def _maybe_clip(self, efitter, dfitter, pars, fracdev):
"""
allow the user to send a s/n above which the clip
is applied. Default is effectively no clipping,
since fracdev_s2n_clip_min defaults to 1.0e9
"""
eres=efitter.get_result()
dres=dfitter.get_result()
s2n_max=max( eres['s2n_w'], dres['s2n_w'] )
clip_min = pars.get('fracdev_s2n_clip_min',1.e9)
if s2n_max > clip_min:
print(" clipping")
frange=pars.get('fracdev_range',[-2.0, 2.0])
fracdev_clipped = fracdev.clip(min=frange[0],max=frange[1])
else:
fracdev_clipped = 0.0 + fracdev
if False and s2n_max > 50:
import images
images.multiview(self.gal_obs.image)
key=raw_input('hit a key: ')
if key=='q':
stop
return fracdev_clipped
def make_galaxy_image(self, dindex, show=False, get_obj=False):
"""
Make the galaxy image object
"""
rindex=self._index_list[dindex]
gal_obj=self.make_galaxy_obj(rindex)
# apply center offset in pixel coords
dx = self._randu() - 0.5
dy = self._randu() - 0.5
gal_im_obj=gal_obj.draw(dx=self._pixel_scale, offset=(dx,dy))
gal_im_obj.addNoise(self._rand_gauss_obj)
#sky_var=gal_obj.noise.applyWhiteningTo(gal_im_obj)
#sky_sigma = numpy.sqrt(sky_var)
sh=gal_im_obj.array.shape
#wt=numpy.zeros(sh) + 1./sky_sigma**2
wt=numpy.zeros(sh) + 1./self._sky_sigma0**2
#print 'sky_sigma0:',self._sky_sigma0,'sky_sigma:',sky_sigma
if show:
import images
print('max image:',gal_im_obj.array.max())
images.multiview(gal_im_obj.array)
if get_obj:
return gal_obj, gal_im_obj, wt
else:
return gal_im_obj, wt
def _make_symmetrized_image(im_input):
"""
add a version of itself roated by 90,180,270 degrees
"""
im = im_input.copy()
im += numpy.rot90(im_input, k=1)
im += numpy.rot90(im_input, k=2)
im += numpy.rot90(im_input, k=3)
im *= (1.0 / 4.0)
if False:
import images
images.multiview(im)
images.compare_images(
im_input,
im,
label1='orig',
label2='symm',
width=1000,
height=1000,
)
if 'q' == raw_input('hit a key: '):
stop
return im
def test():
import images
dguess=[{'p':0.58,'row':15.,'col':15.,'irr':1.3,'irc':-0.05,'icc':1.16},
{'p':0.19,'row':15.,'col':15.,'irr':7.7,'irc':-0.23,'icc':6.8}]
im=gmix_image.gmix2image(dguess, [31,31])
print 'counts:',im.sum()
images.multiview(im)
def _set_sigma_weight(self):
"""
find the largest
"""
N=4.0
sigma_weight=0.0
dk = 0.0
for obslist in self.mb_obs:
for obs in obslist:
kreal = obs.deconvolver.psf_kreal
kmax = util.find_kmax(kreal)
if False:
import images
images.multiview(
kreal.array,
title='psf k real',
)
print(" kmax:",kmax)
tsigma = math.sqrt(2*N)/kmax
if tsigma > sigma_weight:
sigma_weight = tsigma
dk = kreal.scale
self.sigma_weight = sigma_weight/dk
print(" sigma weight:",self.sigma_weight)
def _set_sigma_weight(self):
"""
find the largest
"""
N = 4.0
sigma_weight = 0.0
dk = 0.0
for obslist in self.mb_obs:
for obs in obslist:
kreal = obs.deconvolver.psf_kreal
kmax = util.find_kmax(kreal)
if False:
import images
images.multiview(
kreal.array,
title='psf k real',
)
print(" kmax:", kmax)
tsigma = math.sqrt(2 * N) / kmax
if tsigma > sigma_weight:
sigma_weight = tsigma
dk = kreal.scale
self.sigma_weight = sigma_weight / dk
print(" sigma weight:", self.sigma_weight)
def _make_symmetrized_image(im_input):
"""
add a version of itself roated by 90,180,270 degrees
"""
im = im_input.copy()
im += numpy.rot90(im_input, k=1)
im += numpy.rot90(im_input, k=2)
im += numpy.rot90(im_input, k=3)
im *= (1.0/4.0)
if False:
import images
images.multiview(im)
images.compare_images(
im_input,
im,
label1='orig',
label2='symm',
width=1000,
height=1000,
)
if 'q'==raw_input('hit a key: '):
stop
return im
def test_sub_pixel(ellip,theta):
"""
Round objects for now
"""
import fimage
import images
sigma_vals = linspace(1.0,3.0,10)
nsub_vals = array([1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16],dtype='i4')
data = subpixel_struct(sigma_vals.size, nsub_vals.size)
ct=0
amt=0
for j in xrange(sigma_vals.size):
sigma = sigma_vals[j]
data['sigma'][j] = sigma
Irr,Irc,Icc=util.ellip2mom(2*sigma**2, e=ellip, theta=theta)
print("sigma: %0.2f Irr: %0.2f Irc: %0.2f Icc: %0.2f" % (sigma,Irr,Irc,Icc))
d=int( numpy.ceil( 4.5*sigma*2 ) )
if (d % 2) == 0:
d += 1
cen=[(d-1)/2]*2
dims = [d,d]
# create image with super good subpixel integration
ct0=time.time()
im=fimage.model_image('gauss',dims,cen,[Irr,Irc,Icc],nsub=64)
ct += time.time()-ct0
amt0=time.time()
for i in xrange(nsub_vals.size):
nsub = nsub_vals[i]
res = admom(im, cen[0], cen[1], guess=sigma**2, nsub=nsub)
if res['whyflag'] != 0:
print(" ** Failure:'%s'" % res['whystr'])
images.multiview(im, levels=7)
#key=raw_input('hit a key: ')
#if key == 'q': return
data['Irr'][j,i] = res['Irr']
data['Irc'][j,i] = res['Irc']
data['Icc'][j,i] = res['Icc']
data['a4'][j,i] = res['a4']
data['nsub'][j,i] = nsub
#print(" nsub:",nsub)
amt += time.time()-amt0
#print("time for creation:",ct)
#print("time for admom:",amt)
outf=subpixel_file(ellip,theta,'fits')
eu.io.write(outf,data,verbose=True,clobber=True)
plot_sub_pixel(ellip,theta)
def _process_star(self, iobj, icutout):
from gmix_image import GMixEMBoot
import admom
defres=({'whyflag':ZERO_WEIGHT_PIXELS},
{'flags':ZERO_WEIGHT_PIXELS},
{'flags':ZERO_WEIGHT_PIXELS},
numpy.zeros(len(WIDTHS)) - 9999,
-9999.0)
im,ivar,cen_guess=self._get_star_data(iobj,icutout)
if im is None:
return defres
max_pixel = im.max()
widths = measure_image_width(im, WIDTHS)
cen1_guess=cen_guess
sig1_guess=sqrt(2)
ares = admom.admom(im,
cen_guess[0],
cen_guess[1],
guess=sig1_guess)
gm1=GMixEMBoot(im, 1, cen1_guess, sigma_guess=sig1_guess)
res1=gm1.get_result()
if res1['flags'] == 0:
#print cen_guess
#print res1['pars']
sig2_guess=sqrt( (res1['pars'][3] + res1['pars'][5])/2. )
cen2_guess=[res1['pars'][1], res1['pars'][2] ]
else:
sig2_guess=sig1_guess
cen2_guess=cen1_guess
gm2=GMixEMBoot(im, 2, cen2_guess, sigma_guess=sig2_guess)
res2=gm2.get_result()
if False and res2['flags'] != 0:
import images
images.multiview(im)
resp=raw_input('hit enter (q to quit): ')
if resp.lower() == 'q':
stop
return ares,res1,res2,widths,max_pixel
def _debug_image(self, im, wt):
import biggles
import images
implt=images.multiview(im,show=False,title='image')
wtplt=images.multiview(wt,show=False,title='weight')
arr=biggles.Table(2,1)
arr[0,0] = implt
arr[1,0] = wtplt
arr.show()
key=raw_input('hit a key (q to quit): ')
if key.lower() == 'q':
stop
def find_centroid(image, rng, offset=None, maxiter=200, ntry=4):
"""
use AM to fit a single gaussian
"""
dims = numpy.array(image.shape)
row0, col0 = (dims-1)/2.0
if offset is not None:
row0 += offset['row_offset']
col0 += offset['col_offset']
if False:
import images
images.multiview(image, width=800,height=800)
if 'q'==raw_input('hit a key: '):
stop
j=ngmix.UnitJacobian(row=row0, col=col0)
obs = ngmix.Observation(image, jacobian=j)
guess_T = 4.0
for i in xrange(ntry):
fitter = ngmix.admom.run_admom(obs, guess_T, rng=rng)
res=fitter.get_result()
if res['flags']==0:
break
if res['flags'] != 0:
if False:
import images
images.view(image, width=800,height=800)
if 'q'==raw_input('hit a key: '):
stop
raise TryAgainError("could not fit 1 gauss to get centroid")
pars=res['pars']
row=pars[0]
col=pars[1]
row = row0 + row
col = col0 + col
return row,col
def make_f0conv(self):
"""
Make f0 convolved with epsilon. It is on this image we will
measure new adaptive moments.
f0 will be expanded if epsilon is larger. The center will remain
the same in that expansion
"""
import scipy.signal
import images
if self.f0 is None or self.epsilon is None:
raise ValueError("Create both f0 and epsilon first")
if self.f0 is None:
raise ValueError("f0 is None")
f0 = self.f0
ep = self.epsilon
if self.verbose:
if ep.shape[0] > f0.shape[0] or ep.shape[1] > f0.shape[0]:
print("epsilon is larger than image:")
print(" epsize:",ep.shape)
print(" f0size:",f0.shape)
# if f0 is bigger, it is returned unchanged
f0_expand = images.expand(f0, ep.shape)
if self.conv == 'fft':
f0conv = scipy.signal.fftconvolve(f0_expand, ep, mode='same')
else:
f0conv = scipy.signal.convolve2d(f0_expand,
ep,
#old_behavior=False,
mode='same')
# trim back to the original size
if (f0conv.shape[0] > self.image.shape[0]
or f0conv.shape[1] > self.image.shape[1]):
f0conv = f0conv[ 0:self.image.shape[0], 0:self.image.shape[1] ]
if self.debug:
print("trimming back")
self.f0conv = f0conv
if self.debug:
print("f0.shape",self.f0.shape)
print("f0conv.shape",self.f0conv.shape)
images.multiview(self.f0,title='f0')
images.multiview(self.f0conv, title='f0conv')
def make_psf(self, show=False):
"""
Make the psf image
"""
psf_obj_prepix = galsim.Gaussian(fwhm=self._psf_fwhm,
flux=1.0)
self._psf_obj = galsim.Convolve([psf_obj_prepix, self._pix_obj])
self._psf_image_obj = self._psf_obj.draw(dx=self._pixel_scale)
self._psf_image=self._psf_image_obj.array.astype('f8')
self._psf_cen = [ 0.5*self._psf_image.shape[0] ]*2
if show:
import images
images.multiview(self.psf_image)
def _doplots(self, id, mindex, icut, img_orig, img, bmask, weight):
import images
imdiff = img_orig - img
lbmask = numpy.log10(1.0 + bmask)
imlist = [img_orig, img, imdiff, lbmask, weight]
imlist = [t - t.min() for t in imlist]
titles = ['image', 'imfix', 'image-imfix', 'bmask', 'weight']
title = '%d-%06d-%02d' % (id, mindex, icut)
pname = '%s-%d-%06d-%02d.png' % (self.front, id, mindex, icut)
pname = os.path.join(self.pdir, pname)
tab = images.view_mosaic(
imlist,
titles=titles,
show=False,
)
tab.title = title
ctab = images.multiview(img, show=False)
tab[1, 2] = ctab[0, 1]
print(" writing:", pname)
tab.write_img(800, 800, pname)
def _make_interpolated_image(im, nsub, order=1):
"""
Make a new image linearly interpolated
on a nsubxnsub grid
"""
# mgrid is inclusive at end when step
# is complex and indicates number of
# points
import scipy.ndimage
zimage=scipy.ndimage.zoom(im, nsub, order=order)
if True:
import images
images.multiview(im)
images.multiview(zimage)
return zimage
def test_fit_1gauss_noisy(ellip=0.2, s2n=10000):
import images
import fimage
numpy.random.seed(35)
sigma = 1.4
T=2*sigma**2
nsig=5
dim = int(nsig*T)
if (dim % 2) == 0:
dim += 1
dims=array([dim,dim])
cen=(dims-1.)/2.
theta=23.7*numpy.pi/180.
#eta=-0.7
#ellip=(1+tanh(eta))/2
eta = ellip2eta(ellip+1.e-8)
print >>stderr,'ellip:',ellip
pars=array([cen[0],cen[1],eta,theta,1.,T])
print >>stderr,'pars'
gmix = gmix_fit.pars2gmix_coellip_pick(pars,ptype='eta')
nsub=1
im0=gmix2image_em(gmix,dims,nsub=nsub)
im,skysig = fimage.add_noise(im0, s2n,check=True)
images.multiview(im,title='nsub: %d' % nsub)
p0=pars.copy()
p0[0] += 1*(randu()-0.5) # cen0
p0[1] += 1*(randu()-0.5) # cen1
p0[2] += 0.2*(randu()-0.5) # eta
p0[3] += 0.5*(randu()-0.5) # theta radians
p0[4] += 0.1*(randu()-0.5) # p
p0[5] += 1*(randu()-0.5) # T
print_pars(pars,front='pars: ')
print_pars(p0, front='guess: ')
gf=gmix_fit.GMixFitCoellip(im, p0, ptype='eta',verbose=True)
print 'numiter:',gf.numiter
print_pars(gf.popt, front='popt: ')
print_pars(gf.perr, front='perr: ')
images.imprint(gf.pcov)
def _add_fake_uberseg(self, weight):
"""
for now just a circle
"""
sconf=self['fake_uberseg']
cen = (numpy.array(weight.shape)-1.0)/2.0
rmax = sconf['rmax_frac']*cen[0]
rsq_max = rmax**2
rows,cols = numpy.mgrid[
0:weight.shape[0],
0:weight.shape[1],
]
rows = numpy.array(rows, dtype='f8')
cols = numpy.array(cols,dtype='f8')
if sconf['type']=='circle':
axis_ratio=1.0
elif sconf['type']=='ellipse':
axis_ratio=sconf['axis_ratio']
else:
raise ValueError("bad fake useg type: '%s'" % sconf['type'])
if 'offset' in sconf:
offset = self.rng.normal(scale=sconf['offset']['sigma'],size=2)
print("offset:",offset)
cen += offset
rm = rows-cen[0]
cm = cols-cen[1]
rsq = ( rm*axis_ratio)**2 - 2*rm*cm*0.2 + cm**2
w=numpy.where(rsq > rsq_max)
weight[w] = 0.0
if False:
import images
images.multiview(weight)
if raw_input('hit a key')=='q':
stop
def _show_coadd(self, obslist):
import biggles
import images
for i,obs in enumerate(obslist):
if len(obslist) > 1:
title='image %d' % i
else:
title='coadd'
tab = biggles.Table(2, 1)
ppsf = images.multiview(obs.psf.image, title='psf',width=1000,height=1000,show=False)
pim = images.multiview(obs.image, title=title,width=1000,height=1000,show=False)
tab[0,0] = ppsf
tab[1,0] = pim
tab.show()
if raw_input('hit a key (q to quit): ')=='q':
stop
def _make_obs(self, im, psf_im):
"""
Make new Observation objects for the image and psf.
Copy out the weight maps and jacobians from the original
Observation.
parameters
----------
im: Galsim Image
psf_im: Galsim Image
returns
-------
A new Observation
"""
obs = self.obs
psf_obs = self._make_psf_obs(psf_im)
newobs = Observation(
im.array,
jacobian=obs.jacobian.copy(),
weight=obs.weight.copy(),
psf=psf_obs,
)
if False:
import images
print("orig psf im sum:", self.obs.psf.image.sum())
print("new psf im sum:", psf_im.array.sum())
images.multiview(
psf_im.array,
title='psf',
file='/u/ki/esheldon/public_html/tmp/plots/tmp.png',
)
if 'q' == raw_input('hit a key: '):
stop
return newobs
def _read_data(self):
import fitsio
import images
path=files.get_output_path(ftype='shear-stack',
run=self.run,
psfnum=self.psfnum,
shnum=self.shnum)
print path
with fitsio.FITS(path) as fits:
#res=fits[1].read()
self.psf_stack=fits[2].read()
self.psf_h=fits[2].read_header()
self.psf_skyvar=self.psf_h['skyvar']
self.im_stacks=fits[3].read()
return
subtract_median=False
images.multiview(self.psf_stack,title='psf')
edg=3
if subtract_median:
psf=self.psf_stack
pix2med = tuple( [psf[:,0:edg].ravel(), psf[:, -edg:-1].ravel(), psf[0:edg, :].ravel(), psf[-edg:-1, :].ravel() ])
pix2med=numpy.concatenate(pix2med)
self.psf_stack -= numpy.median(pix2med)
images.multiview(self.psf_stack,title='psf after')
for i in xrange(self.im_stacks.size):
images.multiview(self.im_stacks['images'][i,:,:],title='%d before' % (i+1))
if subtract_median:
im=self.im_stacks['images'][i,:,:]
pix2med = tuple( [im[:,0:edg].ravel(), im[:, -edg:-1].ravel(), im[0:edg, :].ravel(), im[-edg:-1, :].ravel() ])
pix2med=numpy.concatenate(pix2med)
self.im_stacks['images'][i,:,:] -= numpy.median(pix2med)
images.multiview(self.im_stacks['images'][i,:,:],title='%d after' % (i+1))
def _doplots(self, id, mindex, icut, img_orig, img, bmask, weight):
import images
imdiff = img_orig - img
lbmask = numpy.log10(1.0 + bmask)
imlist = [img_orig, img, imdiff, lbmask, weight]
imlist = [t - t.min() for t in imlist]
titles = ["image", "imfix", "image-imfix", "bmask", "weight"]
title = "%d-%06d-%02d" % (id, mindex, icut)
pname = "%s-%d-%06d-%02d.png" % (self.front, id, mindex, icut)
pname = os.path.join(self.pdir, pname)
tab = images.view_mosaic(imlist, titles=titles, show=False)
tab.title = title
ctab = images.multiview(img, show=False)
tab[1, 2] = ctab[0, 1]
print(" writing:", pname)
tab.write_img(800, 800, pname)
def _generate_ring_stacks(self):
sdata={}
gvals=self._get_gvals()
i=0
while i < self.nimage:
if (((i+1) % 100) == 0):
print '%d/%d' % (i+1,self.nimage)
g=gvals[i]
imd1,imd2=self._make_pair(g)
if imd1['flags'] != 0 or imd2['flags'] != 0:
print 'bad'
continue
if i==0:
sdata['image_stack'] = numpy.zeros(imd1['image'].shape)
sdata['psf_stack'] = numpy.zeros(imd1['psf_image'].shape)
sdata['image_skyvar'] = 0.0
sdata['psf_skyvar'] = 0.0
sdata['imsum']=0.0
"""
sdata['rowcen2_sum'] = 0.0
sdata['rowcolcen_sum'] = 0.0
sdata['colcen2_sum'] = 0.0
"""
sdata['row2_sum'] = 0.0
sdata['rowcol_sum'] = 0.0
sdata['col2_sum'] = 0.0
sdata['psf_imsum']=0.0
"""
sdata['psf_rowcen2_sum'] = 0.0
sdata['psf_rowcolcen_sum'] = 0.0
sdata['psf_colcen2_sum'] = 0.0
"""
sdata['psf_row2_sum'] = 0.0
sdata['psf_rowcol_sum'] = 0.0
sdata['psf_col2_sum'] = 0.0
sdata['image_stack'] += imd1['image']
sdata['image_skyvar'] += imd1['image_skyvar']
sdata['psf_stack'] += imd1['psf_image']
sdata['psf_skyvar'] += imd1['psf_skyvar']
sdata['imsum'] += imd1['imsum']
"""
sdata['rowcen2_sum'] += imd1['rowcen2_sum']
sdata['rowcolcen_sum'] += imd1['rowcolcen_sum']
sdata['colcen2_sum'] += imd1['colcen2_sum']
"""
sdata['row2_sum'] += imd1['row2_sum']
sdata['rowcol_sum'] += imd1['rowcol_sum']
sdata['col2_sum'] += imd1['col2_sum']
sdata['psf_imsum'] += imd1['psf_imsum']
"""
sdata['psf_rowcen2_sum'] += imd1['psf_rowcen2_sum']
sdata['psf_rowcolcen_sum'] += imd1['psf_rowcolcen_sum']
sdata['psf_colcen2_sum'] += imd1['psf_colcen2_sum']
"""
sdata['psf_row2_sum'] += imd1['psf_row2_sum']
sdata['psf_rowcol_sum'] += imd1['psf_rowcol_sum']
sdata['psf_col2_sum'] += imd1['psf_col2_sum']
sdata['image_stack'] += imd2['image']
sdata['image_skyvar'] += imd2['image_skyvar']
sdata['psf_stack'] += imd2['psf_image']
sdata['psf_skyvar'] += imd2['psf_skyvar']
sdata['imsum'] += imd2['imsum']
"""
sdata['rowcen2_sum'] += imd2['rowcen2_sum']
sdata['rowcolcen_sum'] += imd2['rowcolcen_sum']
sdata['colcen2_sum'] += imd2['colcen2_sum']
"""
sdata['row2_sum'] += imd2['row2_sum']
sdata['rowcol_sum'] += imd2['rowcol_sum']
sdata['col2_sum'] += imd2['col2_sum']
sdata['psf_imsum'] += imd2['psf_imsum']
"""
sdata['psf_rowcen2_sum'] += imd2['psf_rowcen2_sum']
sdata['psf_rowcolcen_sum'] += imd2['psf_rowcolcen_sum']
sdata['psf_colcen2_sum'] += imd2['psf_colcen2_sum']
"""
sdata['psf_row2_sum'] += imd2['psf_row2_sum']
sdata['psf_rowcol_sum'] += imd2['psf_rowcol_sum']
sdata['psf_col2_sum'] += imd2['psf_col2_sum']
i += 1
self._stack_data=sdata
if False:
import images
images.multiview(image_stack)
def _measure_gals_admom(self):
import admom
nbin=self.im_stacks.size
Tpsf=self.psf_ares['Irr'] + self.psf_ares['Icc']
e1p=self.psf_ares['e1']
e2p=self.psf_ares['e2']
a4p=self.psf_ares['a4']
st=zeros(nbin, dtype=[('s2n','f8'),
('e1','f8'),
('e1err','f8'),
('e2','f8'),
('e2err','f8'),
('T','f8'),
('R','f8')])
self._ares_dicts=[]
for i in xrange(nbin):
print '-'*70
s2n_min=self.im_stacks['s2n_min'][i]
s2n_max=self.im_stacks['s2n_max'][i]
print 's2n: [%.2f,%.2f]' % (s2n_min,s2n_max)
im=self.im_stacks['images'][i,:,:]
skysig=sqrt(self.im_stacks['skyvar'][i])
images.multiview(im)
cen=array(im.shape)/2.
ares = admom.admom(im, cen[0], cen[1],
sigsky=skysig,
guess=4.0,
nsub=self.nsub)
T=ares['Irr']+ares['Icc']
corr=admom.correct(T, ares['e1'], ares['e2'],ares['a4'],
Tpsf, e1p, e2p, a4p)
e1,e2,R,flags=corr
sh1=0.5*e1/self.Rshear
sh2=0.5*e2/self.Rshear
uncer=ares['uncer']/R
err=0.16/sqrt(self.im_stacks['nstack'][i])
mess='sh1: %s +/- %s sh2: %s +/- %s R: %s flags: %s'
mess=mess % (sh1,err,sh2,err,R,flags)
print mess
st['s2n'][i] = (s2n_min+s2n_max)/2.
st['e1'][i] = e1
st['e1err'][i] = err
st['e2'][i] = e2
st['e2err'][i] = err
st['R'][i] = R
st['T'][i] = T
self._ares_dicts.append(ares)
self._admom_shear = st
def show_psf_image(self):
import images
images.multiview(self.psf_image,title='psf')
def go(self, shear=None, doplots=False):
"""
measure moments on all images and perform the sum and mean
"""
fix_noise = self.kw.get('fix_noise', False)
mb_reslist = []
dk = self._dk
if not hasattr(self, 'nx'):
self.nx, self.ny = None, None
#self.nx,self.ny=66,66
for il, obslist in enumerate(self.mb_obs):
reslist = []
for obs in obslist:
ivar = obs.weight
gs_kimage, gs_ikimage = obs.deconvolver.get_kimage(
shear=shear,
dk=dk,
nx=self.nx,
ny=self.ny,
)
self.ny, self.nx = gs_kimage.array.shape
if fix_noise:
nshear = shear
if nshear is not None:
nshear = -nshear
ndk = gs_kimage.scale
rim, iim = obs.noise_deconvolver.get_kimage(
shear=nshear,
dk=ndk,
nx=self.nx,
ny=self.ny,
)
print(obs.deconvolver.dk, obs.noise_deconvolver.dk)
print(gs_kimage.array.shape, rim.array.shape)
gs_kimage += rim
# adding equal noise doubles the variance
ivar = ivar * (1.0 / 2.0)
print("shape:", gs_kimage.array.shape, "dk:", gs_kimage.scale,
"shear:", shear)
if doplots:
import images
dims = gs_kimage.array.shape
kwt = KSigmaWeight(self.sigma_weight * gs_kimage.scale)
cen = util.get_canonical_kcenter(dims)
kweight, rows, cols = kwt.get_weight(dims, cen)
pim = gs_kimage.array * kweight
#pim=gs_kimage.array
#off=int(12*(0.1/obs.deconvolver.dk))
off = int(18 * (0.1 / obs.deconvolver.dk))
pim = pim[cen[0] - off:cen[0] + off + 1,
cen[1] - off:cen[1] + off + 1]
#images.multiview(obs.image,title='image')
#images.multiview(pim,title=str(shear))#,
images.multiview(obs.deconvolver.kreal,
title=str(shear)) #,
#file='/astro/u/esheldon/www/tmp/plots/tmp.png')
res = self._measure_moments(gs_kimage, ivar)
reslist.append(res)
mb_reslist.append(reslist)
if self._force_same:
self._kdims = ny, nx
self._dk = dk
self._mb_reslist = mb_reslist
self._combine_results()
def _show(self, image):
import images
return images.multiview(image)
def process_trial(self, is2, is2n, itrial):
gsum=numpy.zeros(2)
gsum_means=numpy.zeros(2)
nsum_means=numpy.zeros(2)
wsum=numpy.zeros(2)
nsum=0
npair=self.get_npair(is2n)
gvals=self.gprior.sample1d(npair)
gmm_list=[]
for i,g in enumerate(gvals):
imd1,imd2=self._make_ring_pair(g, is2, is2n)
print '%d/%d' % (i+1,npair)
mcm1=self._run_mcm(imd1)
mcm2=self._run_mcm(imd2)
gmm1=mcm1.get_like_mixture(ngauss=3)
gmm2=mcm2.get_like_mixture(ngauss=3)
gmm_list += [gmm1,gmm2]
res1=mcm1.get_result()
res2=mcm2.get_result()
gsum_means += res1['g']
nsum_means += 1
wsum[0] += (1./res1['gcov'][0,0])
wsum[1] += (1./res1['gcov'][1,1])
gsum_means += res2['g']
nsum_means += 1
wsum[0] += (1./res2['gcov'][0,0])
wsum[1] += (1./res2['gcov'][1,1])
err=sqrt(1./wsum)
mn=gsum_means/nsum_means
print ' g1m: %s +/- %s g2m: %s +/- %s' % (mn[0],err[0],mn[1],err[1])
self._ng1=200
self._ng2=200
g1range=[mn[0]-8.*err[0], mn[0]+8.*err[0]]
g2range=[mn[1]-8.*err[1], mn[1]+8.*err[1]]
lnp_surf = self._get_cum_loglike_surf(gmm_list,
g1range=g1range,
g2range=g2range)
self.lnp_surf = lnp_surf - lnp_surf.max()
g1max,g2max=self._get_maxlike_loc(lnp_surf)
print 'maxlike: ',g1max,g2max
self._test_gmm_mean(gmm_list)
if False:
import images
plt=images.multiview(exp(self.lnp_surf),
xdr=g2range,
ydr=g1range,
xlabel='g2', ylabel='g1',
show=False)
plt_log=images.multiview(self.lnp_surf,
xdr=g2range,
ydr=g1range,
xlabel='g2', ylabel='g1',
show=False)
fname='/tmp/test.png'
fname_log='/tmp/test-log.png'
print fname
print fname_log
plt.write_img(1024,1024,fname)
plt_log.write_img(1024,1024,fname_log)
def T_rat_from_fix_rhalf_rat(sigma_psf=1.414, rhalf_exp=4.0, show=False):
"""
Generate models with r_{1/2}^{dev} = 0.6 r_{1/2}^{exp} and
measure the T ratio
parameters
----------
rhalf: float, optional
Half light radius of the exponential component
dependencies
------------
galsim
"""
import galsim
import ngmix
from numpy.random import randn
from esutil.random import srandu
import pprint
flux=100.0
s2n=1.0e6
shape=(40,40)
size=shape[0]*shape[1]
pixel_scale=1.0
j=ngmix.jacobian.UnitJacobian(0.5*shape[0], 0.5*shape[1])
rhalf_dev = 0.6*rhalf_exp
pix = galsim.Pixel(pixel_scale)
psf0=galsim.Gaussian(flux=1.0, sigma=sigma_psf)
exp0=galsim.Exponential(flux=1.0, half_light_radius=rhalf_exp)
dev0=galsim.DeVaucouleurs(flux=1.0, half_light_radius=rhalf_dev)
gal0 = galsim.Add([exp0, dev0])
gal0.setFlux(flux)
gal = galsim.Convolve([gal0, psf0, pix])
psf = galsim.Convolve([psf0, pix])
galsim_image = galsim.ImageD(shape[0], shape[1])
galsim_psf_image = galsim.ImageD(shape[0], shape[1])
gal.draw(galsim_image, dx=pixel_scale)
psf.draw(galsim_psf_image, dx=pixel_scale)
image_nonoise=galsim_image.array
psf_image=galsim_psf_image.array
skysig2 = (image_nonoise**2).sum()/s2n**2
skysig = numpy.sqrt(skysig2)
image = image_nonoise + skysig*randn(size).reshape(shape)
if show:
import images
images.multiview(psf_image,title='psf')
images.multiview(image,title='image')
imsky,sky=ngmix.em.prep_image(psf_image)
em=ngmix.em.GMixEM(imsky, jacobian=j)
Tpsf_guess=2*sigma_psf**2
psf_guess=ngmix.gmix.GMixModel([0.0, 0.0, 0.0, 0.0, Tpsf_guess, 1.0],'gauss')
em.go(psf_guess, sky)
psf_gmix = em.get_gmix()
nwalkers=1000
guess=numpy.zeros( (nwalkers, 8) )
guess[:,0] = 0.1*srandu(nwalkers)
guess[:,1] = 0.1*srandu(nwalkers)
guess[:,2] = 0.1*srandu(nwalkers)
guess[:,3] = 0.1*srandu(nwalkers)
guess[:,4] = 40*(1.0 + 0.1*srandu(nwalkers))
guess[:,5] = 30*(1.0 + 0.1*srandu(nwalkers))
guess[:,6] = flux*(1.0 + 0.1*srandu(nwalkers))
guess[:,7] = flux*(1.0 + 0.1*srandu(nwalkers))
T_prior = ngmix.priors.FlatPrior(0.001, 200)
counts_prior = ngmix.priors.FlatPrior(0.001, 1000)
wt=0*image + 1.0/skysig2
fitter=ngmix.fitting.MCMCBDC(image, wt, j, 'bdc',
psf=psf_gmix,
full_guess=guess,
T_b_prior=T_prior,
T_d_prior=T_prior,
counts_b_prior=counts_prior,
counts_d_prior=counts_prior,
nwalkers=nwalkers,
burnin=800,
nstep=100,
mca_a=3.0,
min_arate=0.3)
fitter.go()
if show:
fitter.make_plots(show=show)
res=fitter.get_result()
pars=res['pars']
perr=res['pars_err']
Tb = pars[4]
Td = pars[5]
Tb_err = perr[4]
Td_err = perr[5]
Trat = Tb/Td
Trat_err = Trat*numpy.sqrt( (Tb_err/Tb)**2 + (Td_err/Td)**2 )
pprint.pprint(res)
print
print '%s +/- %s' % (Trat, Trat_err)
def go(self, shear=None, doplots=False):
"""
measure moments on all images and perform the sum and mean
"""
fix_noise=self.kw.get('fix_noise',False)
mb_reslist=[]
dk = self._dk
if not hasattr(self,'nx'):
self.nx,self.ny=None,None
#self.nx,self.ny=66,66
for il,obslist in enumerate(self.mb_obs):
reslist=[]
for obs in obslist:
ivar=obs.weight
gs_kimage,gs_ikimage = obs.deconvolver.get_kimage(
shear=shear,
dk=dk,
nx=self.nx,
ny=self.ny,
)
self.ny,self.nx=gs_kimage.array.shape
if fix_noise:
nshear=shear
if nshear is not None:
nshear = -nshear
ndk = gs_kimage.scale
rim,iim = obs.noise_deconvolver.get_kimage(
shear=nshear,
dk=ndk,
nx=self.nx,
ny=self.ny,
)
print(obs.deconvolver.dk,obs.noise_deconvolver.dk)
print(gs_kimage.array.shape, rim.array.shape)
gs_kimage += rim
# adding equal noise doubles the variance
ivar = ivar * (1.0/2.0)
print("shape:",gs_kimage.array.shape,"dk:",gs_kimage.scale,"shear:",shear)
if doplots:
import images
dims=gs_kimage.array.shape
kwt = KSigmaWeight(self.sigma_weight*gs_kimage.scale)
cen=util.get_canonical_kcenter(dims)
kweight, rows, cols = kwt.get_weight(dims, cen)
pim=gs_kimage.array*kweight
#pim=gs_kimage.array
#off=int(12*(0.1/obs.deconvolver.dk))
off=int(18*(0.1/obs.deconvolver.dk))
pim = pim[cen[0]-off:cen[0]+off+1,
cen[1]-off:cen[1]+off+1]
#images.multiview(obs.image,title='image')
#images.multiview(pim,title=str(shear))#,
images.multiview(obs.deconvolver.kreal,title=str(shear))#,
#file='/astro/u/esheldon/www/tmp/plots/tmp.png')
res=self._measure_moments(gs_kimage, ivar)
reslist.append(res)
mb_reslist.append( reslist )
if self._force_same:
self._kdims=ny,nx
self._dk = dk
self._mb_reslist=mb_reslist
self._combine_results()
def test_fit_1gauss_galsim(ellip=0.2, s2n=10000):
import images
import galsim
import admom
numpy.random.seed(35)
#sigma = 1.4
sigma = 1
T=2*sigma**2
e = 0.2
theta=23.7
e1,e2 = etheta2e1e2(e,theta)
fimage_cov = ellip2mom(T, e=e, theta=theta)
print 'e: ',e
print 'e1:',e1
print 'e2:',e2
pixel_scale = 1.
nsig=5
dim = int(nsig*T)
if (dim % 2) == 0:
dim += 1
dims=array([dim,dim])
cen=(dims-1)/2
pix = galsim.Pixel(xw=pixel_scale, yw=pixel_scale)
gobj = galsim.Gaussian(sigma=sigma)
gobj.applyDistortion(galsim.Ellipse(e1=e1,e2=e2))
gcobj = galsim.Convolve([gobj,pix])
im0 = galsim.ImageD(int(dims[1]),int(dims[0]))
gcobj.draw(image=im0, dx=pixel_scale)
images.multiview(im0.array)
galsim_nsub=16
ares = admom.admom(im0.array,cen[0],cen[1],guess=T/2,nsub=galsim_nsub)
print 'galsim sigma:',sqrt( (ares['Irr']+ares['Icc'])/2 )
print 'galsim admom e1:',ares['e1']
print 'galsim admom e2:',ares['e2']
print 'galsim center:',ares['row'],ares['col']
fnsub=16
fim0 = model_image('gauss',dims,cen,fimage_cov,nsub=fnsub)
fares = admom.admom(fim0,cen[0],cen[1],guess=T/2,nsub=fnsub)
print 'fimage sigma:',sqrt( (fares['Irr']+fares['Icc'])/2 )
print 'fimage admom e1:',fares['e1']
print 'fimage admom e2:',fares['e2']
print 'fimage center:',fares['row'],fares['col']
return
theta=23.7*numpy.pi/180.
print >>stderr,'ellip:',ellip
pars=array([cen[0],cen[1],eta,theta,1.,T])
print >>stderr,'pars'
gmix = gmix_fit.pars2gmix_coellip_pick(pars,ptype='eta')
nsub=1
im0=gmix2image_em(gmix,dims,nsub=nsub)
im,skysig = fimage.add_noise(im0, s2n,check=True)
images.multiview(im,title='nsub: %d' % nsub)
p0=pars.copy()
p0[0] += 1*(randu()-0.5) # cen0
p0[1] += 1*(randu()-0.5) # cen1
p0[2] += 0.2*(randu()-0.5) # eta
p0[3] += 0.5*(randu()-0.5) # theta radians
p0[4] += 0.1*(randu()-0.5) # p
p0[5] += 1*(randu()-0.5) # T
print_pars(pars,front='pars: ')
print_pars(p0, front='guess: ')
gf=gmix_fit.GMixFitCoellip(im, p0, ptype='eta',verbose=True)
print 'numiter:',gf.numiter
print_pars(gf.popt, front='popt: ')
print_pars(gf.perr, front='perr: ')
images.imprint(gf.pcov)
def show_image(self):
import images
images.multiview(self.image,title='object')
def test_regauss(gal_model='gauss',
gal_T=16.0,
gal_e1=0.0,
gal_e2=0.0,
psf_T=4.0,
psf_e1=0.0,
psf_e2=0.0,
show=True):
from pprint import pprint
import fimage
import scipy.signal
nsigma=5
psf_sigma=numpy.sqrt(psf_T/2.)
psf_imsize = int( round(2*nsigma*psf_sigma) )
if (psf_imsize % 2) == 0:
psf_imsize+=1
psf_dims=[psf_imsize]*2
psf_cen=[(psf_imsize-1)/2.]*2
psf_moms = admom.ellip2mom(psf_T,e1=psf_e1,e2=psf_e2)
psf = fimage.model_image('gauss',
psf_dims,
psf_cen,
psf_moms,
nsub=16)
gal_moms = admom.ellip2mom(gal_T,e1=gal_e1,e2=gal_e2)
# for the image size, make it big enough to easily fit
# the *convolved* galaxy
gal_convolved_sigma = admom.mom2sigma(gal_T+psf_T)
gal_imsize = int( round(2*nsigma*gal_convolved_sigma) )
if gal_imsize < psf_imsize:
gal_imsize = psf_imsize
gal_cen = [ (gal_imsize-1)/2. ]*2
gal_dims = [int(gal_imsize)]*2
gal = fimage.model_image(gal_model,
gal_dims,
gal_cen,
gal_moms,
nsub=16)
levels=7
# now convolve with the PSF
imconv = scipy.signal.fftconvolve(gal, psf, mode='same')
if show:
import images
images.multiview(psf, levels=levels)
images.multiview(gal, levels=levels)
images.multiview(imconv, levels=levels)
rg = ReGauss(imconv, gal_cen[0], gal_cen[1], psf, guess=gal_T/2, guess_psf=psf_T/2)
rg.do_admom()
print("admom stats")
pprint(rg['imstats'])
rg.do_psf_admom()
print("psf admom stats")
pprint(rg['psfstats'])
rg.do_basic_corr()
print("admom corrected stats")
pprint(rg['corrstats'])
rg.do_regauss()
print("rg stats")
pprint(rg['rgstats'])
rg.do_rg_corr()
print("rg corrected stats")
pprint(rg['rgcorrstats'])
def _make_plots(self, obs):
import images
images.multiview(obs.image)
if raw_input('hit a key: ')=='q':
stop
def _measure_gals_admom(self):
import admom
nbin=self.im_stacks.size
Tpsf=self.psf_ares['Irr'] + self.psf_ares['Icc']
e1p=self.psf_ares['e1']
e2p=self.psf_ares['e2']
a4p=self.psf_ares['a4']
st=zeros(nbin, dtype=[('s2n','f8'),
('e1','f8'),
('e1err','f8'),
('e2','f8'),
('e2err','f8'),
('T','f8'),
('R','f8')])
for i in xrange(nbin):
print '-'*70
s2n_min=self.im_stacks['s2n_min'][i]
s2n_max=self.im_stacks['s2n_max'][i]
print 's2n: [%.2f,%.2f]' % (s2n_min,s2n_max)
im_real=self.im_stacks['images_real'][i,:,:]
im_imag=self.im_stacks['images_imag'][i,:,:]
imc = self._make_complex_image(im_real, im_imag)
im=self._make_cspace_image(imc)
if False and i==3:
import images
images.multiview(im)
skysig=sqrt(self.im_stacks['skyvar'][i])
cen=array(im.shape)/2.
ares = admom.admom(im, cen[0], cen[1],
sigsky=skysig,
guess=4.0,
nsub=self.nsub)
T=ares['Irr']+ares['Icc']
corr=admom.correct(T, ares['e1'], ares['e2'],ares['a4'],
Tpsf, e1p, e2p, a4p)
e1,e2,R,flags=corr
uncer=ares['uncer']/R
err2=( 0.32**2 + uncer**2)/self.im_stacks['nstack'][i]
err=sqrt(err2)
mess='e1: %s +/- %s e2: %s +/- %s R: %s flags: %s'
mess=mess % (e1,err,e2,err,R,flags)
print mess
st['s2n'][i] = (s2n_min+s2n_max)/2.
st['e1'][i] = e1
st['e1err'][i] = err
st['e2'][i] = e2
st['e2err'][i] = err
st['R'][i] = R
st['T'][i] = T
self._admom_shear = st