def find_percentile(my_dir, file_dir, filter1, filter2, filter3, flux_filter1,
flux_filter2, flux_filter3, flux_filter1_err,
flux_filter2_err, flux_filter3_err, z):
all_percentile = []
data_point = (flux_filter1, flux_filter2, flux_filter3)
data_point = np.asarray(data_point)
mc_file_name = file_name(my_dir, file_dir, z)
mc_dict = load(mc_file_name)
mc_dict = add_error(mc_dict, filter1, flux_filter1_err)
mc_dict = add_error(mc_dict, filter2, flux_filter2_err)
mc_dict = add_error(mc_dict, filter3, flux_filter3_err)
# print 'redshift: ', z
Ia_percentile = kde3d(mc_dict['type_Ia']['flux'][filter1],
mc_dict['type_Ia']['flux'][filter2],
mc_dict['type_Ia']['flux'][filter3], data_point)
Ibc_percentile = kde3d(mc_dict['type_Ibc']['flux'][filter1],
mc_dict['type_Ibc']['flux'][filter2],
mc_dict['type_Ibc']['flux'][filter3], data_point)
II_percentile = kde3d(mc_dict['type_II']['flux'][filter1],
mc_dict['type_II']['flux'][filter2],
mc_dict['type_II']['flux'][filter3], data_point)
my_percentile = (Ia_percentile, Ibc_percentile, II_percentile)
my_percentile = np.asarray(my_percentile)
all_percentile.append(my_percentile)
all_percentile = np.squeeze(all_percentile)
return all_percentile
def get_data(my_dir, file_dir, filter1, filter2, filter3, flux_filter1_err,
flux_filter2_err, flux_filter3_err, z):
indices = [i for i, x in enumerate(file_dir) if x == '_']
n = int(file_dir[1:indices[0]])
n_Ibc_start = int(n)
n_II_start = 2 * n_Ibc_start
mc_file_name = file_name(my_dir, file_dir, z)
mc_dict = load(mc_file_name)
mc_dict = add_error(mc_dict, filter1, flux_filter1_err)
mc_dict = add_error(mc_dict, filter2, flux_filter2_err)
mc_dict = add_error(mc_dict, filter3, flux_filter3_err)
all_filter1 = np.concatenate(
(mc_dict['type_Ia']['flux'][filter1],
mc_dict['type_Ibc']['flux'][filter1],
mc_dict['type_II']['flux'][filter1]))
all_filter2 = np.concatenate(
(mc_dict['type_Ia']['flux'][filter2],
mc_dict['type_Ibc']['flux'][filter2],
mc_dict['type_II']['flux'][filter2]))
all_filter3 = np.concatenate(
(mc_dict['type_Ia']['flux'][filter3],
mc_dict['type_Ibc']['flux'][filter3],
mc_dict['type_II']['flux'][filter3]))
data = [all_filter1, all_filter2, all_filter3]
data = zip(*data)
all_sources = pd.DataFrame(data, columns=[filter1, filter2, filter3],
index=range(3*n))
all_sources['Type'] = 'Type Ia'
all_sources.ix[n_Ibc_start:n_II_start, 'Type'] = 'Type Ibc'
all_sources.ix[n_II_start:, 'Type'] = 'Type II'
all_features = copy.copy(all_sources)
all_label = all_sources["Type"]
del all_features["Type"]
X = copy.copy(all_features.values)
Y = copy.copy(all_label.values)
return X, Y
def flux_fluxDiff_arrays(my_dir, file_dir, filter1, filter2, z):
mc_file_name = file_name(my_dir, file_dir, z)
mc_dict = load(mc_file_name)
type_Ia_flux_diff = (mc_dict['type_Ia']['flux'][filter2] -
mc_dict['type_Ia']['flux'][filter1])
type_Ibc_flux_diff = (mc_dict['type_Ibc']['flux'][filter2] -
mc_dict['type_Ibc']['flux'][filter1])
type_II_flux_diff = (mc_dict['type_II']['flux'][filter2] -
mc_dict['type_II']['flux'][filter1])
flux = [mc_dict['type_Ia']['flux'][filter1],
mc_dict['type_Ibc']['flux'][filter1],
mc_dict['type_II']['flux'][filter1]]
diff = [type_Ia_flux_diff, type_Ibc_flux_diff, type_II_flux_diff]
for i, item in enumerate(diff):
flux[i] = mask(flux[i], 98)
diff[i] = mask(diff[i], 98)
return flux, diff
def get_data(my_dir, file_dir, filter1, filter2, filter3, flux_filter1_err,
flux_filter2_err, flux_filter3_err, z):
indices = [i for i, x in enumerate(file_dir) if x == '_']
n = int(file_dir[1:indices[0]])
n_Ibc_start = int(n)
n_II_start = 2 * n_Ibc_start
mc_file_name = file_name(my_dir, file_dir, z)
mc_dict = load(mc_file_name)
mc_dict = add_error(mc_dict, filter1, flux_filter1_err)
mc_dict = add_error(mc_dict, filter2, flux_filter2_err)
mc_dict = add_error(mc_dict, filter3, flux_filter3_err)
all_filter1 = np.concatenate((mc_dict['type_Ia']['flux'][filter1],
mc_dict['type_Ibc']['flux'][filter1],
mc_dict['type_II']['flux'][filter1]))
all_filter2 = np.concatenate((mc_dict['type_Ia']['flux'][filter2],
mc_dict['type_Ibc']['flux'][filter2],
mc_dict['type_II']['flux'][filter2]))
all_filter3 = np.concatenate((mc_dict['type_Ia']['flux'][filter3],
mc_dict['type_Ibc']['flux'][filter3],
mc_dict['type_II']['flux'][filter3]))
data = [all_filter1, all_filter2, all_filter3]
data = zip(*data)
all_sources = pd.DataFrame(data,
columns=[filter1, filter2, filter3],
index=range(3 * n))
all_sources['Type'] = 'Type Ia'
all_sources.ix[n_Ibc_start:n_II_start, 'Type'] = 'Type Ibc'
all_sources.ix[n_II_start:, 'Type'] = 'Type II'
all_features = copy.copy(all_sources)
all_label = all_sources["Type"]
del all_features["Type"]
X = copy.copy(all_features.values)
Y = copy.copy(all_label.values)
return X, Y
def flux_fluxDiff_arrays(my_dir, file_dir, filter1, filter2, z):
mc_file_name = file_name(my_dir, file_dir, z)
mc_dict = load(mc_file_name)
type_Ia_flux_diff = (mc_dict['type_Ia']['flux'][filter2] -
mc_dict['type_Ia']['flux'][filter1])
type_Ibc_flux_diff = (mc_dict['type_Ibc']['flux'][filter2] -
mc_dict['type_Ibc']['flux'][filter1])
type_II_flux_diff = (mc_dict['type_II']['flux'][filter2] -
mc_dict['type_II']['flux'][filter1])
flux = [
mc_dict['type_Ia']['flux'][filter1],
mc_dict['type_Ibc']['flux'][filter1],
mc_dict['type_II']['flux'][filter1]
]
diff = [type_Ia_flux_diff, type_Ibc_flux_diff, type_II_flux_diff]
for i, item in enumerate(diff):
flux[i] = mask(flux[i], 98)
diff[i] = mask(diff[i], 98)
return flux, diff
def combined2D(final_pdf,
my_dir,
file_dir,
filter1,
filter2,
flux_filter1,
flux_filter2,
flux_filter1_err,
flux_filter2_err,
outdir,
file_with_RF_and_SF_arrays,
photoz_plot_name,
random_forest_plot_name,
survival_function_plot_name,
final_pdf_plot_name,
photo_z_type=None,
photo_z_file=None,
photo_z_redshift_file=None,
mu=None,
sigma=None):
if not os.path.exists(outdir):
os.makedirs(outdir)
if not os.path.isfile(file_with_RF_and_SF_arrays):
file_name = save_arrays(my_dir, file_dir, filter1, filter2,
flux_filter1, flux_filter2, flux_filter1_err,
flux_filter2_err, outdir)
my_dict = load(file_name)
else:
my_dict = load(file_with_RF_and_SF_arrays)
rf = my_dict['rf']
sf = my_dict['sf']
z = my_dict['rf_z']
rf = np.asarray(rf)
sf = np.asarray(sf)
if photo_z_type is not False:
photo_z, my_z = iterator('photo_z',
my_dir,
file_dir,
filter1,
filter2,
flux_filter1,
flux_filter2,
flux_filter1_err,
flux_filter2_err,
photo_z_type=photo_z_type,
photo_z_file=photo_z_file,
photo_z_redshift_file=photo_z_redshift_file,
mu=mu,
sigma=sigma)
if photo_z_type == 'file':
title_photoz = 'Galaxy Photo-z from file'
else:
title_photoz = ("Galaxy Photo-z: Gaussian( mu=%.2f, sigma=%.2f)" %
(mu, sigma))
if final_pdf == 'RF+SF+photoz':
plt.clf()
title = ['Random Forest', 'Survival Function', 'Photo-z']
plot(z,
rf,
'PDF',
'Random Forest',
outdir,
outname=random_forest_plot_name)
plot(z,
sf,
'1 - CDF',
'Survival Function',
outdir,
outname=survival_function_plot_name)
plot(z, photo_z, 'PDF', title_photoz, outdir, outname=photoz_plot_name)
first_product = np.multiply(sf, rf)
product = (np.multiply(first_product.T, photo_z)).T
x = [z, z, z, z]
y = [rf, sf, photo_z, product]
subplot(x, y, title, outdir, final_pdf_plot_name, title_photoz)
# norm_product = product/product.sum()
elif final_pdf == 'RF+SF':
plt.clf()
title = ['Random Forest', 'Survival Function', '']
plot(z,
rf,
'PDF',
'Random Forest',
outdir,
outname=random_forest_plot_name)
plot(z,
sf,
'1 - CDF',
'Survival Function',
outdir,
outname=survival_function_plot_name)
product = np.multiply(rf, sf)
x = [z, z, [], z]
y = [rf, sf, [], product]
subplot(x, y, title, outdir, final_pdf_plot_name)
# norm_product = product/product.sum()
elif final_pdf == 'RF+photoz':
plt.clf()
title = ['Random Forest', '', 'Photo-z']
plot(z,
rf,
'PDF',
'Random Forest',
outdir,
outname=random_forest_plot_name)
plot(z, photo_z, 'PDF', title_photoz, outdir, outname=photoz_plot_name)
product = np.multiply(rf.T, photo_z)
product = product.T
# norm_product = product/product.sum()
x = [z, [], z, z]
y = [rf, [], photo_z, product]
subplot(x, y, title, outdir, final_pdf_plot_name, title_photoz)
elif final_pdf == 'SF+photoz':
plt.clf()
title = ['', 'Survival Function', 'Photo-z']
plot(z,
sf,
'1 - CDF',
'Survival Function',
outdir,
outname=survival_function_plot_name)
plot(z, photo_z, 'PDF', title_photoz, outdir, outname=photoz_plot_name)
product = np.multiply(sf.T, photo_z)
product = product.T
# norm_product = product/product.sum()
x = [[], z, z, z]
y = [[], sf, photo_z, product]
subplot(x, y, title, outdir, final_pdf_plot_name, title_photoz)
elif final_pdf == 'RF':
plt.clf()
plot(z,
rf,
'PDF',
'Random Forest',
outdir,
outname=random_forest_plot_name)
elif final_pdf == 'SF':
plt.clf()
plot(z,
sf,
'1 - CDF',
'Survival Function',
outdir,
outname=survival_function_plot_name)
elif final_pdf == 'photoz':
plt.clf()
plot(z, photo_z, 'PDF', title_photoz, outdir, outname=photoz_plot_name)
def combined2D(final_pdf, my_dir, file_dir,
filter1, filter2,
flux_filter1, flux_filter2,
flux_filter1_err, flux_filter2_err,
outdir, file_with_RF_and_SF_arrays, photoz_plot_name,
random_forest_plot_name, survival_function_plot_name,
final_pdf_plot_name, photo_z_type=None, photo_z_file=None,
photo_z_redshift_file=None, mu=None, sigma=None):
if not os.path.exists(outdir):
os.makedirs(outdir)
if not os.path.isfile(file_with_RF_and_SF_arrays):
file_name = save_arrays(
my_dir, file_dir, filter1, filter2, flux_filter1,
flux_filter2, flux_filter1_err, flux_filter2_err, outdir)
my_dict = load(file_name)
else:
my_dict = load(file_with_RF_and_SF_arrays)
rf = my_dict['rf']
sf = my_dict['sf']
z = my_dict['rf_z']
rf = np.asarray(rf)
sf = np.asarray(sf)
if photo_z_type is not False:
photo_z, my_z = iterator('photo_z', my_dir, file_dir, filter1,
filter2, flux_filter1, flux_filter2,
flux_filter1_err,
flux_filter2_err,
photo_z_type=photo_z_type,
photo_z_file=photo_z_file,
photo_z_redshift_file=photo_z_redshift_file,
mu=mu, sigma=sigma)
if photo_z_type == 'file':
title_photoz = 'Galaxy Photo-z from file'
else:
title_photoz = ("Galaxy Photo-z: Gaussian( mu=%.2f, sigma=%.2f)"
% (mu, sigma))
if final_pdf == 'RF+SF+photoz':
plt.clf()
title = ['Random Forest', 'Survival Function', 'Photo-z']
plot(z, rf, 'PDF', 'Random Forest', outdir,
outname=random_forest_plot_name)
plot(z, sf, '1 - CDF', 'Survival Function', outdir,
outname=survival_function_plot_name)
plot(z, photo_z, 'PDF', title_photoz, outdir, outname=photoz_plot_name)
first_product = np.multiply(sf, rf)
product = (np.multiply(first_product.T, photo_z)).T
x = [z, z, z, z]
y = [rf, sf, photo_z, product]
subplot(x, y, title, outdir, final_pdf_plot_name, title_photoz)
# norm_product = product/product.sum()
elif final_pdf == 'RF+SF':
plt.clf()
title = ['Random Forest', 'Survival Function', '']
plot(z, rf, 'PDF', 'Random Forest', outdir,
outname=random_forest_plot_name)
plot(z, sf, '1 - CDF', 'Survival Function', outdir,
outname=survival_function_plot_name)
product = np.multiply(rf, sf)
x = [z, z, [], z]
y = [rf, sf, [], product]
subplot(x, y, title, outdir, final_pdf_plot_name)
# norm_product = product/product.sum()
elif final_pdf == 'RF+photoz':
plt.clf()
title = ['Random Forest', '', 'Photo-z']
plot(z, rf, 'PDF', 'Random Forest', outdir,
outname=random_forest_plot_name)
plot(z, photo_z, 'PDF', title_photoz, outdir, outname=photoz_plot_name)
product = np.multiply(rf.T, photo_z)
product = product.T
# norm_product = product/product.sum()
x = [z, [], z, z]
y = [rf, [], photo_z, product]
subplot(x, y, title, outdir, final_pdf_plot_name, title_photoz)
elif final_pdf == 'SF+photoz':
plt.clf()
title = ['', 'Survival Function', 'Photo-z']
plot(z, sf, '1 - CDF', 'Survival Function', outdir,
outname=survival_function_plot_name)
plot(z, photo_z, 'PDF', title_photoz, outdir, outname=photoz_plot_name)
product = np.multiply(sf.T, photo_z)
product = product.T
# norm_product = product/product.sum()
x = [[], z, z, z]
y = [[], sf, photo_z, product]
subplot(x, y, title, outdir, final_pdf_plot_name, title_photoz)
elif final_pdf == 'RF':
plt.clf()
plot(z, rf, 'PDF', 'Random Forest', outdir,
outname=random_forest_plot_name)
elif final_pdf == 'SF':
plt.clf()
plot(z, sf, '1 - CDF', 'Survival Function', outdir,
outname=survival_function_plot_name)
elif final_pdf == 'photoz':
plt.clf()
plot(z, photo_z, 'PDF', title_photoz, outdir, outname=photoz_plot_name)