa.allnorm,
make_plots = args.make_plots,
#cov_scale = sed_config['cov_scale'],
#use_local_index = sed_config['use_local_index'],
#use_local_index = True, # sed_config['use_local_index'],
#bin_index = sed_config['bin_index']
)
logging.info("SED covariance: {0}".format(sed['param_covariance']))
# generate additional SEDs
for src in fit_config.get('additional_seds', []):
sed = gta.sed(src,
prefix = args.state,
#outfile = 'sed.fits',
#free_radius = #sed_config['free_radius'],
#free_background= #sed_config['free_background'],
#free_pars = fa.allnorm,
make_plots = args.make_plots,
#cov_scale = sed_config['cov_scale'],
#use_local_index = sed_config['use_local_index'],
#use_local_index = True, # sed_config['use_local_index'],
#bin_index = sed_config['bin_index']
)
if args.srcprob:
logging.info("Running srcprob with srcmdl {0:s}".format(args.state))
gta.compute_srcprob(xmlfile=args.state, overwrite = True)
return f, gta, fit_config
