def test_power():
conv_map = ConvergenceMap.load("Data/unmasked.fit")
mask_profile = Mask.load("Data/mask.fit")
l_edges = np.arange(200.0,50000.0,200.0)
fig,ax = plt.subplots()
l,P_original = conv_map.powerSpectrum(l_edges)
l,P_masked = (conv_map*mask_profile).powerSpectrum(l_edges)
l,P_mask = mask_profile.powerSpectrum(l_edges)
ax.plot(l,l*(l+1)*P_original/(2*np.pi),label="Unmasked")
ax.plot(l,l*(l+1)*P_masked/(2*np.pi),label="Masked")
ax.plot(l,l*(l+1)*P_masked/(2*np.pi*(1.0 - mask_profile.maskedFraction)**2),label="Re-scaled")
ax.plot(l,l*(l+1)*P_mask/(2*np.pi),label="Mask")
ax.set_xscale("log")
ax.set_yscale("log")
ax.set_xlabel(r"$l$")
ax.set_ylabel(r"$l(l+1)P_l/2\pi$")
ax.legend(loc="lower left")
plt.savefig("power_mask.png")
plt.clf()
def maskedFraction(self): if "mask_filename" in self.kwargs.keys(): mask_profile = Mask.load(self.kwargs["mask_filename"],format=cfht_fits_loader) return mask_profile.maskedFraction else: return 0.0
def test_moments():
conv_map = ConvergenceMap.load("Data/unmasked.fit")
mask_profile = Mask.load("Data/mask.fit")
masked_map = conv_map.mask(mask_profile)
#Compute the moments in both masked and unmasked cases
mom_original = conv_map.moments(connected=True)
mom_masked = masked_map.moments(connected=True)
rel_difference = np.abs(mom_masked/mom_original - 1.0)
#Save the values and relative differences to file
np.savetxt("masked_moments.txt",np.array([mom_original,mom_masked,rel_difference]),fmt="%.2e")
def test_visualize():
conv_map = ConvergenceMap.load("Data/unmasked.fit")
mask_profile = Mask.load("Data/mask.fit")
masked_fraction = conv_map.mask(mask_profile,inplace=True)
fig,ax = plt.subplots(1,2,figsize=(16,8))
mask_profile.visualize(fig,ax[0],cmap=plt.cm.binary)
conv_map.visualize(fig,ax[1])
ax[0].set_title("Mask")
ax[1].set_title("Masked map: masking fraction {0:.2f}".format(masked_fraction))
fig.tight_layout()
fig.savefig("mask.png")
def test_minkowski():
th_minkowski = np.ogrid[-0.15:0.15:50j]
#Set up plots
fig,ax = plt.subplots(1,3,figsize=(24,8))
conv_map = ConvergenceMap.load("Data/unmasked.fit")
mask_profile = Mask.load("Data/mask.fit")
#Compute and plot the MFs for the unmasked map
v,V0,V1,V2 = conv_map.minkowskiFunctionals(th_minkowski)
ax[0].plot(v,V0,label="Unmasked")
ax[1].plot(v,V1,label="Unmasked")
ax[2].plot(v,V2,label="Unmasked")
#Compute and plot the MFs for the masked, zero padded mask
v,V0,V1,V2 = (conv_map*mask_profile).minkowskiFunctionals(th_minkowski)
ax[0].plot(v,V0,linestyle="--",label="Zero padded")
ax[1].plot(v,V1,linestyle="--",label="Zero padded")
ax[2].plot(v,V2,linestyle="--",label="Zero padded")
#Compute and plot the MFs for the masked map
masked_fraction = conv_map.mask(mask_profile,inplace=True)
v,V0,V1,V2 = conv_map.minkowskiFunctionals(th_minkowski)
ax[0].plot(v,V0,label="Masked {0:.1f}%".format(masked_fraction*100))
ax[1].plot(v,V1,label="Masked {0:.1f}%".format(masked_fraction*100))
ax[2].plot(v,V2,label="Masked {0:.1f}%".format(masked_fraction*100))
#Labels
ax[0].set_xlabel(r"$\kappa$")
ax[0].set_ylabel(r"$V_0(\kappa)$")
ax[1].set_xlabel(r"$\kappa$")
ax[1].set_ylabel(r"$V_1(\kappa)$")
ax[2].set_xlabel(r"$\kappa$")
ax[2].set_ylabel(r"$V_2(\kappa)$")
ax[0].legend(loc="upper right")
fig.tight_layout()
plt.savefig("masked_minkowski.png")
plt.clf()
def test_boundaries():
conv_map = ConvergenceMap.load("Data/unmasked.fit")
mask_profile = Mask.load("Data/mask.fit")
masked_map = conv_map.mask(mask_profile)
assert hasattr(masked_map,"_mask")
assert masked_map._masked
assert masked_map.side_angle == conv_map.side_angle
assert masked_map.data.shape == conv_map.data.shape
#Compute boundaries
perimeter_area = masked_map.maskBoundaries()
fig,ax = plt.subplots(1,3,figsize=(24,8))
#Plot gradient boundary
ax[0].imshow(masked_map._gradient_boundary,origin="lower",cmap=plt.cm.binary,interpolation="nearest",extent=[0,conv_map.side_angle.value,0,conv_map.side_angle.value])
#Plot hessian (but not gradient) boundary
ax[1].imshow(masked_map._gradient_boundary ^ masked_map._hessian_boundary,origin="lower",cmap=plt.cm.binary,interpolation="nearest",extent=[0,conv_map.side_angle.value,0,conv_map.side_angle.value])
#Plot gradient and hessian boundary
ax[2].imshow(masked_map.boundary,origin="lower",cmap=plt.cm.binary,interpolation="nearest",extent=[0,conv_map.side_angle.value,0,conv_map.side_angle.value])
ax[0].set_xlabel(r"$x$(deg)")
ax[0].set_ylabel(r"$y$(deg)")
ax[0].set_title("Gradient boundary")
ax[1].set_xlabel(r"$x$(deg)")
ax[1].set_ylabel(r"$y$(deg)")
ax[1].set_title("Hessian overhead")
ax[2].set_xlabel(r"$x$(deg)")
ax[2].set_ylabel(r"$y$(deg)")
ax[2].set_title("Full boundary: perimeter/area={0:.3f}".format(perimeter_area))
fig.tight_layout()
plt.savefig("boundaries.png")
plt.clf()
def test_pdf():
th_pdf = np.ogrid[-0.15:0.15:50j]
conv_map = ConvergenceMap.load("Data/unmasked.fit")
mask_profile = Mask.load("Data/mask.fit")
masked_map = conv_map.mask(mask_profile)
v,p_original = conv_map.pdf(th_pdf)
v,p_masked = masked_map.pdf(th_pdf)
#Plot the two histograms
plt.plot(v,p_original,label="Unmasked")
plt.plot(v,p_masked,label="Masked {0:.1f}%".format(mask_profile.maskedFraction * 100))
#Labels
plt.xlabel(r"$\kappa$")
plt.ylabel(r"$P(\kappa)$")
plt.legend(loc="upper right")
plt.savefig("masked_pdf.png")
plt.clf()
def test_peak_locations():
th_peaks = np.arange(0.24,0.5,0.01)
conv_map = ConvergenceMap.load("Data/unmasked.fit")
mask_profile = Mask.load("Data/mask.fit")
masked_map = conv_map.mask(mask_profile)
#Locate the peaks on the map
values,location = masked_map.locatePeaks(th_peaks)
#Visualize the map and the peak locations
fig,ax = plt.subplots(1,2,figsize=(16,8))
masked_map.visualize(fig=fig,ax=ax[0],colorbar=True)
masked_map.visualize(fig=fig,ax=ax[1])
ax[1].scatter(location[:,0].value,location[:,1].value,color="black")
ax[1].set_xlim(0.0,masked_map.side_angle.value)
ax[1].set_ylim(0.0,masked_map.side_angle.value)
#Save the figure
fig.tight_layout()
fig.savefig("masked_peak_locations.png")
def test_peaks():
th_peaks = np.ogrid[-0.04:0.12:50j]
conv_map = ConvergenceMap.load("Data/unmasked.fit")
mask_profile = Mask.load("Data/mask.fit")
masked_map = conv_map.mask(mask_profile)
v,pk_orig = conv_map.peakCount(th_peaks)
v,pk_masked = masked_map.peakCount(th_peaks)
#Plot the difference
plt.plot(v,pk_orig,label=r"Unmasked: $N_p=${0}".format(int(integrate.simps(pk_orig,x=v))))
plt.plot(v,pk_masked,label=r"With {0:.1f}% area masking: $N_p=${1}".format(mask_profile.maskedFraction * 100,int(integrate.simps(pk_masked,x=v))))
plt.plot(v,pk_masked/(1.0 - mask_profile.maskedFraction),label="Re-scaled")
#Labels
plt.xlabel(r"$\kappa$")
plt.ylabel(r"$dN/d\kappa$")
plt.legend(loc="upper left")
plt.savefig("masked_peaks.png")
plt.clf()