def set_cmap(self, cmap):
""" Change the colormap of the image """
if cmap == 'autumn':
plt.autumn()
elif cmap == 'spring':
plt.spring()
self.canvas.show()
def main():
observations = [pfd_snr.Observation(pfdfn) for pfdfn in sys.argv[1:]]
data = np.array([(o.snr, o.ra, o.dec) for o in observations])
if debug:
print "data:"
for z in data:
print "\tSNR:", z[0], "RA:", z[1], "Dec:", z[2]
# Init guess is max SNR, weighted avg of RA, weighted avg of Dec
data_T = data.transpose()
init_params = (data_T[0].max(), (data_T[0]*data_T[1]).sum()/data_T[0].sum(), \
(data_T[0]*data_T[2]).sum()/data_T[0].sum())
if debug:
print "initial parameters:"
print "\tSNR:", init_params[0], "RA:", init_params[1], "Dec:", init_params[2]
global beam_profile
# Use gain = 1
beam_profile = estimate_snr.EstimateFWHMSNR(3.35/2.0, 1420, 100, 2, 1, 24)
result = fit(init_params, data)
if debug:
print "results:"
print "\tSNR:", result[0], "RA:", result[1], "Dec:", result[2]
psrsnr, psrra, psrdec = result
snrs, ras, decs = data.transpose()
plt.figure(figsize=(8.5,11))
plt.subplot(211)
plt.title("Fitting gridding observations to determine pulsar position")
plt.scatter((ras-psrra)*60/15.0, (decs-psrdec)*60, c=snrs, marker='o', label='_nolegend_')
plt.spring()
cbar = plt.colorbar()
cbar.set_label(r"$SNR$")
plt.scatter(np.array([0]), np.array([0]), s=100, c='k', marker=(5,1,0), \
label='Best PSR posn')
if debug:
plt.scatter(np.array([init_params[1]-psrra])*60/15.0, \
np.array([init_params[2]-psrdec])*60, \
s=100, c='w', marker=(5,1,0), label='Init PSR posn')
plt.legend(loc='best')
plt.xlabel("RA (sec) + %02.0f:%02.0f:%07.4f" % psr_utils.rad_to_hms(psrra/60.0*psr_utils.DEGTORAD))
plt.ylabel("Dec (arcsec) + %02.0f:%02.0f:%07.4f" % psr_utils.rad_to_dms(psrdec/60.0*psr_utils.DEGTORAD))
obsangseps = np.zeros(len(snrs))
for ii in range(len(snrs)):
obsangseps[ii] = angsep_arcmin(psrra, psrdec, ras[ii], decs[ii])
maxangsep = obsangseps.max()
angseps = np.linspace(0,maxangsep*1.1, 1000)
plt.subplot(212)
plt.plot(angseps, psrsnr*beam_profile.gain_at_angular_offset(angseps), 'k', zorder=-1)
plt.scatter(obsangseps, snrs, c=snrs, zorder=1)
plt.xlabel("Angular separation (arcmin)")
plt.ylabel("SNR")
plt.savefig('gridding.tmp.ps', papertype='letter', orientation='portrait')
cid_keypress = plt.gcf().canvas.mpl_connect('key_press_event', \
keypress)
plt.show()
def graph_plotting(dates_temperatures, city):
plt.figure(figsize=(15, 7))
plt.bar(range(len(dates_temperatures)),
[value - 273 for value in dates_temperatures.values()],
align='edge',
width=0.5,
color='red') # Kelvins to Celsius
plt.xticks(range(len(dates_temperatures)), list(dates_temperatures.keys()))
plt.title('{} Weather Forecast {} - {}'.format(
city,
list(dates_temperatures.keys())[0],
list(dates_temperatures.keys())[-1]))
plt.tick_params(axis='x', rotation=70)
plt.spring()
plt.show()
def plot4DGraph(clusters):
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
iter = 0
for cluster in clusters:
u_val = [obj[0] for obj in clusters[cluster]]
v_val = [obj[1] for obj in clusters[cluster]]
w_val = [obj[2] for obj in clusters[cluster]]
x_val = [obj[3] for obj in clusters[cluster]]
if iter == 0:
img1 = ax.scatter(u_val,
v_val,
w_val,
s=75,
c=x_val,
cmap=plt.winter(),
label='cluster1')
cbar = fig.colorbar(img1, shrink=0.5, aspect=10)
elif iter == 1:
img2 = ax.scatter(u_val,
v_val,
w_val,
s=75,
c=x_val,
cmap=plt.spring(),
label='cluster2')
cbar = fig.colorbar(img2, shrink=0.5, aspect=10)
else:
img3 = ax.scatter(u_val,
v_val,
w_val,
s=75,
c=x_val,
cmap=plt.gray(),
label='cluster3')
cbar = fig.colorbar(img3, shrink=0.5, aspect=10)
iter += 1
cbar.ax.get_yaxis().labelpad = 15
cbar.ax.set_ylabel('petal width in cm')
cbar.ax.get_xaxis().labelpad = 15
cbar.ax.set_xlabel('cluster' + str(iter))
ax.set_xlabel('sepal length in cm', rotation=150)
ax.set_ylabel('sepal width in cm')
ax.set_zlabel(r'petal length in cm', rotation=60)
plt.title("4D representation of clustering solution")
plt.show()
t = data[:, 1]
rvm = RVM()
rvm.learn(X, t)
print "a=", rvm.a
# 描画
import matplotlib.pyplot as plt
x = sp.linspace(0, 1, 50)
# +-標準偏差1つ分の幅の表示
meshx, meshy = sp.meshgrid(sp.linspace(0, 1, 200), sp.linspace(-1.5, 1.5, 200))
meshz = [[abs(rvm.mean(meshx[j][i]) - meshy[j][i]) <= sp.sqrt(rvm.variance(meshx[j][i]))
for i in range(len(meshx[0]))] for j in range(len(meshx))]
plt.contour(meshx, meshy, meshz, 1)
plt.spring()
# 入力
plt.scatter(X, t, label="input")
# 関連ベクトルの描画
plt.scatter(X[rvm.rv_index], t[rvm.rv_index], marker='d', color='r', label="relevance vector")
# 元の曲線を表示
y = sp.sin(2*sp.pi*x)
plt.plot(x, y, label="sin(2pix)")
# RVMの予測の平均
y_ = [rvm.mean(xi) for xi in x]
plt.plot(x, y_, label="RVM regression")
def main():
observations = [pfd_snr.Observation(pfdfn) for pfdfn in sys.argv[1:]]
data = np.array([(o.snr, o.ra, o.dec) for o in observations])
if debug:
print "data:"
for z in data:
print "\tSNR:", z[0], "RA:", z[1], "Dec:", z[2]
# Init guess is max SNR, weighted avg of RA, weighted avg of Dec
data_T = data.transpose()
init_params = (data_T[0].max(), (data_T[0]*data_T[1]).sum()/data_T[0].sum(), \
(data_T[0]*data_T[2]).sum()/data_T[0].sum())
if debug:
print "initial parameters:"
print "\tSNR:", init_params[0], "RA:", init_params[
1], "Dec:", init_params[2]
global beam_profile
# Use gain = 1
beam_profile = estimate_snr.EstimateFWHMSNR(3.35 / 2.0, 1420, 100, 2, 1,
24)
result = fit(init_params, data)
if debug:
print "results:"
print "\tSNR:", result[0], "RA:", result[1], "Dec:", result[2]
psrsnr, psrra, psrdec = result
snrs, ras, decs = data.transpose()
plt.figure(figsize=(8.5, 11))
plt.subplot(211)
plt.title("Fitting gridding observations to determine pulsar position")
plt.scatter((ras - psrra) * 60 / 15.0, (decs - psrdec) * 60,
c=snrs,
marker='o',
label='_nolegend_')
plt.spring()
cbar = plt.colorbar()
cbar.set_label(r"$SNR$")
plt.scatter(np.array([0]), np.array([0]), s=100, c='k', marker=(5,1,0), \
label='Best PSR posn')
if debug:
plt.scatter(np.array([init_params[1]-psrra])*60/15.0, \
np.array([init_params[2]-psrdec])*60, \
s=100, c='w', marker=(5,1,0), label='Init PSR posn')
plt.legend(loc='best')
plt.xlabel("RA (sec) + %02.0f:%02.0f:%07.4f" %
psr_utils.rad_to_hms(psrra / 60.0 * psr_utils.DEGTORAD))
plt.ylabel("Dec (arcsec) + %02.0f:%02.0f:%07.4f" %
psr_utils.rad_to_dms(psrdec / 60.0 * psr_utils.DEGTORAD))
obsangseps = np.zeros(len(snrs))
for ii in range(len(snrs)):
obsangseps[ii] = angsep_arcmin(psrra, psrdec, ras[ii], decs[ii])
maxangsep = obsangseps.max()
angseps = np.linspace(0, maxangsep * 1.1, 1000)
plt.subplot(212)
plt.plot(angseps,
psrsnr * beam_profile.gain_at_angular_offset(angseps),
'k',
zorder=-1)
plt.scatter(obsangseps, snrs, c=snrs, zorder=1)
plt.xlabel("Angular separation (arcmin)")
plt.ylabel("SNR")
plt.savefig('gridding.tmp.ps', papertype='letter', orientation='portrait')
cid_keypress = plt.gcf().canvas.mpl_connect('key_press_event', \
keypress)
plt.show()
