import numpy as np
import matplotlib.cm as cm
from matplotlib.colors import rgb2hex, XKCD_COLORS
# up to ~1000 colors
ID_TO_COL = {i: col for i, col in enumerate(XKCD_COLORS.values())}
DIST_TO_COL = np.vectorize(lambda x: rgb2hex(cm.Blues_r(x)))
frame = event.frame
image_feat = recog_net.feat_extract(frame)
image_feature[counter, :] = image_feat.data.squeeze().unsqueeze_(0).numpy()
counter += 1
event = env.step(action=dict(action='RotateRight', rotation=90.0))
# low dimensional embedding
svd = TruncatedSVD(n_components=50, n_iter=10)
image_feature_SVD = svd.fit_transform(image_feature)
image_embedded = TSNE(n_components=2).fit_transform(image_feature_SVD)
# plot embeddings
ff1 = plt.figure(figsize=(8, 6))
counter = 0
for i in range(100):
plt.scatter(image_embedded[counter,0], image_embedded[counter,1],c=cm.Purples_r(i / 100.0))
counter += 1
for i in range(100):
plt.scatter(image_embedded[counter,0], image_embedded[counter,1],c=cm.Greens_r(i / 100.0))
counter += 1
for i in range(100):
plt.scatter(image_embedded[counter,0], image_embedded[counter,1],c=cm.Blues_r(i / 100.0))
counter += 1
for i in range(100):
plt.scatter(image_embedded[counter,0], image_embedded[counter,1],c=cm.Oranges_r(i / 100.0))
counter += 1
plt.tight_layout()
plt.savefig('embedding3_ResNet3.png')
plt.close(ff1)
def plot_convergence(npy, file_name, model):
converged_idx = 0
if model == 'befavor':
linspace = [
np.linspace(3.4, 14.6, 8),
np.linspace(1.0, 1.45, 6),
np.linspace(0.0, 1.00, 5),
np.linspace(0.0, 1.0, 5),
np.linspace(50, 130, 5),
np.linspace(0, 0.1, 5)
]
if model == 'aara' or model == 'acol' or model == 'bcmi':
linspace = [
np.linspace(3.4, 14.6, 8),
np.linspace(1.0, 1.45, 6),
np.linspace(0.0, 1.00, 5),
np.linspace(12., 14.0, 5),
np.linspace(0., 30.0, 5),
np.linspace(0.5, 2.0, 4),
np.linspace(0.0, 1.0, 5),
np.linspace(50, 130, 5),
np.linspace(0, 0.1, 5)
]
if model == 'beatlas':
linspace = [
np.linspace(3.8, 14.6, 5),
np.linspace(1.0, 1.45, 6),
np.linspace(0.0, 1.00, 5),
np.linspace(3.0, 4.5, 4),
np.linspace(0.0, 1.0, 5),
np.linspace(50, 130, 5),
np.linspace(0, 0.1, 5)
]
# Map the codified parameter names to their sexy latex equivalents
if model == 'befavor':
param_to_latex = dict(mass=r'$M\,[M_\odot]$',
oblat=r"$R_\mathrm{eq} / R_\mathrm{pole}$",
age=r"$H_\mathrm{frac}$",
inc=r'$\cos i$',
dis=r'$d\,[pc]$',
ebv=r'$E(B-V)$')
params = ["mass", "oblat", "age", "inc", "dis", "ebv"]
fig = plt.figure(figsize=(16, 20.6))
if model == 'aara' or model == 'acol' or model == 'bcmi':
param_to_latex = dict(mass=r'$M\,[M_\odot]$',
oblat=r"$R_\mathrm{eq} / R_\mathrm{pole}$",
age=r"$H_\mathrm{frac}$",
logn0=r'$\log \, n_0 \, [\mathrm{cm}^{-3}]$',
rdk=r'$R_\mathrm{D}\, [R_\star]$',
inc=r'$\cos i$',
nix=r'$m$',
dis=r'$d\,[\mathrm{pc}]$',
ebv=r'$E(B-V)$')
params = [
"mass", "oblat", "age", "logn0", "rdk", "nix", "inc", "dis", "ebv"
]
fig = plt.figure(figsize=(16, 24.6))
if model == 'beatlas':
param_to_latex = dict(mass=r'$M\,[M_\odot]$',
oblat=r"$R_\mathrm{eq} / R_\mathrm{pole}$",
sig0=r'$\Sigma_0$',
nix=r'$n$',
inc=r'$\cos i$',
dis=r'$d\,[pc]$',
ebv=r'$E(B-V)$')
params = ["mass", "oblat", "sig0", "nix", "inc", "dis", "ebv"]
fig = plt.figure(figsize=(16, 20.6))
chain = np.load(npy)
# chain = chain[(acceptance_fractions > 0.20) &
# (acceptance_fractions < 0.5)]
gs = gridspec.GridSpec(len(params), 3)
# gs.update(hspace=0.10, wspace=0.025, top=0.85, bottom=0.44)
gs.update(hspace=0.25)
for ii, param in enumerate(params):
these_chains = chain[:, :, ii]
max_var = max(np.var(these_chains[:, converged_idx:], axis=1))
ax1 = plt.subplot(gs[ii, :2])
ax1.axvline(0, color="#67A9CF", alpha=0.7, linewidth=2)
for walker in these_chains:
ax1.plot(np.arange(len(walker)) - converged_idx,
walker,
drawstyle="steps",
color=cm.Blues_r(
np.var(walker[converged_idx:]) / max_var),
alpha=0.5)
ax1.set_ylabel(param_to_latex[param],
fontsize=30,
labelpad=18,
rotation="vertical",
color=font_color)
# Don't show ticks on the y-axis
ax1.yaxis.set_ticks([])
# For the plot on the bottom, add an x-axis label. Hide all others
if ii == len(params) - 1:
ax1.set_xlabel("step number",
fontsize=24,
labelpad=18,
color=font_color)
else:
ax1.xaxis.set_visible(False)
ax2 = plt.subplot(gs[ii, 2])
ax2.hist(np.ravel(these_chains[:, converged_idx:]),
bins=np.linspace(ax1.get_ylim()[0],
ax1.get_ylim()[1], 20),
orientation='horizontal',
facecolor="#67A9CF",
edgecolor="none",
normed=True,
histtype='barstacked')
ax2.xaxis.set_visible(False)
ax2.yaxis.tick_right()
# print(ii)
# print(param)
ax2.set_yticks(linspace[ii])
ax1.set_ylim(ax2.get_ylim())
if ii == 0:
t = ax1.set_title("Walkers", fontsize=30, color=font_color)
t.set_y(1.01)
t = ax2.set_title("Posterior", fontsize=30, color=font_color)
t.set_y(1.01)
ax1.tick_params(axis='x',
pad=2,
direction='out',
colors=tick_color,
labelsize=22)
ax2.tick_params(axis='y',
pad=2,
direction='out',
colors=tick_color,
labelsize=22)
ax1.get_xaxis().tick_bottom()
fig.subplots_adjust(hspace=0.0,
wspace=0.0,
bottom=0.075,
top=0.9,
left=0.12,
right=0.88)
plt.savefig(file_name + '.png')
view.camera = 'arcball'
# view.camera = 'fly'
#
Globe = np.load(sys.argv[1])
Elevation = np.load(sys.argv[2])
Temps = np.load(sys.argv[3])
if len(sys.argv) > 4:
Rads =np.ones(Globe[0].shape)+0.01*Elevation
else:
Rads = np.ones(Globe[0].shape)
colors = np.empty((*Elevation.shape,4))
colors[Elevation>0.45] = cm.YlGn_r(Elevation)[Elevation>0.45]
colors[Elevation<=0.45] = cm.Blues_r(Elevation)[Elevation<=0.45]
Tempcolors = np.zeros_like(colors)
Tempcolors[np.where(Temps < 0.15)] = cm.cool(Temps, alpha = 0.8)[np.where(Temps < 0.15)]
# print(Tempcolors)
def toCart(p,t,r):
return r*np.cos(t)*np.sin(p),r*np.sin(t)*np.sin(p),r*np.cos(p)
BumpySphereCart = toCart(*Globe,Rads)
BumpySphereCart2 = toCart(*Globe,Rads+0.001)
# bumpy = geometry.MeshData(vertices = BumpySphereCart)
scene.visuals.GridMesh(*BumpySphereCart,parent=view.scene, colors = colors,shading = None)
scene.visuals.GridMesh(*BumpySphereCart2,parent=view.scene, colors = Tempcolors,shading = None)
def main(args):
df1 = pd.read_csv(args.top_csv)
df2 = pd.read_csv(args.bot_csv)
top = df1.values[:, 1:]
top = np.array(top)
max_top = np.ptp(top)
bot = df2.values[:, 1:]
bot = np.array(bot)
max_bot = np.ptp(bot)
# I need f(y) : f(mx) = T, f(0) = 0
# Scale 0 -> T to 0 to mx
def rescale(m, mx):
nm = np.where(m > 0, m, int(1e9))
T = 0.35
mn = np.min(nm)
#print(mn)
def alter(x):
if x == 0:
return 1.0
#print(x, mn)
shift_x = x - mn
sc = shift_x / mx
assert 0 <= sc and sc <= 1
rsc = sc * T
return 1 - rsc
#return 1-((x-mn)/mx)*T
#scale = lambda x: alter(x)
return alter
top_scale = rescale(top, max_top)
bot_scale = rescale(bot, max_bot)
# print(scale(m))
#keys = df1.columns[1:]
def f(c, v):
r, g, b, a = c
g = lambda x: '{:3f}'.format(x)
cstring = ','.join(map(g, c[:3]))
color = '\\cellcolor[rgb]{{{c}}}{{{v}}}'.format(c=cstring, v=v)
return color
for ii in range(1, len(df1.index) + 1):
vals = []
i = ii - 1
for j in range(0, len(df2.index) + 1):
if j == 0:
vals.append(str(df1.values[i, j]))
elif ii < j:
x = int(top_scale(df1.values[i, j]) * 255)
color = cm.Blues_r(x)
cell = f(c=color, v=df1.values[i, j])
vals.append(cell)
elif ii > j:
x = int(bot_scale(df2.values[i, j]) * 255)
color = cm.Reds_r(x)
cell = f(c=color, v=df2.values[i, j])
vals.append(cell)
else:
vals.append('')
#print(df1.values[i,j],df2.values[i,j],i,j)
# for idx, row in df1.iterrows():
#
# for key in keys:
# v = row[key]
# cell = ('\\cc{{{c:.3f}}}{{{v}}}'
# .format(c=scale(v), v=v))
# vals.append(cell)
print('&'.join(vals), end='')
print('\\\\')
def plotConvergence(npy, filename, ranges, inclFlag):
converged_idx = 0
if inclFlag: #If multiple inclinations are given
linspace = [
np.linspace(ranges[0][0], ranges[0][1], 5),
np.linspace(ranges[1][0], ranges[1][1], 5),
np.linspace(ranges[2][0], ranges[2][1], 5),
np.linspace(ranges[3][0], ranges[3][1], 5)
]
param_to_latex = dict(inc=r'$i [^{\circ}$]',
f_e=r'$f_{e}$',
v_h=r'$v_{h} [km/s]$',
v_e=r'$v_{e}$ [km/s]')
params = ["inc", "f_e", "v_h", "v_e"]
fig = plt.figure(num=None, figsize=(6, 8), dpi=100, facecolor='w',\
edgecolor='k')
else: #If only one inclination is given
linspace = [
np.linspace(ranges[0][0], ranges[0][1], 5),
np.linspace(ranges[1][0], ranges[1][1], 5),
np.linspace(ranges[2][0], ranges[2][1], 5)
]
param_to_latex = dict(f_e=r'$f_{e}$', v_h=r'$v_{h} [km/s]$',\
v_e=r'$v_{e} [km/s]$')
params = ["f_e", "v_h", "v_e"]
fig = plt.figure(num=None, figsize=(6, 8), dpi=100, facecolor='w',\
edgecolor='k')
chain = np.load(npy)
gs = gridspec.GridSpec(len(params), 3)
gs.update(hspace=0.25)
for ii, param in enumerate(params):
these_chains = chain[:, :, ii]
max_var = max(np.var(these_chains[:, converged_idx:], axis=1))
ax1 = plt.subplot(gs[ii, :2])
ax1.axvline(0, color="#67A9CF", alpha=0.7, linewidth=2)
for walker in these_chains:
ax1.plot(np.arange(len(walker)) - converged_idx,
walker,
drawstyle="steps",
color=cm.Blues_r(
np.var(walker[converged_idx:]) / max_var),
alpha=0.5)
ax1.set_ylabel(param_to_latex[param],
fontsize='xx-large',
labelpad=18,
rotation="vertical",
color=font_color)
# Don't show ticks on the y-axis
ax1.yaxis.set_ticks([])
# For the plot on the bottom, add an x-axis label. Hide all others
if ii == len(params) - 1:
ax1.set_xlabel("step number",
fontsize='x-large',
labelpad=18,
color=font_color)
else:
ax1.xaxis.set_visible(False)
ax2 = plt.subplot(gs[ii, 2])
ax2.hist(np.ravel(these_chains[:, converged_idx:]),
bins=np.linspace(ax1.get_ylim()[0],
ax1.get_ylim()[1], 20),
orientation='horizontal',
facecolor="#67A9CF",
edgecolor="none",
normed=True,
histtype='barstacked')
ax2.xaxis.set_visible(False)
ax2.yaxis.tick_right()
ax2.set_yticks(linspace[ii])
ax1.set_ylim(ax2.get_ylim())
if ii == 0:
t = ax1.set_title("Walkers", fontsize='x-large', color=font_color)
t.set_y(1.01)
t = ax2.set_title("Posterior",
fontsize='x-large',
color=font_color)
t.set_y(1.01)
ax1.tick_params(axis='x',
pad=2,
direction='out',
colors=tick_color,
labelsize='large')
ax2.tick_params(axis='y',
pad=2,
direction='out',
colors=tick_color,
labelsize='large')
ax1.get_xaxis().tick_bottom()
fig.subplots_adjust(hspace=0.0,
wspace=0.0,
bottom=0.10,
top=0.93,
left=0.12,
right=0.87)
#plt.show()
plt.savefig(filename + '.png')
def MakeColArr(Elv, LI, newL):
colorArr = np.zeros((newL[1], newL[0], 4))
colorArr[LI.T] = cm.YlGn_r(Elv.T)[LI]
colorArr[np.invert(LI.T)] = cm.Blues_r(Elv.T)[np.invert(LI.T)]
return colorArr
t = np.linspace(0, 20, tn + 1).reshape((1, tn + 1))
Space, Time = np.meshgrid(s, t)
s_i = s - c * (t)
wave_surf = wave_phi(s_i)
wave_surf = wave_surf.T
"""
plot original wave
"""
figcont += 1
fig = plt.figure(figsize=(8, 4))
ax = Axes3D(fig)
ax.plot_surface(Space,
Time,
wave_surf,
facecolors=cm.Blues_r(wave_surf / 4 * wave_surf.max()),
antialiased=False,
shade=False)
ax.view_init(elev=40, azim=235)
ax.set_zlim(-2, 2)
ax.set_zticks(np.array([0, 1, 2]))
ax.grid(False)
plt.xlim(-0.05, 1)
plt.ylim(-0.5, 20)
plt.xlabel("space(s)", size=12)
plt.ylabel("time(t)", size=12)
plt.xticks(np.linspace(0, 1, 6))
plt.yticks(np.linspace(0, tn, tn // 5 + 1))
plt.tick_params(labelsize='10')
plt.title("true wave", fontsize=15)
plt.grid()
ax3.plot(exc_range,
inh_exc_stn_ff[inh_count],
color=cm.Greens_r(inh_count * 30),
lw=3.,
label=str(inh_val))
ax2.plot(exc_range,
inh_exc_gpe_fr[inh_count],
color=cm.Reds_r(inh_count * 30),
lw=3.)
ax4.plot(exc_range,
inh_exc_gpe_ff[inh_count],
color=cm.Greens_r(inh_count * 30),
lw=3.)
ax5.plot(exc_range,
inh_exc_stn_spec[inh_count],
color=cm.Blues_r(inh_count * 30),
lw=3.,
label=str(inh_val))
ax6.plot(exc_range,
inh_exc_gpe_spec[inh_count],
color=cm.Blues_r(inh_count * 30),
lw=3.)
ax1.legend(prop={'size': 7.}, title='BS frac. GPe')
ax1.set_ylabel('Firing rate (Bq)')
ax3.legend(prop={'size': 7.}, title='BS frac. GPe')
ax3.set_ylabel('Fano Factor')
ax5.legend(prop={'size': 7.}, title='BS frac. GPe')
ax5.set_ylabel('spectral entropy')
ax6.set_xlabel('BS ratio STN')
