stats[fitname[redshift_bin]] = []
output[fitname[redshift_bin]] = []
c[redshift_bin].add_chain(df,
weights=weight,
**extra,
plot_contour=True,
plot_point=False,
show_as_1d_prior=False)
else:
c[redshift_bin].add_marker(params, **extra)
# Compute some summary statistics and add them to a dictionary
mean, cov = weighted_avg_and_cov(
df[[
"$\\alpha_\\parallel$",
"$\\alpha_\\perp$",
]],
weight,
axis=0,
)
corr = cov[1, 0] / np.sqrt(cov[0, 0] * cov[1, 1])
stats[fitname[redshift_bin]].append([
mean[0], mean[1],
np.sqrt(cov[0, 0]),
np.sqrt(cov[1, 1]), corr, new_chi_squared
])
output[fitname[redshift_bin]].append(
f"{realisation:s}, {mean[0]:6.4f}, {mean[1]:6.4f}, {np.sqrt(cov[0, 0]):6.4f}, {np.sqrt(cov[1, 1]):6.4f}, {corr:7.3f}, {r_s:7.3f}, {new_chi_squared:7.3f}, {dof:4d}"
)
truth = {
figname = pfn + "_" + fitname + "_bestfit.pdf" if counter == 3 else None
new_chi_squared, dof, bband, mods, smooths = model.plot(
params, display=False, figname=figname)
ps = chain[max_post, :]
best_fit = {}
for l, p in zip(model.get_labels(), ps):
best_fit[l] = p
mean, cov = weighted_avg_and_cov(
df[[
"$\\alpha_\\parallel$",
"$\\alpha_\\perp$",
"$\\Sigma_s$",
"$\\Sigma_{nl,||}$",
"$\\Sigma_{nl,\\perp}$",
]],
weight,
axis=0,
)
corr = cov[1, 0] / np.sqrt(cov[0, 0] * cov[1, 1])
print(fitname)
if "3-Poly" in fitname:
if "No-Hexa" in fitname:
output[fitname].append(
f"{kmax:3s}, {mean[0]:6.4f}, {mean[1]:6.4f}, {np.sqrt(cov[0,0]):6.4f}, {np.sqrt(cov[1,1]):6.4f}, {corr:7.3f}, {r_s:7.3f}, {chi2:7.3f}, {dof:4d}, {mean[3]:7.3f}, {mean[4]:7.3f}, {mean[2]:7.3f}, {bband[0]:7.3f}, {bband[1]:8.1f}, {bband[2]:8.1f}, {bband[3]:8.1f}, {bband[4]:8.1f}, {bband[5]:8.1f}, {bband[6]:8.1f}"
)
else:
output[fitname].append(