inj_opts=inj_opts,
random_state=rndgen)
sig_injs[name].fit(srcs_rec, src_maps=src_maps, MC=mc, livetime=livetime)
# Strip unused fields from data and mc to better connect to skylab
keep = sig_injs[name].provided_data
drop = filter(lambda s: s not in keep, exp.dtype.names)
exp = np.lib.recfunctions.drop_fields(exp, drop)
print("Stripped '{}' data from fields:\n {}".format(name, drop))
# Setup the energy LLH model with fixed index, only ns is fitted
settings = _loader.settings_loader(name, skylab_bins=use_skylab_bins)[name]
llh_model = EnergyLLH(**settings["llh_model_opts"])
llh = PointSourceLLH(exp,
mc,
livetime,
llh_model,
scramble=True,
**settings["llh_opts"])
multillh.add_sample(name, llh)
del exp
del mc
gc.collect()
multiinj.fit(sig_injs)
del src_maps
gc.collect()
# Do performance trials by injecting on a grid
print(":: Starting {} performance trials ::".format(ntrials))
def config(alert_ind,
seed=1,
scramble=True,
e_range=(0, np.inf),
g_range=[1., 5.],
gamma=2.0,
E0=1 * TeV,
remove=False,
ncpu=20,
nside=256,
poisson=False,
injector=True,
verbose=True,
smear=True):
r""" Configure point source likelihood and injector.
Parameters
----------
alert_ind: int
index of IceCube alert event
seed : int
Seed for random number generator
Returns
-------
llh : PointSourceLLH
Point source likelihood object
inj : PriorInjector
Point source injector object
"""
seasons = [("GFUOnline_v001p02", "IC86, 2011-2018"),
("GFUOnline_v001p02", "IC86, 2019")]
#seasons = [("GFU_v002_p05", "IC86, 2011-2018"),
# (GFU_v002_p05", "IC86, 2019")]
skymap_files = sorted(
glob('/data/ana/realtime/alert_catalog_v2/fits_files/Run1*.fits.gz'))
skymap_fits, skymap_header = hp.read_map(skymap_files[alert_ind],
h=True,
verbose=False)
skymap_header = {name: val for name, val in skymap_header}
run_id, ev_id = skymap_header['RUNID'], skymap_header['EVENTID']
# Check to make sure there aren't any indexing errors
alert_df = pd.read_csv(f_path + 'icecube_misc/alert_dataframe.csv')
df_entry = alert_df.iloc[alert_ind]
assert df_entry['Run ID'] == run_id, \
"Dataframe run ID does not match the fits file run ID"
assert df_entry['Event ID'] == ev_id, \
"Dataframe event ID does not match the fits file event ID"
ev_mjd = skymap_header['EVENTMJD']
ev_iso = skymap_header['START']
signalness = skymap_header['SIGNAL']
ev_en = skymap_header['ENERGY']
ev_ra, ev_dec = np.radians(skymap_header['RA']), np.radians(
skymap_header['DEC'])
ev_stream = skymap_header['I3TYPE']
skymap_llh = skymap_fits.copy()
skymap_fits = np.exp(-1. * skymap_fits /
2.) #Convert from 2LLH to unnormalized probability
skymap_fits = np.where(skymap_fits > 1e-12, skymap_fits, 0.0)
skymap_fits = skymap_fits / np.sum(skymap_fits)
if smear:
ninety_msk = skymap_llh < 64.2
init_nside = hp.get_nside(skymap_llh)
cdf = np.cumsum(np.sort(skymap_fits[ninety_msk][::-1]))
pixs_above_ninety = np.count_nonzero(cdf > 0.1)
original_ninety_area = hp.nside2pixarea(init_nside) * pixs_above_ninety
new_ninety_area = hp.nside2pixarea(init_nside) * np.count_nonzero(
skymap_fits[ninety_msk])
original_ninety_radius = np.sqrt(original_ninety_area / np.pi)
new_ninety_radius = np.sqrt(new_ninety_area / np.pi)
scaled_probs = scale_2d_gauss(skymap_fits, original_ninety_radius,
new_ninety_radius)
skymap_fits = scaled_probs
if hp.pixelfunc.get_nside(skymap_fits) != nside:
skymap_fits = hp.pixelfunc.ud_grade(skymap_fits, nside)
skymap_fits = skymap_fits / skymap_fits.sum()
spatial_prior = SpatialPrior(skymap_fits, containment=0.99)
llh = [] # store individual llh as lists to prevent pointer over-writing
multillh = MultiPointSourceLLH(ncpu=1)
if verbose:
print("\n seasons:")
for season in np.atleast_1d(seasons):
sample = season[0]
name = season[1]
dataset = Datasets[sample]
exp, mc, livetime = Datasets[sample].season(name,
floor=np.radians(0.2))
if remove:
run_msk = exp['run'] == run_id
ev_msk = exp['event'] == ev_id
if np.count_nonzero(run_msk * ev_msk) > 0:
mjd_keys = exp['time'][run_msk * ev_msk]
exp = dataset.remove_ev(exp, mjd_keys=mjd_keys[0])
sinDec_bins = Datasets[sample].sinDec_bins(name)
energy_bins = Datasets[sample].energy_bins(name)
msg = " - % 15s (" % season
msg += "livetime %7.2f days, %6d events" % (livetime, exp.size)
msg += ", mjd0 %.2f" % min(exp['time'])
msg += ", mjd1 %.2f)" % max(exp['time'])
if verbose:
print(msg)
llh_model = EnergyLLH(twodim_bins=[energy_bins, sinDec_bins],
allow_empty=True,
bounds=g_range,
seed=gamma,
kernel=1,
ncpu=ncpu)
llh.append(
PointSourceLLH(exp,
mc,
livetime,
mode="box",
scramble=scramble,
llh_model=llh_model,
nsource_bounds=(0., 1e3),
nsource=1.))
multillh.add_sample(sample + " : " + name, llh[-1])
# save a little RAM by removing items copied into LLHs
del exp, mc
# END for (season)
if injector is False:
return multillh, spatial_prior
else:
inj = PriorInjector(spatial_prior,
seed=seed,
gamma=gamma,
E0=1 * TeV,
bunchsize=10)
inj.fill(multillh.exp, multillh.mc, multillh.livetime)
if verbose:
print("\n injected spectrum:")
print(" - %s" % str(inj.spectrum))
return multillh, spatial_prior, inj
def config(seasons, seed=1, scramble=True, verbose=True, dec=0.):
r""" Configure multi season point source likelihood and injector.
Parameters
----------
seasons : list
List of season names
seed : int
Seed for random number generator
Returns
-------
multillh : MultiPointSourceLLH
Multi year point source likelihood object
inj : PointSourceInjector
Point source injector object
"""
# store individual llh as lists to prevent pointer over-writing
llh = []
multillh = FastMultiPointSourceLLH(seed=seed, ncpu=25)
# setup likelihoods
if verbose: print("\n seasons:")
for season in np.atleast_1d(seasons):
#sample = 'PointSourceTracks7yr_galactic_plane'
#sample = 'PointSourceTracks7yr'
#sample = 'PointSourceTracks_v002p01b'
sample = 'PointSourceTracks'
exp, mc, livetime = Datasets[sample].season(season)
sinDec_bins = Datasets[sample].sinDec_bins(season)
energy_bins = Datasets[sample].energy_bins(season)
mc = mc[mc["logE"] > 1.]
if verbose:
print(" - %-15s (livetime %7.2f days, %6d events)" %
(season, livetime, exp.size))
llh_model = EnergyLLH(twodim_bins=[energy_bins, sinDec_bins],
allow_empty=True,
bounds=[1., 4.],
seed=2.,
kernel=1)
llh.append(
PointSourceLLH(exp,
mc,
livetime,
mode="box",
seed=seed,
scramble=scramble,
llh_model=llh_model,
nsource_bounds=(0., 1e3),
delta_ang=np.deg2rad(1.),
nsource=15.))
multillh.add_sample(season, llh[-1])
# save a little RAM by removing items copied into LLHs
del exp, mc
# END for (season)
#######
# LLH #
#######
#############################################################################
############
# INJECTOR #
############
'''
inj = PointSourceInjector(gamma = 2., E0 = 1*TeV, seed = seed)
inj.fill(dec, multillh.exp, multillh.mc, multillh.livetime)
'''
inj = None
############
# INJECTOR #
############
#############################################################################
'''
if verbose:
print("\n fitted spectrum:")
vals = (inj.E0/TeV)
print(" - dN/dE = A (E / %.1f TeV)^-index TeV^-1cm^-2s^-1" % vals)
print(" - index is *fit*")
print("\n injected spectrum:")
vals = (inj.E0/TeV, inj.gamma)
print(" - dN/dE = A (E / %.1f TeV)^-%.2f TeV^-1cm^-2s^-1" % vals)
'''
return (multillh, inj)
def config(seasons,
seed=1,
scramble=True,
e_range=(0, np.inf),
verbose=True,
gamma=2.0,
dec=0.,
remove=False,
src_w=None):
r""" Configure multi season point source likelihood and injector.
Parameters
----------
seasons : list
List of season names
seed : int
Seed for random number generator
Returns
-------
multillh : MultiPointSourceLLH
Multi year point source likelihood object
inj : PointSourceInjector
Point source injector object
"""
print("Scramble is %s" % str(scramble))
# store individual llh as lists to prevent pointer over-writing
llh = []
multillh = MultiPointSourceLLH(seed=seed, ncpu=25)
# setup likelihoods
if verbose: print("\n seasons:")
for season in np.atleast_1d(seasons):
sample = season[0]
name = season[1]
exp, mc, livetime = Datasets[sample].season(name)
sinDec_bins = Datasets[sample].sinDec_bins(name)
energy_bins = Datasets[sample].energy_bins(name)
if sample == "GFU" and remove:
exp = Datasets['GFU'].remove_ev(
exp, 58018.87118560489) # remove EHE 170922A
mc = mc[mc["logE"] > 1.]
if verbose:
print(" - % 15s : % 15s" % (sample, name))
vals = (livetime, exp.size, min(exp['time']), max(exp['time']))
print(
" (livetime %7.2f days, %6d events, mjd0 %.2f, mjd1 %.2f)"
% vals)
llh_model = EnergyLLH(twodim_bins=[energy_bins, sinDec_bins],
allow_empty=True,
bounds=[1., 4.],
seed=2.,
kernel=1)
llh.append(
PointSourceLLH(exp,
mc,
livetime,
mode="box",
scramble=scramble,
llh_model=llh_model,
nsource_bounds=(0., 1e3),
nsource=15.))
multillh.add_sample(sample + " : " + name, llh[-1])
# save a little RAM by removing items copied into LLHs
del exp, mc
# END for (season)
#######
# LLH #
#######
#############################################################################
############
# INJECTOR #
############
inj = PointSourceInjector(gamma=gamma,
E0=1 * TeV,
seed=seed,
e_range=e_range)
inj.fill(dec, multillh.exp, multillh.mc, multillh.livetime, src_w)
############
# INJECTOR #
############
#############################################################################
if verbose:
print("\n fitted spectrum:")
vals = (inj.E0 / TeV)
print(" - dN/dE = A (E / %.1f TeV)^-index TeV^-1cm^-2s^-1" % vals)
print(" - index is *fit*")
print("\n injected spectrum:")
vals = (inj.E0 / TeV, inj.gamma)
print(" - dN/dE = A (E / %.1f TeV)^-%.2f TeV^-1cm^-2s^-1" % vals)
return (multillh, inj)