-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtools_delens.py
executable file
·295 lines (203 loc) · 10.5 KB
/
tools_delens.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
# Linear template delensing
import numpy as np
import healpy as hp
import pickle
import tqdm
# from cmblensplus
import curvedsky
import misctools
# local
import prjlib
import tools_lens
import tools_multitracer
# Define delensing filenames
class delens:
def __init__(self,olmax=2048,elmin=20,elmax=2048,klmin=20,klmax=2048,nside=2048,klist=['TT'],kfltr='',etype=''):
conf = misctools.load_config('DELENSING')
# etype
self.etype = etype
# kappa type
self.klist = klist
self.kfltr = kfltr
#Newton method iteration number for obtaining anti-deflection angle in remapping
self.nremap = 3
# minimum/maximum multipole of E and kappa in lensing template construction
self.elmin = conf.getint('elmin',elmin)
self.elmax = conf.getint('elmax',elmax)
self.klmin = conf.getint('klmin',klmin)
self.klmax = conf.getint('klmax',klmax)
# output template
self.olmax = conf.getint('dklmax',olmax)
#remapping Nside/Npix for lensing template construction / remapping
self.nside = nside
self.npix = 12*self.nside**2
self.l = np.linspace(0,self.olmax,self.olmax+1)
def fname(self,qtag,mlist,etag,doreal):
#set directory
d = prjlib.data_directory()
ids = prjlib.rlz_index(doreal=doreal)
# delensing internal tag
ltag = 'le'+str(self.elmin)+'-'+str(self.elmax)+'_lk'+str(self.klmin)+'-'+str(self.klmax)
ttag = ltag + '_' + self.kfltr + '_' + qtag + '_' + '-'.join(mlist.keys()) + '_' + etag
#alm of lensing template B-modes
self.falm, self.fwlm, self.cl = {}, {}, {}
for k in self.klist:
self.falm[k] = [d['del']+'alm/alm_'+k+'_'+ttag+'_'+x+'.pkl' for x in ids]
self.cl[k] = [d['del']+'aps/rlz/cl_'+k+'_'+ttag+'_'+x+'.dat' for x in ids]
self.gtag = '_ideal'
# correlation coeff of templates
self.frho = d['del'] + 'aps/rho_' + '-'.join(self.klist) + '_' + ttag
def init_template(qtag,mlist,etag,doreal,**kwargs):
# setup parameters for lensing reconstruction (see cmblensplus/utils/quad_func.py)
dobj = delens(**kwargs)
delens.fname(dobj,qtag,mlist,etag,doreal)
return dobj
def diag_wiener(pqf,clkk,dlmin,dlmax,kL=None,Al=None,klist=['TT','TE','EE','EB']): #kappa filter (including kappa->phi conversion)
wlk = {}
for k in tqdm.tqdm(klist,ncols=100,desc='load diag wiener filter'):
wlk[k] = np.zeros((dlmax+1))
if k in ['TT','TE','EE','EB']:
if Al is None:
Nl = np.loadtxt(pqf[k].al,unpack=True)[1]
else:
Nl = Al[k]
for l in range(dlmin,dlmax+1):
wlk[k][l] = clkk[l]/(clkk[l]+Nl[l])
if kL is not None:
wlk[k][l] /= kL[l]
elif k == 'comb':
l = np.linspace(0,dlmax,dlmax+1)
kL = l*(l+1.)*.5
wlk[k][dlmin:dlmax+1] = 1./kL[dlmin:dlmax+1]
else:
wlk[k][dlmin:dlmax+1] = 1.
return wlk
def template_alm(rlz,klist,qf,elmin,elmax,klmin,klmax,fElm,fdlm,wlk,fgalm='',olmax=2048,glmax=2008,klist_cmb=['TT','TE','EE','EB'],**kwargs_ov):
for k in tqdm.tqdm(klist,ncols=100,desc='template:'):
for i in tqdm.tqdm(rlz,ncols=100,desc='each rlz ('+k+')',leave=False):
if misctools.check_path(fdlm[k][i],**kwargs_ov): continue
# load E mode
wElm = pickle.load(open(fElm[i],"rb"))[:elmax+1,:elmax+1]
#wElm = pickle.load(open(fElm[i].replace('base_maskv3_a5.0deg','base'),"rb"))[:elmax+1,:elmax+1]
wElm[:elmin,:] = 0.
# load kappa
if k in klist_cmb:
if qf[k].mfb is not None:
wplm = wlk[k][:klmax+1,None] * tools_lens.load_klms( qf[k].alm[i], klmax, fmlm=qf[k].mfb[i] )
else:
wplm = wlk[k][:klmax+1,None] * tools_lens.load_klms( qf[k].alm[i], klmax )
elif k == 'ALLid':
Glm = np.load( fgalm[i] )
glm = 0.*wElm
glm[20:glmax+1,:glmax+1] = curvedsky.utils.lm_healpy2healpix( len(Glm), Glm, glmax )[20:,:]
wplm = glm * wlk[k][:klmax+1,None] #* kL[:dlmax+1,None]
wplm[:klmin,:] = 0.
elif k == 'comb':
glm = 0.*wElm
glm[:glmax,:glmax] = pickle.load(open(fgalm[i],"rb"))
wplm = glm * wlk[k][:klmax+1,None]
wplm[:klmin,:] = 0.
# construct lensing B-mode template
dalm = curvedsky.delens.lensingb( olmax, elmin, elmax, klmin, klmax, wElm, wplm )
# save to file
pickle.dump((dalm),open(fdlm[k][i],"wb"),protocol=pickle.HIGHEST_PROTOCOL)
def template_aps(rlz,fdlm,fBlm,fcl,W,olmax=2048,klist=['TT','TE','EE','EB'],**kwargs_ov):
npix = len(W)
nside = int(np.sqrt(npix/12.))
for k in tqdm.tqdm(klist,ncols=100,desc='template aps'):
for i in tqdm.tqdm(rlz,ncols=100,desc='each rlz ('+k+')',leave=False):
if misctools.check_path(fcl[k][i],**kwargs_ov): continue
dalm = pickle.load(open(fdlm[k][i],"rb"))[0:olmax+1,0:olmax+1]
wdlm = curvedsky.utils.mulwin_spin(nside,olmax,olmax,2,0*dalm,dalm,W)[1]
Balm = pickle.load(open(fBlm[i],"rb"))[:olmax+1,:olmax+1]
wBlm = curvedsky.utils.mulwin_spin(nside,olmax,olmax,2,0*Balm,Balm,W)[1]
clbb = curvedsky.utils.alm2cl(olmax,wBlm)
cldd = curvedsky.utils.alm2cl(olmax,wdlm)
clbd = curvedsky.utils.alm2cl(olmax,wdlm,wBlm)
np.savetxt(fcl[k][i],np.array((clbb,cldd,clbd)).T)
def compute_coeff(rlz,fdlm,fblm,frho,W,olmax=1024,klist=['TT','TE','EE','EB']):
npix = len(W)
nside = int(np.sqrt(npix/12.))
cbb = np.zeros((len(rlz),olmax+1))
vec = np.zeros((len(rlz),len(klist),olmax+1))
mat = np.zeros((len(rlz),len(klist),len(klist),olmax+1))
#bb = 0.
#mvec = np.zeros((len(klist),olmax+1))
#mmat = np.zeros((len(klist),len(klist),olmax+1))
for ii, i in enumerate(tqdm.tqdm(rlz,ncols=100,desc='compute coeff')):
dalm = {}
for k in klist:
dalm[k] = pickle.load(open(fdlm[k][i],"rb"))[0:olmax+1,0:olmax+1]
dalm[k] = curvedsky.utils.mulwin_spin(nside,olmax,olmax,2,0*dalm[k],dalm[k],W)[1]
Balm = pickle.load(open(fblm[i],"rb"))[0:olmax+1,0:olmax+1]
wBlm = curvedsky.utils.mulwin_spin(nside,olmax,olmax,2,0*Balm,Balm,W)[1]
cbb[ii,:] = curvedsky.utils.alm2cl(olmax,wBlm)
#bb += curvedsky.utils.alm2cl(olmax,wBlm)/len(rlz)
for ki, k0 in enumerate(klist):
vec[ii,ki,:] = curvedsky.utils.alm2cl(olmax,dalm[k0],wBlm)
#mvec[ki,:] += curvedsky.utils.alm2cl(olmax,dalm[k0],wBlm)/len(rlz)
for kj, k1 in enumerate(klist):
mat[ii,ki,kj,:] = curvedsky.utils.alm2cl(olmax,dalm[k0],dalm[k1])
#mmat[ki,kj,:] += curvedsky.utils.alm2cl(olmax,dalm[k0],dalm[k1])/len(rlz)
bb, mvec, mmat = np.mean(cbb,axis=0), np.mean(vec,axis=0), np.mean(mat,axis=0)
# compute correlation coefficients
rho = np.zeros(olmax+1)
for l in range(2,olmax):
rho[l] = np.dot(mvec[:,l],np.dot(np.linalg.inv(mmat[:,:,l]),mvec[:,l]))
# save to file
L = np.linspace(0,olmax,olmax+1)
np.savetxt(frho,np.array((L,bb,rho)).T)
def interface(run_del=[],kwargs_ov={},kwargs_cmb={},kwargs_qrec={},kwargs_mass={},kwargs_del={},klist_cmb=['TT','TE','EE','EB']):
freq = kwargs_cmb.pop('freq')
# //// prepare E modes //// #
if kwargs_del['etype'] == 'id':
pE = prjlib.analysis_init(t='id',freq='cocom',ntype=kwargs_cmb['ntype'])
if kwargs_del['etype'] == 'co':
#pE = prjlib.analysis_init(t='co',freq='com',fltr='cinv',ntype='base')
pE = prjlib.analysis_init(t='co',freq='com',fltr='cinv',ntype=kwargs_cmb['ntype'].replace('_iso',''))
if kwargs_del['etype'] == 'la': # LAT-only E-modes
#pE = prjlib.analysis_init(t='la',freq='com',fltr='cinv',ntype=kwargs_cmb['ntype']) # Computationally expensive!
pE = prjlib.analysis_init(t='la',freq='com',fltr='none',ntype=kwargs_cmb['ntype']) # Diagonal filtering
# //// prepare phi //// #
# define object
glob = prjlib.analysis_init( freq='com', **kwargs_cmb )
qobj = tools_lens.init_qobj( glob.stag, glob.doreal, **kwargs_qrec )
mobj = tools_multitracer.mass_tracer( glob, qobj, **kwargs_mass )
dobj = init_template( glob.stag+qobj.ltag, mobj.klist, pE.stag, glob.doreal, **kwargs_del )
# change TT to none filter case
#if glob.fltr == 'cinv':
# kwargs_cmb['fltr'] = 'none'
# P = prjlib.analysis_init( freq='com', **kwargs_cmb )
# Qobj = tools_lens.init_qobj( P.stag, P.doreal, **kwargs_qrec )
# qobj.f['TT'] = Qobj.f['TT']
# pre-filtering for CMB phi
wlk = diag_wiener( qobj.f, glob.kk, dobj.klmin, dobj.klmax, kL=glob.kL, klist=dobj.klist )
# only kcinv for TT is used
if dobj.kfltr == 'cinv':
print('does not support kfltr = cinv')
# for k in ['TT']:
# wlk[k] = 1./(1e-30+p.kL[:dobj.klmax+1])
# qobj.f[k].alm = qobj.f[k].walm # replaced with kcinv
# qobj.f[k].mfb = None
# fullsky isotropic noise
if 'iso' in glob.ntype:
for k in dobj.klist:
if k in klist_cmb:
qobj.f[k].mfb = None
# //// compute lensing template alm //// #
if 'alm' in run_del:
template_alm( glob.rlz, dobj.klist, qobj.f, dobj.elmin, dobj.elmax, dobj.klmin, dobj.klmax, pE.fcmb.alms['o']['E'], dobj.falm, wlk, fgalm=mobj.fcklm, olmax=dobj.olmax, **kwargs_ov )
if 'aps' in run_del or 'rho' in run_del:
# prepare fullsky idealistic B mode
kwargs_cmb['t'] = 'id'
kwargs_cmb['ntype'] = 'cv'
pid = prjlib.analysis_init(**kwargs_cmb)
Wsa, __ = prjlib.window('sa')
Wla, __ = prjlib.window('la',ascale=0.)
Wsa *= hp.pixelfunc.ud_grade(Wla,512)
if 'aps' in run_del:
# compute lensing template spectrum projected on SAT area #
template_aps( glob.rlz, dobj.falm, pid.fcmb.alms['o']['B'], dobj.cl, Wsa, olmax=dobj.olmax, klist=dobj.klist, **kwargs_ov ) # ignore E-to-B leakage
if 'rho' in run_del:
# compute optimal combination weights
compute_coeff( glob.rlz, dobj.falm, pid.fcmb.alms['o']['B'], dobj.frho, Wsa, olmax=dobj.olmax, klist=dobj.klist )