noe/pipeline
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

bugs + refactoring

Browse Source
This commit is contained in:
Noe Brucy
2019-06-13 12:17:14 +02:00
parent fe5c9c6ac8
commit 84fcf81dfc
2 changed files with 130 additions and 204 deletions
Download Patch File Download Diff File Expand all files Collapse all files
disk_postprocess.py
+129 -203
View File
@@ -825,7 +825,7 @@ def disk_pdf(
interactive=False,
nb_bin_hist=50,
tag="",
rad_min=0.05,
rad_min=0.075,
rad_max=0.3,
put_title=True,
do_speed=True,
@@ -863,10 +863,14 @@ def disk_pdf(
f.close()
print("maps file loaded")
# Properties
time = prop_disk["time"]
im_extent = maps_disk["im_extent"]
title = tag.split("_")[1] + " t=" + str(time)[0:5] + " (code)"
# Load coldens
coldens_map = maps_disk["coldens_z"]
# radius of the corner of the box plus a margin
rad_box = (
np.sqrt((im_extent[1] - pos_star[0]) ** 2 + (im_extent[3] - pos_star[1]) ** 2)
@@ -877,13 +881,23 @@ def disk_pdf(
rad_bins = 0.5 * (rad_bins[0:-1] + rad_bins[1:])
rad_bins = np.concatenate(([0.0], rad_bins, [rad_box]))
coldens_map = maps_disk["coldens_z"]
x = np.linspace(im_extent[0], im_extent[1], coldens_map.shape[0])
y = np.linspace(im_extent[2], im_extent[3], coldens_map.shape[1])
xx, yy = np.meshgrid(x, y)
rr = np.sqrt((xx - pos_star[0]) ** 2 + (yy - pos_star[1]) ** 2)
# Find appropriate bin for each coordinate set
bins = np.zeros(rr.shape, dtype=int)
for r in rad_bins[1:]:
bins = bins + (rr >= r).astype(int)
bins_flat = bins.flatten()
rr_flat = rr.flatten()
# Mask selecting the zone of interest
mask_map = (rr > rad_min) & (rr < rad_max)
mask_flat = mask_map.flatten()
# Additionnal maps
rho_map = maps_disk["rho_z"]
cs_map = np.sqrt(maps_disk["T_z"])
@@ -896,49 +910,45 @@ def disk_pdf(
vrad_map = (vx_map * xx_star + vy_map * yy_star) / rr
vaz_map = (vy_map * xx_star - vx_map * yy_star) / rr
maps = [coldens_map, rho_map, cs_map, v_map, vaz_map]
maps = {
"coldens": coldens_map,
"rho": rho_map,
"cs": cs_map,
"v": v_map,
"vaz": vaz_map,
}
# Find appropriate bin for each coordinate set
bins = np.zeros(rr.shape, dtype=int)
for r in rad_bins[1:]:
bins = bins + (rr >= r).astype(int)
bins_flat = bins.flatten()
rr_flat = rr.flatten()
avg_maps = {}
fluct_maps = {}
coldens_rad = np.zeros(len(rad_bins) - 1)
rho_rad = np.zeros(len(rad_bins) - 1)
cs_rad = np.zeros(len(rad_bins) - 1)
v_rad = np.zeros(len(rad_bins) - 1)
vaz_rad = np.zeros(len(rad_bins) - 1)
rads = [coldens_rad, rho_rad, cs_rad, v_rad, vaz_rad]
means = [None] * len(rads)
for i, rad_arr in enumerate(rads):
map_arr = maps[i]
for cur_map in maps:
map_arr = maps[cur_map]
# mean of all the cells in the bin
mean_bin = np.zeros(len(rad_bins) - 1)
for j in range(len(rad_bins) - 1):
rad_arr[j] = np.mean(map_arr[bins == j])
mean_bin[j] = np.mean(map_arr[bins == j])
# Add value for borders
rad_arr = np.concatenate(([rad_arr[0]], rad_arr))
mean_bin = np.concatenate(([mean_bin[0]], mean_bin))
# Retrieve the mean of the bins
# Compute the map azimuthally averaged
# use linear interpolation to improve accuracy
means[i] = (rad_bins[bins_flat + 1] - rr_flat) * rad_arr[bins_flat]
means[i] = means[i] + (rr_flat - rad_bins[bins_flat]) * rad_arr[bins_flat + 1]
means[i] = means[i] / (rad_bins[bins_flat + 1] - rad_bins[bins_flat])
avg_flat = (rad_bins[bins_flat + 1] - rr_flat) * mean_bin[bins_flat]
avg_flat = avg_flat + (rr_flat - rad_bins[bins_flat]) * mean_bin[bins_flat + 1]
avg_flat = avg_flat / (rad_bins[bins_flat + 1] - rad_bins[bins_flat])
avg_maps[cur_map] = np.reshape(avg_flat, rr.shape)
coldens_mean = means[0]
coldens_mean_map = np.reshape(coldens_mean, coldens_map.shape)
# Kill fluctuations for r < rad_min and r > rad_max
mask_map = (rr > rad_min) & (rr < rad_max)
coldens_map[np.logical_not(mask_map)] = np.nan
# Select zone of interest
avg_maps[cur_map][np.logical_not(mask_map)] = np.nan
# Compute fluctuation
fluct_maps[cur_map] = map_arr / avg_maps[cur_map]
# Histogramm : density fluctuation distribution function
dcoldens = np.log10(coldens_map.flatten() / coldens_mean)
mask_hist = (rr_flat > rad_min) & (rr_flat < rad_max)
nb_cells = np.sum(mask_hist)
dcoldens = np.log10(fluct_maps["coldens"]).flatten()
nb_cells = np.sum(mask_flat)
P.grid()
P.yscale("log")
P.ylim([0.5 / nb_cells, 1.0])
@@ -947,7 +957,7 @@ def disk_pdf(
if put_title:
P.title(title)
values, edges, _ = P.hist(
dcoldens[mask_hist],
dcoldens[mask_flat],
nb_bin_hist,
range=(-1, 3),
weights=np.ones(nb_cells) / nb_cells,
@@ -955,7 +965,7 @@ def disk_pdf(
centers = 0.5 * (edges[1:] + edges[:-1])
# Variance
var = np.var(dcoldens[mask_hist])
var = np.var(dcoldens[mask_flat])
# Compute the slope of the right part of the histogramm
mask_fit = (centers > 0.25) & (centers < 1.25) & (values > 0)
if np.sum(mask_fit > 0):
@@ -988,26 +998,27 @@ def disk_pdf(
P.savefig(path + "/dcol_hist_" + tag + "_" + str(num).zfill(5) + out_ext)
P.close()
# Derived quantities
drho = fluct_maps["rho"].flatten()
dcs = fluct_maps["cs"].flatten()
dv = fluct_maps["v"].flatten()
dvaz_kepl = abs(maps["vaz"] - v_kepl) / v_kepl
fluct_maps["vaz_kepl"] = dvaz_kepl
dmach = abs(maps["v"] - avg_maps["v"]) / maps["cs"]
fluct_maps["mach"] = dmach
dmach_mean = np.mean(dmach[mask_map].flatten())
# Fluctuations plots
rho_mean = means[1]
cs_mean = means[2]
v_mean = means[3]
vaz_mean = means[4]
drho = abs(rho_map.flatten() / rho_mean)
dcs = abs(cs_map.flatten() / cs_mean)
dvaz = abs(vaz_map.flatten() / vaz_mean)
dv = abs(v_map.flatten() / v_mean)
dmach = abs((v_map.flatten() - v_mean) / cs_map.flatten())
dmach_mean = np.mean(dmach[mask_hist])
f = open(name_prop, "w")
prop_disk["dmach_mean"] = dmach_mean
print("dmach_mean = {}".format(dmach_mean))
pickle.dump(prop_disk, f)
f.close()
# P.scatter(drho[mask_hist], dcs[mask_hist], s=0.1)
# dcs = f(drho)
P.hist2d(
np.log10(drho[mask_hist]),
np.log10(dcs[mask_hist]),
np.log10(drho[mask_flat]),
np.log10(dcs[mask_flat]),
(1000, 1000),
norm=mpl.colors.LogNorm(),
)
@@ -1023,9 +1034,10 @@ def disk_pdf(
P.savefig(path + "/dcs_" + tag + "_" + str(num).zfill(5) + out_ext)
P.close()
# dv = f(drho)
P.hist2d(
np.log10(drho[mask_hist]),
dvaz[mask_hist],
np.log10(drho[mask_flat]),
dv[mask_flat],
(1000, 1000),
norm=mpl.colors.LogNorm(),
)
@@ -1037,160 +1049,67 @@ def disk_pdf(
P.figure()
else:
P.tight_layout(pad=1)
P.savefig(path + "/dvaz_" + tag + "_" + str(num).zfill(5) + out_ext)
P.savefig(path + "/dv_" + tag + "_" + str(num).zfill(5) + out_ext)
P.close()
# Fluctuations maps
dcoldens_map = np.reshape(dcoldens, coldens_map.shape)
im = P.imshow(dcoldens_map, extent=im_extent, origin="lower", cmap="viridis")
if put_title:
P.title(title)
P.xlabel(r"$x$")
P.ylabel(r"$y$")
cbar = P.colorbar(im)
cbar.set_label(r"$log(N/\bar{N})$ (code)")
name = path + "/dcoldensmap_" + tag + "_" + format(num, "05")
name_im = name + out_ext
if interactive:
P.figure()
else:
P.tight_layout(pad=1)
P.savefig(name_im)
P.close()
dcs_map = np.reshape(dcs, coldens_map.shape)
dcs_map[np.logical_not(mask_map)] = np.nan
im = P.imshow(
np.log10(dcs_map),
extent=im_extent,
origin="lower",
cmap="RdBu_r",
vmin=-0.6,
vmax=0.6,
)
if put_title:
P.title(title)
P.xlabel(r"$x$")
P.ylabel(r"$y$")
cbar = P.colorbar(im)
cbar.set_label(r"$log(c_s/\bar{c_s})$ (code)")
name = path + "/dcsmap_" + tag + "_" + format(num, "05")
name_im = name + out_ext
if interactive:
P.figure()
else:
P.tight_layout(pad=1)
P.savefig(name_im)
P.close()
drho_map = np.reshape(drho, coldens_map.shape)
drho_map[np.logical_not(mask_map)] = np.nan
im = P.imshow(
np.log10(drho_map),
extent=im_extent,
origin="lower",
cmap="RdBu_r",
vmin=-2.0,
vmax=2.0,
)
if put_title:
P.title(title)
P.xlabel(r"$x$")
P.ylabel(r"$y$")
cbar = P.colorbar(im)
cbar.set_label(r"$log(\rho/\bar{\rho})$ (code)")
name = path + "/drhomap_" + tag + "_" + format(num, "05")
name_im = name + out_ext
if interactive:
P.figure()
else:
P.tight_layout(pad=1)
P.savefig(name_im)
P.close()
if do_speed:
dvel_kepl = abs(v_map - v_kepl) / v_kepl
dvaz_kepl = abs(vaz_map - v_kepl) / v_kepl
dvrad_vaz = abs(vrad_map / vaz_map)
dvaz_map = np.reshape(dvaz, coldens_map.shape)
dvaz_map[np.logical_not(mask_map)] = np.nan
im = P.imshow(
dvaz_kepl,
extent=im_extent,
origin="lower",
cmap="RdBu_r",
norm=mpl.colors.LogNorm(),
vmax=1.0,
vmin=0.01,
)
cbar = P.colorbar(im)
cbar.set_label(r"$|v_\varphi - v_{kepl}|/v_{kepl}$")
P.xlabel(r"$x$")
P.ylabel(r"$y$")
for cur_map in fluct_maps:
fluct_map = fluct_maps[cur_map]
if cur_map == "coldens":
im = P.imshow(
fluct_map,
norm=mpl.colors.LogNorm(),
extent=im_extent,
origin="lower",
cmap="viridis",
)
label = r"$log(N/\bar{N})$"
elif cur_map == "cs":
im = P.imshow(
np.log10(fluct_map),
extent=im_extent,
origin="lower",
cmap="RdBu_r",
vmin=-0.6,
vmax=0.6,
)
label = r"$log(c_s/\bar{c_s})$"
elif cur_map == "rho":
im = P.imshow(
np.log10(fluct_map),
extent=im_extent,
origin="lower",
cmap="RdBu_r",
vmin=-2.0,
vmax=2.0,
)
label = r"$log(\rho/\bar{\rho})$"
elif cur_map == "vaz_kepl":
im = P.imshow(
fluct_map,
extent=im_extent,
origin="lower",
cmap="RdBu_r",
norm=mpl.colors.LogNorm(),
vmax=1.0,
vmin=0.01,
)
label = r"$|v_\varphi - v_{kepl}|/v_{kepl}$"
elif cur_map == "vaz":
im = P.imshow(fluct_map, extent=im_extent, origin="lower", cmap="RdBu_r")
label = r"$v_\varphi / \bar{v_\varphi}$"
elif cur_map == "mach":
im = P.imshow(fluct_map, extent=im_extent, origin="lower", cmap="RdBu_r")
label = r"$|v - \bar{v}| / c_s$"
if put_title:
P.title(title)
name = path + "/dvaz_kepl_" + tag + "_" + format(num, "05")
name_im = name + out_ext
if interactive:
P.figure()
else:
P.tight_layout(pad=1)
P.savefig(name_im)
P.close()
im = P.imshow(
vrad_map,
extent=im_extent,
origin="lower",
cmap="RdBu_r",
norm=mpl.colors.SymLogNorm(linthresh=0.03),
vmax=8.0,
vmin=-8,
)
cbar = P.colorbar(im)
cbar.set_label(r"$v_r$")
P.xlabel(r"$x$")
P.ylabel(r"$y$")
if put_title:
P.title(title)
name = path + "/vradmap_" + tag + "_" + format(num, "05")
name_im = name + out_ext
if interactive:
P.figure()
else:
P.tight_layout(pad=1)
P.savefig(name_im)
P.close()
im = P.imshow(dvaz_map, extent=im_extent, origin="lower", cmap="RdBu_r")
cbar = P.colorbar(im)
cbar.set_label(r"$v_\varphi / \bar{v_\varphi}$")
P.xlabel(r"$x$")
P.ylabel(r"$y$")
if put_title:
P.title(title)
name = path + "/dvazmap_" + tag + "_" + format(num, "05")
cbar.set_label(label)
name = path + "/d" + cur_map + "map_" + tag + "_" + format(num, "05")
name_im = name + out_ext
if interactive:
@@ -1242,6 +1161,9 @@ def compare(
dmach = np.zeros(len(runs))
dmach_std = np.zeros(len(runs))
all_var = []
all_dmach = []
for i, run in enumerate(runs):
path_run = path + "/" + run
nb_outputs = 0
@@ -1276,6 +1198,9 @@ def compare(
var_run.append(fit["var"])
dmach_run.append(prop_disk["dmach_mean"])
all_var = all_var + var_run
all_dmach = all_dmach + dmach_run
nb_outputs = nb_outputs + 1
print(run, num, nb_outputs)
@@ -1387,17 +1312,18 @@ def compare(
P.savefig(path_out + "/varbeta_" + nums_name + out_ext)
P.close()
# dmach = f(var)
# var = f(log(dmach)
P.grid()
P.errorbar(var, dmach, xerr=var_std, yerr=dmach_std, fmt="o")
P.plot(all_dmach, all_var, "o")
P.legend()
P.ylabel(r"$<(v - \bar{v}) / c_s>$")
P.xlabel(r"$Var(\log(N / \bar(N))$")
P.xlabel(r"$<(v - \bar{v}) / c_s>$")
P.ylabel(r"$Var(\log(N / \bar(N))$")
P.yscale("log")
(a, b, rho, _, stderr) = linregress(var, dmach)
P.plot(var, a * var + b, "--", linewidth=1.5)
print("a=%e, b=%e, rho^2=%e" % (a, b, rho ** 2))
# (a, b, rho, _, stderr) = linregress(var, log(dmach)
# P.plot(var, a*var + b, '--', linewidth=1.5)
# print("a=%e, b=%e, rho^2=%e"% (a,b,rho**2))
if interactive:
P.figure()