-
Notifications
You must be signed in to change notification settings - Fork 0
/
analyse_intrinsic_colour.py
205 lines (173 loc) · 8.89 KB
/
analyse_intrinsic_colour.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
import os
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
import scipy.optimize
from astropy.io import fits, ascii
from scipy.stats import pearsonr
from scipy import odr
import matplotlib as mpl
from matplotlib import rc
import pickle
rc('font', **{'family': 'serif', 'serif': ['cmr10']})
mpl.rcParams['axes.unicode_minus'] = False
mpl.rcParams['mathtext.fontset'] = 'cm'
plt.rcParams.update({'font.size': 22})
from lmfit.models import LinearModel
# mpl.use('macosx')
def line(x, m, c):
return m * x + c
def correlation_step_plot(df, x_param, y_param, xlabel, ylabel, split_point=None):
param_df = df.dropna(subset=(x_param, y_param))
if f'b_{x_param}' in param_df.columns:
xerr = np.mean([param_df[x_param] - param_df[f'b_{x_param}'],
param_df[f'B_{x_param}'] - param_df[x_param]], axis=0)
floor = xerr[xerr > 0].min()
xerr = np.max([xerr, np.ones_like(xerr) * floor], axis=0)
xerr = np.nan_to_num(xerr, nan=floor)
else:
xerr = param_df[f'{x_param}_err']
floor = xerr[xerr > 0].min()
xerr = np.max([xerr, np.ones_like(xerr) * floor], axis=0)
xerr = np.nan_to_num(xerr, nan=floor)
if f'b_{y_param}' in param_df.columns:
yerr = np.mean([param_df[y_param] - param_df[f'b_{y_param}'],
param_df[f'B_{y_param}'] - param_df[y_param]], axis=0)
floor = yerr[yerr > 0].min()
yerr = np.max([yerr, np.ones_like(yerr) * floor], axis=0)
yerr = np.nan_to_num(yerr, nan=floor)
else:
yerr = param_df[f'{y_param}_err']
floor = yerr[yerr > 0].min()
yerr = np.max([yerr, np.ones_like(yerr) * floor], axis=0)
yerr = np.nan_to_num(yerr, nan=floor)
param_df['xerr'] = xerr
param_df['yerr'] = yerr
plt.figure(figsize=(12, 8))
plt.errorbar(param_df[x_param], param_df[y_param], yerr=yerr, fmt='x', alpha=0.4)
plt.xlabel(xlabel)
plt.ylabel(ylabel)
if split_point is None:
split_point = param_df[x_param].median()
df1, df2 = param_df[param_df[x_param] < split_point], param_df[param_df[x_param] > split_point]
wmu1, wmu2 = (df1[y_param] / df1.yerr).sum() / (1 / df1.yerr).sum(), \
(df2[y_param] / df1.yerr).sum() / (1 / df1.yerr).sum()
N1, N2 = df1.shape[0], df2.shape[0]
wstd1 = np.sqrt((np.power(df1[y_param] - wmu1, 2) * (1 / df1.yerr)).sum() / (
((N1 - 1) / N1) * (1 / df1.yerr).sum()))
wstd2 = np.sqrt((np.power(df2[y_param] - wmu2, 2) * (1 / df2.g_r_at_max_err)).sum() / (
((N2 - 1) / N2) * (1 / df2.g_r_at_max_err).sum()))
means = [df1[x_param].mean(), df2[x_param].mean()]
plt.errorbar(means, [wmu1, wmu2], yerr=[wstd1, wstd2], fmt='kx')
plt.vlines(split_point, param_df.g_r_at_max.min(), param_df.g_r_at_max.max(), ls='--', color='k')
p, r = pearsonr(param_df[x_param], param_df[y_param])
print(f'{x_param}: ', p, r)
data = odr.Data(param_df[x_param], param_df[y_param], wd=1/xerr, we=1/yerr)
fit_odr = odr.ODR(data, model=odr.unilinear)
result = fit_odr.run()
m, c = result.beta
merr, cerr = result.sd_beta
print(f'{x_param}: ', m, merr, m / merr)
popt, pcov = scipy.optimize.curve_fit(line, param_df[x_param], param_df[y_param], p0=(m, c), sigma=yerr)
m, c, = popt
merr, cerr = np.sqrt(np.diag(pcov))
print(f'{x_param}: ', m, merr, m / merr)
plot_x = np.linspace(param_df[x_param].min(), param_df[x_param].max(), 3)
plot_y = m * plot_x + c
plt.plot(plot_x, plot_y, color='b')
plt.plot(plot_x, (m + merr) * plot_x + c - cerr, color='b', ls='-.')
plt.plot(plot_x, (m - merr) * plot_x + c + cerr, color='b', ls='-.')
def main():
# Load host properties-------------------
hdu = fits.open('data/host/J_ApJ_867_108_localsn.dat.fits')
host_data = hdu[1].data
df = pd.DataFrame.from_records(host_data)
df['SN'] = df.SN.apply(lambda x: x.rstrip())
sn_list = pd.read_csv('data/lcs/Foundation_DR1/Foundation_DR1/Foundation_DR1.LIST', names=['file'])
sn_list['sn'] = sn_list.file.apply(lambda x: x[x.rfind('_') + 1: x.rfind('.')])
meta_file = pd.read_csv('data/lcs/meta/T21_training_set_meta.txt', delim_whitespace=True)
sn_list = sn_list.merge(meta_file, left_on='sn', right_on='SNID')
df = sn_list.merge(df, how='left', left_on='sn', right_on='SN')
df = df.replace(-99.0, np.nan).copy()
host_data = pd.read_csv('data/host/GPC1v3_hosts.txt', delim_whitespace=True)
host_data = host_data.replace('-', np.nan).replace('None', np.nan)
for col in host_data.columns:
if col not in ['ID', 'ra', 'dec', 'hostra', 'hostdec']:
host_data[col] = host_data[col].astype(float)
df = df.merge(host_data, how='left', left_on='sn', right_on='ID').copy()
df['host_g-r'] = df.PS1gMag_local - df.PS1rMag_local
df['host_g-r_err'] = np.sqrt(np.power(df.PS1gMagErr_local, 2) + np.power(df.PS1rMagErr_local, 2))
df['host_u-r'] = df.SDSSuMag_local - df.SDSSrMag_local
df['host_u-r_err'] = np.sqrt(np.power(df.SDSSuMagErr_local, 2) + np.power(df.SDSSuMagErr_local, 2))
hres = np.load(os.path.join('results', 'foundation_fit_T21freeRv', 'hres.npy'))
df['Hres_bayesn'] = hres[1, :]
df['Hres_bayesn_err'] = hres[2, :]
df['theta'] = hres[3, :]
df['theta_err'] = hres[4, :]
df['Av'] = hres[5, :]
df['Av_err'] = hres[6, :]
df['Hres_err'] = df.e_Hres
mags = np.load(os.path.join('results', 'foundation_fit_T21freeRv', 'rf_mags.npy'))
colours = np.zeros((mags.shape[0], mags.shape[1] - 1, *mags.shape[2:]))
for i in range(colours.shape[1]):
colours[:, i, ...] = mags[:, i, ...] - mags[:, i + 1, ...]
c, cerr = colours.mean(axis=0), colours.std(axis=0)
eps0_mags = np.load(os.path.join('results', 'foundation_fit_T21freeRv', 'rf_mags_eps0.npy'))
eps0colours = np.zeros((eps0_mags.shape[0], eps0_mags.shape[1] - 1, *eps0_mags.shape[2:]))
for i in range(eps0colours.shape[1]):
eps0colours[:, i, ...] = eps0_mags[:, i, ...] - eps0_mags[:, i + 1, ...]
eps0c, eps0cerr = eps0colours.mean(axis=0), eps0colours.std(axis=0)
g_r_at_max, g_r_at_max_err = c[0, 2, :], cerr[0, 2, :]
c_eps = c - eps0c
c_eps_err = np.sqrt(cerr * cerr + eps0cerr * eps0cerr)
g_r_at_max, g_r_at_max_err = c_eps[0, 2, :], c_eps_err[0, 2, :]
# Load delta mus------------
df['g_r_at_max'] = g_r_at_max
df['g_r_at_max_err'] = g_r_at_max_err
# Load dist mods
with open('results/foundation_fit_T21/chains.pkl', 'rb') as file:
eps_chains = pickle.load(file)
with open('results/foundation_fit_noeps/chains.pkl', 'rb') as file:
no_eps_chains = pickle.load(file)
eps_mu, eps_mu_err = eps_chains['mu'].mean(axis=(0, 1)), eps_chains['mu'].std(axis=(0, 1))
no_eps_mu, no_eps_mu_err = no_eps_chains['mu'].mean(axis=(0, 1)), no_eps_chains['mu'].std(axis=(0, 1))
delta_mu = eps_mu - no_eps_mu
print('Delta_mu dispersion: ', np.std(delta_mu))
df['delta_mu'] = delta_mu
df['delta_mu_err'] = 0.01
# Plots-------------------
correlation_step_plot(df, 'Av', 'Hres', r'A$_V$', 'SALT2 Hubble Residual')
plt.savefig('plots/Av_vs_SALT_Hres.png')
plt.show()
correlation_step_plot(df, 'host_g-r', 'g_r_at_max', r'Local PS1 g-r colour', 'Intrinsic g-r SN colour at peak')
plt.savefig('plots/intrinsic_g-r_vs_host_g-r.png')
plt.show()
correlation_step_plot(df, 'Av', 'g_r_at_max', r'A$_V$', 'Intrinsic g-r SN colour at peak')
plt.xscale('log')
plt.savefig('plots/intrinsic_g-r_vs_Av.png')
plt.show()
correlation_step_plot(df, 'host_u-r', 'g_r_at_max', r'Local SDSS u-r', 'Intrinsic g-r SN colour at peak')
plt.savefig('plots/intrinsic_g-r_vs_host_u-r.png')
plt.show()
correlation_step_plot(df, 'Mass', 'g_r_at_max', r'$\log_{10}$(Global mass)', 'Intrinsic g-r SN colour at peak')
plt.savefig('plots/intrinsic_g-r_vs_global_mass.png')
plt.show()
correlation_step_plot(df, 'Massloc', 'g_r_at_max', r'$\log_{10}$(Local mass)', 'Intrinsic g-r SN colour at peak')
plt.savefig('plots/intrinsic_g-r_vs_local_mass.png')
plt.show()
correlation_step_plot(df, '(u-g)loc', 'g_r_at_max', r'Local u-g colour', 'Intrinsic g-r SN colour at peak')
plt.savefig('plots/intrinsic_g-r_vs_local_colour.png')
plt.show()
correlation_step_plot(df, 'SFRloc', 'g_r_at_max', r'$\log_{10}$(Local SFR)', 'Intrinsic g-r SN colour at peak')
plt.savefig('plots/intrinsic_g-r_vs_local_SFR.png')
plt.show()
correlation_step_plot(df, 'Hres', 'g_r_at_max', r'Hubble residual (Jones+18)', 'Intrinsic g-r SN colour at peak')
plt.savefig('plots/intrinsic_g-r_vs_Hres.png')
plt.show()
correlation_step_plot(df, 'Hres_bayesn', 'g_r_at_max', r'Hubble residual (BayeSN)', 'Intrinsic g-r SN colour at peak')
plt.savefig('plots/intrinsic_g-r_vs_Hres_bayesn.png')
plt.show()
correlation_step_plot(df, 'delta_mu', 'g_r_at_max', r'$\Delta \mu$', 'Intrinsic g-r SN colour at peak')
plt.savefig('plots/intrinsic_g-r_vs_delta_mu.png')
plt.show()
main()