def calcProbs(entries, covar):
invcovar = np.linalg.inv(covar)
probs = np.zeros(len(entries))
for i, (zbins, median) in enumerate(entries):
mask = median == 0.
bincenters = (zbins[:-1] + zbins[1:])/2.
expected = sr.shearScaling(bincenters)
delta = median - expected
delta[mask] = 0.
probs[i] = np.exp(-0.5*np.dot(delta, np.dot(invcovar, delta)))
probs = probs/np.sum(probs)
return probs
def stackClusters(data = None, doPlot = True, cosmology = nfwutils.std_cosmology):
workdir = '/u/ki/dapple/ki06/catalog_backup_2012-02-08'
outdir = '/u/ki/dapple/subaru/doug/publication/baseline_2012-02-08'
if data is None:
data = {}
if 'items' not in data:
data['items'] = readtxtfile('worklist')
items = data['items']
if 'zbins' not in data:
data['zbins'] = np.unique(np.hstack([np.linspace(0., 1., 3.), np.linspace(1., 5., 10.), np.linspace(5., 10., 5.)]))
zbins = data['zbins']
bincenters = (zbins[1:] + zbins[:-1])/2.
if 'clusters' not in data:
clusters = {}
for cluster, filter, image in items:
key = (cluster, filter, image)
clusters[key] = {}
controller = driver.makeController()
options, args = controller.modelbuilder.createOptions()
options, args = controller.filehandler.createOptions(options = options, args = args,
workdir = workdir,
incatalog = '%s/%s.%s.%s.lensingbase.cat' % (workdir, cluster, filter, image),
cluster = cluster, filter = filter, image = image,
shapecut = True,
redseqcat = '%s/%s.%s.%s.redsequence.cat' % (workdir, cluster, filter, image))
controller.load(options, args)
stats = cPickle.load(open('%s/%s.%s.%s.out.mass15mpc.mass.summary.pkl' % (outdir, cluster, filter, image)))
mass = stats['quantiles'][50]
rs = nfwutils.RsMassInsideR(mass, 4.0, controller.zcluster, 1.5)
scaledZ, estimators = sr.scaleShear(controller, rs, 4.0, cosmology = cosmology)
bins, weights, aveshear = sr.calcZBinWeights(scaledZ, controller.pdz, estimators, zbins)
clusters[key]['weights'] = weights
clusters[key]['shears'] = aveshear
data['clusters'] = clusters
else:
clusters = data['clusters']
if 'median' not in data:
allweights = np.vstack([clusters[tuple(item)]['weights'] for item in items])
allshears = np.vstack([clusters[tuple(item)]['shears'] for item in items])
median, sig1, sig2 = sr.calcBinDistro(zbins, allweights, allshears)
data['median'] = median
data['sig1'] = sig1
data['sig2'] = sig2
else:
median = data['median']
sig1 = data['sig1']
sig2 = data['sig2']
if doPlot:
fig = pylab.figure()
ax = fig.add_axes([0.12, 0.12, 0.95 - 0.12, 0.95 - 0.12])
ax.errorbar(bincenters, median, sig2, fmt='bo')
ax.errorbar(bincenters, median, sig1, fmt='ro')
xplot = np.arange(0., np.max(zbins), 0.01)
ax.plot(xplot, sr.shearScaling(xplot), 'k-', linewidth=1.5)
ax.set_xlabel('Scaled Redshift')
ax.set_ylabel('Lensing Power')
ax.set_title('All Cluster Stack -- Restricted to Fit Data')
fig.savefig('notes/shearratio/stack_restricted_manypoints.pdf')
return fig, data
return data
def stackClusters(data = None, cosmology = nfwutils.std_cosmology, outdir = '/u/ki/dapple/subaru/doug/publication/baseline_2012-05-17'):
workdir = '/u/ki/dapple/ki06/catalog_backup_2012-05-17'
if data is None:
data = {}
if 'items' not in data:
data['items'] = readtxtfile('worklist')
items = data['items']
if 'clusters' not in data:
clusters = {}
for cluster, filter, image in items:
key = (cluster, filter, image)
clusters[key] = {}
controller = driver.makeController()
options, args = controller.modelbuilder.createOptions(zcut= None)
options, args = controller.filehandler.createOptions(options = options, args = args,
workdir = workdir,
incatalog = '%s/%s.%s.%s.lensingbase.cat' % (workdir, cluster, filter, image),
cluster = cluster, filter = filter, image = image,
shapecut = True,
redseqcat = '%s/%s.%s.%s.redsequence.cat' % (workdir, cluster, filter, image))
controller.load(options, args)
stats = cPickle.load(open('%s/%s.%s.%s.out.mass15mpc.mass.summary.pkl' % (outdir, cluster, filter, image)))
mass = stats['quantiles'][50]
rs = nfwutils.RsMassInsideR(mass, 4.0, controller.zcluster, 1.5)
scaledZ, estimators = sr.scaleShear(controller, rs, 4.0, cosmology = cosmology)
clusters[key]['scaledZ'] = scaledZ
clusters[key]['estimators'] = estimators
clusters[key]['pdz'] = controller.pdz
data['clusters'] = clusters
else:
clusters = data['clusters']
maxScaledZ = -1
for key in clusters.keys():
localMax = np.max(clusters[key]['scaledZ'])
maxScaledZ = max(localMax, maxScaledZ)
if 'zbins' not in data:
zbins = np.unique(np.hstack([np.linspace(0., 1., 3.), np.logspace(0.1, np.log10(maxScaledZ), 12.)]))
data['zbins'] = np.hstack([zbins[:-3], zbins[-1]])
zbins = data['zbins']
bincenters = (zbins[1:] + zbins[:-1])/2.
for key in clusters.keys():
bins, weights, aveshear = sr.calcZBinWeights(clusters[key]['scaledZ'], clusters[key]['estimators'], zbins)
clusters[key]['weights'] = weights
clusters[key]['shears'] = aveshear
if 'maxlike' not in data:
allweights = np.vstack([clusters[tuple(item)]['weights'] for item in items])
allshears = np.vstack([clusters[tuple(item)]['shears'] for item in items])
pointest = sr.calcBinDistro(zbins, allweights, allshears)
data['pointest'] = pointest
maxlike, sig1, sig2, maxlike_ests = sr.bootstrapBinDistro(zbins, allweights, allshears)
data['maxlike'] = maxlike
data['sig1'] = sig1
data['sig2'] = sig2
data['maxlike_ests'] = maxlike_ests
else:
maxlike = data['maxlike']
sig1 = data['sig1']
sig2 = data['sig2']
fig = pylab.figure()
ax = fig.add_axes([0.12, 0.12, 0.95 - 0.12, 0.95 - 0.12])
ax.errorbar(bincenters, maxlike, sig2, fmt='bo')
ax.errorbar(bincenters, maxlike, sig1, fmt='ro')
xplot = np.arange(0.01, np.max(zbins), 0.01)
ax.plot(xplot, sr.shearScaling(xplot), 'k-', linewidth=1.5)
ax.set_xlabel('$x = \omega_s/\omega_l$')
ax.set_ylabel('$\gamma(z)/\gamma(\infty)$')
ax.set_title('All Cluster Stack -- Maxlike Point Est')
fig.savefig('notes/shearratio/stack_maxlike.pdf')
return fig, data
def plotOneCluster(cluster, filter, image, workdir = '/u/ki/dapple/ki06/catalog_backup_2012-02-08',
outdir = '/u/ki/dapple/subaru/doug/publication/baseline_2012-02-08', data = None):
if data is None:
data = {}
if 'controller' not in data:
controller = driver.makeController()
options, args = controller.modelbuilder.createOptions(zcut = None)
options, args = controller.filehandler.createOptions(options = options, args = args,
workdir = workdir,
incatalog = '%s/%s.%s.%s.lensingbase.cat' % (workdir, cluster, filter, image),
cluster = cluster, filter = filter, image = image,
shapecut = True,
redseqcat = '%s/%s.%s.%s.redsequence.cat' % (workdir, cluster, filter, image))
controller.load(options, args)
data['controller'] = controller
else:
controller = data['controller']
if 'rs' not in data:
stats = cPickle.load(open('%s/%s.%s.%s.out.mass15mpc.mass.summary.pkl' % (outdir, cluster, filter, image)))
mass = stats['quantiles'][50]
rs = nfwutils.RsMassInsideR(mass, 4.0, controller.zcluster, 1.5)
data['rs'] = rs
else:
rs = data['rs']
if 'median' not in data:
scaledZ, estimators = sr.scaleShear(controller, rs, 4.0)
bins = np.unique(np.hstack([np.linspace(0., 1., 3.), np.linspace(1., np.max(scaledZ), 5.)]))
scaledZbins, weights, aveEst = sr.calcZBinWeights(scaledZ, controller.pdz, estimators, bins)
median, sig1, sig2 = sr.calcBinDistro(scaledZbins, weights, aveEst)
data['bins'] = bins
data['median'] = median
data['sig1'] = sig1
data['sig2'] = sig2
else:
bins = data['bins']
median = data['median']
sig1 = data['sig1']
sig2 = data['sig2']
fig = pylab.figure()
ax = fig.add_axes([0.12, 0.12, 0.95 - 0.12, 0.95 - 0.12])
bincenters = (bins[1:] + bins[:-1])/2.
ax.errorbar(bincenters, median, sig2, fmt='bo')
ax.errorbar(bincenters, median, sig1, fmt='ro')
xplot = np.arange(0., np.max(bins), 0.01)
ax.plot(xplot, sr.shearScaling(xplot), 'k-', linewidth=1.5)
ax.set_xlabel('Scaled Redshift')
ax.set_ylabel('Lensing Power')
ax.set_title('%s %s %s' % (cluster, filter, image))
fig.savefig('notes/shearratio/%s.%s.%s.pdf' % (cluster, filter, image))
return fig, data
def publicationStack(data = None):
if data is None:
data = {}
if 'standard' not in data:
stddata = {}
stdfig, stddata = stackClusters(data = stddata)
data['standard'] = stddata
else:
stddata = data['standard']
if 'eds' not in data:
edsdata = {}
edsfig, edsdata = stackClusters(data = edsdata, cosmology = nfwutils.Cosmology(omega_m = 1.0, omega_l = 0.0), outdir = '/u/ki/dapple/subaru/doug/publication/eds_cosmos_2012-05-17')
data['eds'] = edsdata
else:
edsdata = data['eds']
stdzbins = stddata['zbins']
stdbincenters = (stdzbins[1:] + stdzbins[:-1])/2.
edszbins = edsdata['zbins']
edsbincenters = (edszbins[1:] + edszbins[:-1])/2.
stdmaxlike = stddata['maxlike']
stdsig1 = stddata['sig1']
stdsig2 = stddata['sig2']
edsmaxlike = edsdata['maxlike']
edssig1 = edsdata['sig1']
edssig2 = edsdata['sig2']
fig = pylab.figure()
ax = fig.add_axes([0.125, 0.12, 0.95 - 0.125, 0.95 - 0.12])
xplot = np.arange(0.01, np.max(np.hstack([stdzbins, edszbins])), 0.01)
ax.axhline(0.0, linestyle=':', linewidth=1.25, color='k')
ax.plot(xplot, sr.shearScaling(xplot), 'k-', linewidth=1.5, label='Expected Scaling')
ax.errorbar(edsbincenters, edsmaxlike, edssig1, fmt='rs', mfc = 'None',label='$\mathbf{\Omega_{\Lambda} = 0.0}$', elinewidth=1.5,
markeredgewidth=1.5, markeredgecolor='r', marker='None')
ax.plot(edsbincenters, edsmaxlike, 'rs', markerfacecolor='None', markeredgecolor='r', markeredgewidth=1.5)
ax.errorbar(stdbincenters, stdmaxlike, stdsig1, fmt='bo', label='$\mathbf{\Omega_{\Lambda} = 0.7}$', elinewidth=1.5, markeredgewidth=1.5,
markeredgecolor='b', marker='None')
ax.plot(stdbincenters, stdmaxlike, 'bo')
thexrange = (0., 5.0)
theyrange = (-0.15, 1.75)
ax.legend(loc='upper left', numpoints=1)
ax.set_xlabel('$x = \omega_s/\omega_l$', size='x-large')
ax.set_ylabel('$\gamma_t(x)/\gamma_t(x=\infty)$', size='x-large')
fig.savefig('publication/pubstack.pdf')
######
fig2 = pylab.figure()
ax = fig2.add_axes([0.125, 0.12, 0.95 - 0.125, 0.95 - 0.12])
ax.axhline(0.0, linestyle=':', linewidth=1.25, color='k')
ax.plot(xplot, sr.shearScaling(xplot), 'k-', linewidth=1.5, label='Expected Scaling')
ax.errorbar(stdbincenters, stdmaxlike, stdsig1, fmt='bo', label='$\mathbf{\Omega_{\Lambda} = 0.7}$', elinewidth=1.5, markeredgewidth=1.5,
markeredgecolor='b', marker='None')
ax.plot(stdbincenters, stdmaxlike, 'bo')
ax.set_xlim(*thexrange)
ax.set_ylim(*theyrange)
ax.text(0.3, 1.5, '$\mathbf{\Omega_{\Lambda} = 0.7}$', size='xx-large', weight='bold')
ax.set_xlabel('$x = \omega_s/\omega_l$ ', size='x-large')
ax.set_ylabel('$\gamma_t(x)/\gamma_t(x=\infty)$', size='x-large')
fig2.savefig('publication/pubstack_lcdm.pdf')
###################
fig3 = pylab.figure()
ax = fig3.add_axes([0.125, 0.12, 0.95 - 0.125, 0.95 - 0.12])
ax.axhline(0.0, linestyle=':', linewidth=1.25, color='k')
ax.plot(xplot, sr.shearScaling(xplot), 'k-', linewidth=1.5, label='Expected Scaling')
ax.errorbar(edsbincenters, edsmaxlike, edssig1, fmt='bo', label='$\mathbf{\Omega_{\Lambda} = 0.0}$', elinewidth=1.5, markeredgewidth=1.5,
markeredgecolor='b', marker='None')
ax.plot(edsbincenters, edsmaxlike, 'bo')
ax.set_xlim(*thexrange)
ax.set_ylim(*theyrange)
ax.text(0.3, 1.5, '$\mathbf{\Omega_{\Lambda} = 0.0}$', size='xx-large', weight='bold')
ax.set_xlabel('$x = \omega_s/\omega_l$', size='x-large')
ax.set_ylabel('$\gamma_t(x)/\gamma_t(x=\infty)$', size='x-large')
fig3.savefig('publication/pubstack_eds.pdf')
return fig, fig2, fig3, data