def parallel_datasets(number_of_simulations=n_sim / n_cores):
print 'Number of simulations per core:' + str(number_of_simulations)
np.random.seed() # changes the seed
Sn_datasets = []
ra = np.asarray(tdata['RA'])
dec = np.asarray(tdata['DEC'])
pa = np.asarray(tdata['position_angle'])
length = len(tdata)
max_ra = np.max(ra)
min_ra = np.min(ra)
max_dec = np.max(dec)
min_dec = np.min(dec)
rdata = Table()
for i in range(number_of_simulations):
np.random.shuffle(pa)
rdata['RA'] = ra
rdata['DEC'] = dec
rdata['position_angle'] = pa
Sn = angular_dispersion_vectorized_n(rdata,
n=n) # up to n nearest neighbours
Sn_datasets.append(Sn)
return Sn_datasets
def parallel_datasets(number_of_simulations=n_sim / n_cores):
print 'Number of simulations per core:' + str(number_of_simulations)
np.random.seed() # changes the seed
Sn_datasets = []
ra = np.asarray(tdata['RA'])
dec = np.asarray(tdata['DEC'])
pa = np.asarray(tdata['final_PA'])
length = len(tdata)
if redshift: z_best = np.asarray(tdata['z_best'])
max_ra = np.max(ra)
min_ra = np.min(ra)
max_dec = np.max(dec)
min_dec = np.min(dec)
rdata = Table()
for i in range(number_of_simulations):
np.random.shuffle(pa)
rdata['RA'] = ra
rdata['DEC'] = dec
if redshift: rdata['z_best'] = z_best
rdata['final_PA'] = pa
if parallel_transport:
Sn = angular_dispersion_vectorized_n_parallel_transport(
rdata, n=n, redshift=redshift) # up to n nearest neighbours
else:
Sn = angular_dispersion_vectorized_n(
rdata, n=n, redshift=redshift) # up to n nearest neighbours
Sn_datasets.append(Sn)
return Sn_datasets
def produce_Sn_plots():
Sn_data = angular_dispersion_vectorized_n(tdata, n)
prop_cycle = plt.rcParamsDefault['axes.prop_cycle']
colors = prop_cycle.by_key()['color']
if n == 60:
plt.hist(Sn_mc['S_%i' % n],
bins=25,
histtype=u'step',
color=colors[0],
label=r'$S_{%i}\vert_{MC}$' % n)
elif n == 140:
plt.hist(Sn_mc['S_%i' % n],
bins=25,
histtype=u'step',
color=colors[1],
label=r'$S_{%i}\vert_{MC}$' % n)
# print (Sn_data[n-1])
# plt.axvline(x=Sn_data[n-1],ymin=0,ymax=1,ls='dashed', label='Data no PT')
Sn_data = angular_dispersion_vectorized_n_parallel_transport(tdata, n)
print(Sn_data[n - 1])
if n == 60:
plt.axvline(x=Sn_data[n - 1],
ymin=0,
ymax=1,
ls='dashed',
label=r'$S_{60}$',
color=colors[0])
elif n == 140:
plt.axvline(x=Sn_data[n - 1],
ymin=0,
ymax=1,
ls='dashed',
label=r'$S_{140}$',
color=colors[1])
ax = plt.gca()
ax.tick_params(labelsize=12)
ax.set_xlabel(r'$S_n$', fontsize=14)
ax.set_ylabel(r'Counts', fontsize=14)
ax.legend(fontsize=14)
title = 'Distribution of S_%i' % n
title = ' '
plt.title(title, fontsize=16)
# if n == 140: plt.xlim(0.040,0.12151429413439714)
Sn_mc_n = np.asarray(Sn_mc['S_' + str(n)])
Sn_data_n = Sn_data[n - 1]
av_Sn_mc_n = np.mean(Sn_mc_n)
sigma = np.std(Sn_mc_n)
print sigma
print('S_%i = %.3f' % (n, Sn_data_n))
SL = 1 - norm.cdf((Sn_data_n - av_Sn_mc_n) / (sigma))
print('log10 SL = %.3f for n= %i, calculated with PT' %
(np.log10(SL), n))
def statistics(filename, tdata, redshift=False):
if parallel_transport:
if n < 998:
Sn_mc = Table(
fits.open('./data/2D/Sn_monte_carlo_PT' + filename +
'.fits')[1].data)
else:
Sn_mc = ascii.read('./data/2D/Sn_monte_carlo_PT' + filename +
'.csv')
else:
if n < 998:
Sn_mc = Table(
fits.open('./data/2D/Sn_monte_carlo_' + filename +
'.fits')[1].data)
else:
Sn_mc = ascii.read('./data/2D/Sn_monte_carlo_' + filename + '.csv')
# calculate angular radius for number of nn,
angular_radius = angular_radius_vs_n(tdata, filename, n)
if parallel_transport:
Sn_data = angular_dispersion_vectorized_n_parallel_transport(
tdata, n, redshift)
else:
Sn_data = angular_dispersion_vectorized_n(tdata, n, redshift)
Sn_vs_n(tdata, Sn_mc, Sn_data, filename, angular_radius, n)
def produce_Sn_plots():
Sn_data = angular_dispersion_vectorized_n(tdata, n)
plt.hist(Sn_mc['S_%i' % n],
bins=25,
histtype=u'step',
color='black',
label=r'$S_n\vert_{MC}$')
# print (Sn_data[n-1])
# plt.axvline(x=Sn_data[n-1],ymin=0,ymax=1,ls='dashed', label='Data no PT')
Sn_data = angular_dispersion_vectorized_n_parallel_transport(tdata, n)
print(Sn_data[n - 1])
plt.axvline(x=Sn_data[n - 1],
ymin=0,
ymax=1,
ls='dashed',
color='k',
label='Data')
ax = plt.gca()
ax.tick_params(labelsize=12)
ax.set_xlabel(r'$S_{%i}$' % n, fontsize=14)
ax.set_ylabel(r'Counts', fontsize=14)
ax.legend(fontsize=14)
title = 'Distribution of S_%i' % n
title = ' '
plt.title(title, fontsize=16)
if n == 140: plt.xlim(0.040, 0.12151429413439714)
plt.tight_layout()
plt.show()
Sn_mc_n = np.asarray(Sn_mc['S_' + str(n)])
Sn_data_n = Sn_data[n - 1]
av_Sn_mc_n = np.mean(Sn_mc_n)
sigma = np.std(Sn_mc_n)
print sigma
print('S_%i = %.3f' % (n, Sn_data_n))
SL = 1 - norm.cdf((Sn_data_n - av_Sn_mc_n) / (sigma))
print('log10 SL = %.3f for n= %i, calculated with PT' %
(np.log10(SL), n))
def statistics(filename, tdata, load=True, redshift=False):
Sn_mc = Table(
fits.open('./data/Sn_monte_carlo_' + filename + '.fits')[1].data)
# calculate angular radius for number of nn (does not include redshift obviously)
angular_radius = angular_radius_vs_n(tdata, filename, n)
if redshift: # same number of sources, so only have to plot this once
plt.plot((-10, -10), (-10, -10),
c='white',
label='Num_sources %i' % len(tdata))
Sn_data = angular_dispersion_vectorized_n(tdata, n, redshift)
line = Sn_vs_n(tdata,
Sn_mc,
Sn_data,
filename,
angular_radius,
n,
load=load,
redshift=redshift)
plot_additional_info(tdata, filename, redshift)
return line
plt.savefig('./angular_radius_vs_n_' + filename + '.pdf', overwrite=True)
plt.close()
return median
n = 180
for i in range(5):
# Open the size data
filename = 'size_bins2_' + str(i)
tdata = Table(fits.open('./' + filename + '.fits')[1].data)
tdata = deal_with_overlap(tdata)
Sn_mc = Table(fits.open('./Sn_monte_carlo_' + filename + '.fits')[1].data)
# Calculate the Sn and the plots
angular_radius = angular_radius_vs_n(tdata, filename, n)
Sn_data = angular_dispersion_vectorized_n(tdata, n)
Sn_vs_n(tdata, Sn_mc, Sn_data, filename, angular_radius, n)
break
# Open the flux data
filename = 'flux_bins2_' + str(i)
tdata = Table(fits.open('./' + filename + '.fits')[1].data)
tdata = deal_with_overlap(tdata)
Sn_mc = Table(fits.open('./Sn_monte_carlo_' + filename + '.fits')[1].data)
# Caluclate the Sn and the plots
Sn_data = angular_dispersion_vectorized_n(tdata, n)
Sn_vs_n(tdata, Sn_mc, Sn_data, filename, n)
angular_radius_vs_n(tdata, filename, n)
