-
Notifications
You must be signed in to change notification settings - Fork 2
/
intensity_profiles.py
260 lines (219 loc) · 10 KB
/
intensity_profiles.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
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
import os
import pickle
import numpy as np
from uncertainties import ufloat
import tqdm
import matplotlib
from matplotlib import pyplot as plt
from astropy.io import fits
from astropy import table
import get_lds_with_errors_v3 as glds
LINECOLOR = matplotlib.rcParams["ytick.color"]
# Compute and store intensity profile interpolators
def intensity_profile_interpolators(star, RF_list, models=None, main_dir="",
overwrite=False, nsig=4):
Teff = star["Teff"]
logg = star["logg"]
M_H = star["M_H"]
vturb = ufloat(2, 0.5) if star["vturb"] is None else star["vturb"]
name = star["Name"].replace(" ", "_")
fn = os.path.join(main_dir,
"{}_intensity_profile_interpolators.pkl".format(name))
if not os.path.isfile(fn) or overwrite:
bounds = {"Teff": (max(0, Teff.n-nsig*Teff.s), Teff.n+nsig*Teff.s),
"logg": (logg.n-nsig*logg.s, logg.n+nsig*logg.s),
"M_H": (M_H.n-nsig*M_H.s, M_H.n+nsig*M_H.s),
"vturb": (max(0, vturb.n-nsig*vturb.s),
vturb.n+nsig*vturb.s)}
glds.download_files(**bounds, models=models, force_download=False)
subgrids = glds.get_subgrids(**bounds, models=models)
ip_interp = glds.get_profile_interpolators(subgrids, RF_list,
models=models,
interpolation_order=1,
atlas_correction=True,
photon_correction=True,
max_bad_RF=0.0,
overwrite_pck=False,
atlas_hdu=1)
with open(fn, "wb") as f:
pickle.dump(ip_interp, f)
else:
with open(fn, "rb") as f:
ip_interp = pickle.load(f)
return ip_interp
# Drawing stellar parameters from normal distributions
def get_samples(star, RF_list, nsamples=10000):
Teff = star["Teff"]
logg = star["logg"]
M_H = star["M_H"]
vturb = ufloat(2, 0.5) if star["vturb"] is None else star["vturb"]
vals = np.full((nsamples, 4), np.nan)
for i, p in enumerate([Teff, logg, M_H, vturb]):
vals[:, i] = np.random.normal(p.n, p.s, nsamples)
if (i == 0) or (i == 3): # removing negative values
n_max = 100
j = 0
while np.any(vals[:, i] < 0):
idx = np.where(vals[:, i] < 0, True, False)
vals[:, i][idx] = np.random.normal(p.n, p.s, sum(idx))
if j >= n_max:
raise RuntimeError("failed to draw stellar parameters")
j += 1
return vals
# Interpolating ATLAS and PHOENIX LD curves
def intensity_profiles(star, ip_interp, samples, main_dir="", save=False):
intensity_profiles_dict = {}
RF_list = list(ip_interp.keys())
pbar = tqdm.tqdm(total=len(RF_list)*(len(ip_interp[RF_list[0]])),
desc="Computing intensity profiles", dynamic_ncols=True)
for RF in RF_list:
intensity_profiles_RF = {}
for FT in ip_interp[RF]:
lndi_mu, lndi_I = ip_interp[RF][FT]
if lndi_I is None:
continue
if FT.startswith("A"):
mu = lndi_mu
I = lndi_I(samples)
mu = np.tile(mu, (len(I), 1))
elif FT.startswith("P"):
mu = lndi_mu(samples[:, :-1])
I = lndi_I(samples[:, :-1])
idx = np.isfinite(mu) & np.isfinite(I)
mu = mu[idx]
I = I[idx]
idx = np.argsort(mu)
mu = mu[idx]
I = I[idx]
intensity_profiles_RF[FT] = (mu, I)
pbar.update()
intensity_profiles_dict[RF] = intensity_profiles_RF
if save:
# saving intensity profiles
rf_name = os.path.splitext(os.path.basename(RF))[0]
fn = os.path.join(main_dir,
"{}_intensity_profiles__{}.fits"
.format(star["Name"].replace(" ", "_"), rf_name))
hdu = fits.PrimaryHDU()
for k, v in star.items():
hdu.header[k] = str(v)
hdu.header["RF"] = RF
hdulist = fits.HDUList([hdu])
for mn, (mu, I) in intensity_profiles_RF.items():
tbl = table.Table(data=[mu, I], names=["mu", "Intensity"])
hdu = fits.table_to_hdu(tbl)
hdu.name = mn
hdulist.append(hdu)
hdulist.writeto(fn, overwrite=True)
os.system("gzip '{}'".format(fn))
pbar.close()
return intensity_profiles_dict
# plotting and computing log-likelihood
def ld_linear_model(mu, a):
return 1.0-a*(1.0-mu)
def ld_square_root_model(mu, s1, s2):
return 1.0-s1*(1.0-mu)-s2*(1.0-np.sqrt(mu))
def ld_quadratic_model(mu, u1, u2):
return 1.0-u1*(1.0-mu)-u2*(1.0-mu)**2
def ld_kipping2013_model(mu, q1, q2):
u1 = 2*np.sqrt(q1)*q2
u2 = np.sqrt(q1)*(1-2*q2)
return 1.0-u1*(1.0-mu)-u2*(1.0-mu)**2
def ld_three_parameter_model(mu, b1, b2, b3):
return 1.0-b1*(1.0-mu)-b2*(1.0-np.sqrt(mu)**3)-b3*(1-mu**2)
def ld_non_linear_model(mu, c1, c2, c3, c4):
return (1.0-c1*(1.0-np.sqrt(mu))-c2*(1.0-mu)-c3*(1.0-np.sqrt(mu)**3)
- c4*(1-mu**2))
def ld_logarithmic_model(mu, l1, l2):
return 1.0-l1*(1.0-mu)-l2*mu*np.log(mu)
def ld_exponential_model(mu, e1, e2):
return 1.0-e1*(1.0-mu)-e2/(1.0-np.exp(mu))
def ld_power2_model(mu, p1, p2):
return 1.0-p1*(1.0-mu**p2)
def chi2(mu, I, u1, u2, ld_model):
return np.sum((I-ld_model(mu, u1, u2))**2)
def lnlike(mu, I, u1, u2, fit_func):
u1_best, u2_best = fit_func(mu, I)
chi2_min = chi2(mu, I, u1_best.n, u2_best.n)
chi2_val = chi2(mu, I, u1, u2)
ll = -0.5*(chi2_val/chi2_min-1)
return ll
ld_laws = {"linear": (ld_linear_model, glds.lds.fit_linear),
"square-root": (ld_square_root_model, glds.lds.fit_square_root),
"quadratic": (ld_quadratic_model, glds.lds.fit_quadratic),
"kipping2013": (ld_kipping2013_model, glds.lds.fit_kipping2013),
"three-parameter": (ld_three_parameter_model,
glds.lds.fit_three_parameter),
"non-linear": (ld_non_linear_model, glds.lds.fit_non_linear),
"logarithmic": (ld_logarithmic_model, glds.lds.fit_logarithmic),
"exponential": (ld_exponential_model, glds.lds.fit_exponential),
"power-2": (ld_power2_model, glds.lds.fit_power2)}
def plot_intensity_profiles(star, intensity_profiles, ld_law="quadratic",
main_dir="", FT_idx=2):
ld_model, fit_func = ld_laws[ld_law]
FT_names = ["", "(discarded points with $\\mu \\leq 0.05$ $-$ Sing 2010)",
"(100-point interpolation $-$ Claret & Bloemen 2011)"]
fn_suffix = ["", "_mu>0.05", "_interp100"]
kwargs = dict(s=5, alpha=.2, zorder=1)
RF_list = list(intensity_profiles.keys())
for RF in RF_list:
rf_name = os.path.splitext(os.path.basename(RF))[0]
fn = ("{}_intensity_profiles__{}{}.png"
.format(star["Name"].replace(" ", "_"), rf_name,
fn_suffix[FT_idx]))
if rf_name.lower().startswith("uniform"):
rf_name = ("top-hat filter from {} A to {} A"
.format(*rf_name.split("_")[1:]))
elif rf_name.lower().startswith("gaussian"):
rf_name = ("gaussian filter centered on {} A (FWHM = {} A)"
.format(*rf_name.split("_")[1:]))
ttl = ("Intensity profile for {}\nwith {}\n"
.format(star["Name"], rf_name.replace("_", " ")))
fig, ax = plt.subplots(figsize=(6, 4.5), gridspec_kw=dict(top=0.83))
fig.suptitle(ttl, fontsize=13)
fig.text(0.5, 0.87, FT_names[FT_idx], fontsize=10,
ha="center", va="center")
mu_all, I_all = np.array([]), np.array([])
n = 0
for FT in intensity_profiles[RF]:
FT_flags = [FT in ["A17", "P"], # original data points
FT.endswith("S"), # Sing 2010 (mu ≥ 0.05)
FT.endswith("100")] # Claret & Bloemen 2011 (interp.)
if not FT_flags[FT_idx]:
continue
if "A" in FT:
mn, c, ls = "ATLAS", "C0", ":"
elif "P" in FT:
mn, c, ls = "PHOENIX", "C1", "--"
else:
mn, c, ls = "??", "C2", "-."
mu, I = intensity_profiles[RF][FT]
mu_all = np.append(mu_all, mu)
I_all = np.append(I_all, I)
ldc = [ldc_i.n for ldc_i in fit_func(mu, I)]
ax.scatter(mu, I, c=c, label="{} data".format(mn), **kwargs)
lbl_pattern = ("{{}} {{}} model: ({0})"
.format(", ".join(["{:.3f}" for ldc_i in ldc])))
lbl = lbl_pattern.format(mn, ld_law, *ldc)
ax.plot(mu, ld_model(mu, *ldc), c=LINECOLOR, ls=ls, label=lbl,
zorder=100)
n += 1
if n > 1:
ldc = [ldc_i.n for ldc_i in fit_func(mu_all, I_all)]
lbl_pattern = ("Joint {{}} model: ({0})"
.format(", ".join(["{:.3f}" for ldc_i in ldc])))
lbl = lbl_pattern.format(ld_law, *ldc)
ax.plot(mu, ld_model(mu, *ldc), c=LINECOLOR, ls="-", label=lbl,
zorder=100)
i = 1.1
ax.set_ylabel(r"$\frac{\mathcal{I}\left(\mu\right)}{\mathcal{I}_0}$",
fontsize=14*i, rotation=0, labelpad=15)
ax.set_xlabel(r"$\mu=\cos\!\left(\theta\right)$", fontsize=10*i)
ax.set_ylim(-0.01, 1.01)
ax.set_xlim(-0.01, 1.01)
for i, txt in enumerate(["limb", "center"]):
ax.text(i, -0.11*np.ptp(ax.get_ylim()), "({})".format(txt),
ha="center", va="center", fontsize=10, style="italic")
ax.legend(loc="lower right", fontsize=8)
fig.savefig(os.path.join(main_dir, fn), dpi=300)
return fig