forked from mike-a-yen/bpz-1.99.3
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy path#bpz.py#
1427 lines (1238 loc) · 46.8 KB
/
#bpz.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
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
"""
bpz: Bayesian Photo-Z estimation
Reference: Benitez 2000, ApJ, 536, p.571
Usage:
python bpz.py catalog.cat
Needs a catalog.columns file which describes the contents of catalog.cat
"""
from useful import *
rolex=watch()
rolex.set()
#from Numeric import *
from numpy import *
from bpz_tools import *
from string import *
import os,glob,sys
import time
import pickle
import shelve
from coetools import pause, params_cl
def seglist(vals, mask=None):
"""Split vals into lists based on mask > 0"""
if mask == None:
mask = greater(vals, 0)
lists = []
i = 0
lastgood = False
list1 = []
for i in range(len(vals)):
if mask[i] == False:
if lastgood:
lists.append(list1)
list1 = []
lastgood = False
if mask[i]:
list1.append(vals[i])
lastgood = True
if lastgood:
lists.append(list1)
return lists
# Initialization and definitions#
#Current directory
homedir=os.getcwd()
#Parameter definition
pars=params()
pars.d={
'SPECTRA':'CWWSB4.list', # template list
#'PRIOR': 'hdfn_SB', # prior name
'PRIOR': 'hdfn_gen', # prior name
'NTYPES':None, # Number of Elliptical, Spiral, and Starburst/Irregular templates Default: 1,2,n-3
'DZ': 0.01, # redshift resolution
'ZMIN': 0.01, # minimum redshift
'ZMAX': 10., # maximum redshift
'MAG': 'yes', # Data in magnitudes?
'MIN_MAGERR': 0.001, # minimum magnitude uncertainty --DC
'ODDS': 0.95, # Odds threshold: affects confidence limits definition
'INTERP': 0, # Number of interpolated templates between each of the original ones
'EXCLUDE': 'none', # Filters to be excluded from the estimation
'NEW_AB': 'no', # If yes, generate new AB files even if they already exist
'CHECK': 'yes', # Perform some checks, compare observed colors with templates, etc.
'VERBOSE': 'yes', # Print estimated redshifts to the standard output
'PROBS': 'no', # Save all the galaxy probability distributions (it will create a very large file)
'PROBS2': 'no', # Save all the galaxy probability distributions P(z,t) (but not priors) -- Compact
'PROBS_LITE': 'yes', # Save only the final probability distribution
'GET_Z': 'yes', # Actually obtain photo-z
'ONLY_TYPE':'no', # Use spectroscopic redshifts instead of photo-z
'MADAU':'yes', #Apply Madau correction to spectra
'Z_THR':0, #Integrate probability for z>z_thr
'COLOR':'no', #Use colors instead of fluxes
'PLOTS':'no', #Don't produce plots
'INTERACTIVE':'yes', #Don't query the user
'PHOTO_ERRORS':'no', #Define the confidence interval using only the photometric errors
'MIN_RMS':0.05, #"Intrinsic" photo-z rms in dz /(1+z) (Change to 0.05 for templates from Benitez et al. 2004
'N_PEAKS':1,
'MERGE_PEAKS':'no',
'CONVOLVE_P':'yes',
'P_MIN':1e-2,
'SED_DIR': sed_dir,
'AB_DIR': ab_dir,
'FILTER_DIR': fil_dir,
'DELTA_M_0': 0.,
'ZP_OFFSETS':0.,
'ZC': None,
'FC':None,
"ADD_SPEC_PROB":None,
"ADD_CONTINUOUS_PROB":None,
"NMAX": None # Useful for testing
}
if pars.d['PLOTS']=='no': plots=0
if plots:
# If pylab installed show plots
plots='pylab'
try:
import matplotlib
matplotlib.use('TkAgg')
from pylab import *
# from coeplot2a import *
plot([1])
title('KILL THIS WINDOW!')
show()
ioff()
except:
try:
from biggles import *
plots='biggles'
except:
plots=0
#Define the default values of the parameters
pars.d['INPUT']=sys.argv[1] # catalog with the photometry
obs_file=pars.d['INPUT']
#root=os.path.splitext(pars.d['INPUT'])[0]
### Mike's edits ###
root=os.path.splitext(pars.d['INPUT'])[0]
### End Mike's Edits ###
pars.d['COLUMNS']=root+'.columns' # column information for the input catalog
pars.d['OUTPUT']= root+'.bpz' # output
nargs=len(sys.argv)
ipar=2
if nargs>2: #Check for parameter file and update parameters
if sys.argv[2]=='-P':
pars.fromfile(sys.argv[3])
ipar=4
# Update the parameters using command line additions
#pars.fromcommandline(sys.argv[ipar:])
#for key in pars.d:
# print key, pars.d[key]
#pause()
pars.d.update(params_cl()) # allows for flag only (no value after), e.g., -CHECK
def updateblank(var, ext):
global pars
if pars.d[var] in [None, 'yes']:
pars.d[var] = root+'.'+ext
updateblank('CHECK', 'flux_comparison')
updateblank('PROBS_LITE', 'probs')
updateblank('PROBS', 'full_probs')
updateblank('PROBS2', 'chisq')
#if pars.d['CHECK'] in [None, 'yes']:
# pars.d['CHECK'] = root+'.flux_comparison'
#This allows to change the auxiliary directories used by BPZ
if pars.d['SED_DIR']<>sed_dir:
print "Changing sed_dir to ",pars.d['SED_DIR']
sed_dir=pars.d['SED_DIR']
if sed_dir[-1]<>'/': sed_dir+='/'
if pars.d['AB_DIR']<>ab_dir:
print "Changing ab_dir to ",pars.d['AB_DIR']
ab_dir=pars.d['AB_DIR']
if ab_dir[-1]<>'/': ab_dir+='/'
if pars.d['FILTER_DIR']<>fil_dir:
print "Changing fil_dir to ",pars.d['FILTER_DIR']
fil_dir=pars.d['FILTER_DIR']
if fil_dir[-1]<>'/': fil_dir+='/'
#Better safe than sorry
if pars.d['OUTPUT']==obs_file or pars.d['PROBS']==obs_file or pars.d['PROBS2']==obs_file or pars.d['PROBS_LITE']==obs_file:
print "This would delete the input file!"
sys.exit()
if pars.d['OUTPUT']==pars.d['COLUMNS'] or pars.d['PROBS_LITE']==pars.d['COLUMNS'] or pars.d['PROBS']==pars.d['COLUMNS']:
print "This would delete the .columns file!"
sys.exit()
#Assign the intrinsin rms
if pars.d['SPECTRA']=='CWWSB.list':
print 'Setting the intrinsic rms to 0.067(1+z)'
pars.d['MIN_RMS']=0.067
pars.d['MIN_RMS']=float(pars.d['MIN_RMS'])
pars.d['MIN_MAGERR']=float(pars.d['MIN_MAGERR'])
if pars.d['INTERACTIVE']=='no': interactive=0
else: interactive=1
if pars.d['VERBOSE']=='yes':
print "Current parameters"
view_keys(pars.d)
pars.d['N_PEAKS']=int(pars.d['N_PEAKS'])
if pars.d["ADD_SPEC_PROB"]<>None:
specprob=1
specfile=pars.d["ADD_SPEC_PROB"]
spec=get_2Darray(specfile)
ns=spec.shape[1]
if ns/2<>(ns/2.):
print "Number of columns in SPEC_PROB is odd"
sys.exit()
z_spec=spec[:,:ns/2]
p_spec=spec[:,ns/2:]
# Write output file header
header="#ID "
header+=ns/2*" z_spec%i"
header+=ns/2*" p_spec%i"
header+="\n"
header=header % tuple(list(range(ns/2))+list(range(ns/2)))
specout=open(specfile.split()[0]+".p_spec","w")
specout.write(header)
else:
specprob=0
pars.d['DELTA_M_0']=float(pars.d['DELTA_M_0'])
#Some misc. initialization info useful for the .columns file
#nofilters=['M_0','OTHER','ID','Z_S','X','Y']
nofilters=['M_0','OTHER','ID','Z_S']
#Numerical codes for nondetection, etc. in the photometric catalog
unobs=-99. #Objects not observed
undet= 99. #Objects not detected
#Define the z-grid
zmin=float(pars.d['ZMIN'])
zmax=float(pars.d['ZMAX'])
if zmin > zmax : raise 'zmin < zmax !'
dz=float(pars.d['DZ'])
linear=1
if linear:
z=arange(zmin,zmax+dz,dz)
else:
if zmax<>0.:
zi=zmin
z=[]
while zi<=zmax:
z.append(zi)
zi=zi+dz*(1.+zi)
z=array(z)
else: z=array([0.])
#Now check the contents of the FILTERS,SED and A diBrectories
#Get the filters in stock
filters_db=[]
filters_db=glob.glob(fil_dir+'*.res')
for i in range(len(filters_db)):
filters_db[i]=os.path.basename(filters_db[i])
filters_db[i]=filters_db[i][:-4]
#Get the SEDs in stock
sed_db=[]
sed_db=glob.glob(sed_dir+'*.sed')
for i in range(len(sed_db)):
sed_db[i]=os.path.basename(sed_db[i])
sed_db[i]=sed_db[i][:-4]
#Get the ABflux files in stock
ab_db=[]
ab_db=glob.glob(ab_dir+'*.AB')
for i in range(len(ab_db)):
ab_db[i]=os.path.basename(ab_db[i])
ab_db[i]=ab_db[i][:-3]
#Get a list with the filter names and check whether they are in stock
col_file=pars.d['COLUMNS']
filters=get_str(col_file,0)
for cosa in nofilters:
if filters.count(cosa):filters.remove(cosa)
if pars.d['EXCLUDE']<>'none':
if type(pars.d['EXCLUDE'])==type(' '):
pars.d['EXCLUDE']=[pars.d['EXCLUDE']]
for cosa in pars.d['EXCLUDE']:
if filters.count(cosa):filters.remove(cosa)
for filter in filters:
if filter[-4:]=='.res': filter=filter[:-4]
if filter not in filters_db:
print 'filter ', filter, 'not in database at',fil_dir, ':'
if ask('Print filters in database?'):
for line in filters_db: print line
sys.exit()
#Get a list with the spectrum names and check whether they're in stock
#Look for the list in the home directory first,
#if it's not there, look in the SED directory
spectra_file=os.path.join(homedir,pars.d['SPECTRA'])
if not os.path.exists(spectra_file):
spectra_file=os.path.join(sed_dir,pars.d['SPECTRA'])
spectra=get_str(spectra_file,0)
for i in range(len(spectra)):
if spectra[i][-4:]=='.sed': spectra[i]=spectra[i][:-4]
nf=len(filters)
nt=len(spectra)
nz=len(z)
#Get the model fluxes
f_mod=zeros((nz,nt,nf))*0.
abfiles=[]
for it in range(nt):
for jf in range(nf):
if filters[jf][-4:]=='.res': filtro=filters[jf][:-4]
else: filtro=filters[jf]
model=join([spectra[it],filtro,'AB'],'.')
model_path=os.path.join(ab_dir,model)
abfiles.append(model)
#Generate new ABflux files if not present
# or if new_ab flag on
if pars.d['NEW_AB']=='yes' or model[:-3] not in ab_db:
if spectra[it] not in sed_db:
print 'SED ', spectra[it], 'not in database at',sed_dir
# for line in sed_db:
# print line
sys.exit()
#print spectra[it],filters[jf]
print ' Generating ',model,'....'
ABflux(spectra[it],filtro,madau=pars.d['MADAU'])
#z_ab=arange(0.,zmax_ab,dz_ab) #zmax_ab and dz_ab are def. in bpz_tools
# abflux=f_z_sed(spectra[it],filters[jf], z_ab,units='nu',madau=pars.d['MADAU'])
# abflux=clip(abflux,0.,1e400)
# buffer=join(['#',spectra[it],filters[jf], 'AB','\n'])
#for i in range(len(z_ab)):
# buffer=buffer+join([`z_ab[i]`,`abflux[i]`,'\n'])
#open(model_path,'w').write(buffer)
#zo=z_ab
#f_mod_0=abflux
#else:
#Read the data
zo,f_mod_0=get_data(model_path,(0,1))
#Rebin the data to the required redshift resolution
f_mod[:,it,jf]=match_resol(zo,f_mod_0,z)
#if sometrue(less(f_mod[:,it,jf],0.)):
if less(f_mod[:,it,jf],0.).any():
print 'Warning: some values of the model AB fluxes are <0'
print 'due to the interpolation '
print 'Clipping them to f>=0 values'
#To avoid rounding errors in the calculation of the likelihood
f_mod[:,it,jf]=clip(f_mod[:,it,jf],0.,1e300)
#We forbid f_mod to take values in the (0,1e-100) interval
#f_mod[:,it,jf]=where(less(f_mod[:,it,jf],1e-100)*greater(f_mod[:,it,jf],0.),0.,f_mod[:,it,jf])
#Here goes the interpolacion between the colors
ninterp=int(pars.d['INTERP'])
ntypes = pars.d['NTYPES']
if ntypes == None:
nt0 = nt
else:
nt0 = list(ntypes)
for i, nt1 in enumerate(nt0):
print i, nt1
nt0[i] = int(nt1)
if (len(nt0) <> 3) or (sum(nt0) <> nt):
print
print '%d ellipticals + %d spirals + %d ellipticals' % tuple(nt0)
print 'does not add up to %d templates' % nt
print 'USAGE: -NTYPES nell,nsp,nsb'
print 'nell = # of elliptical templates'
print 'nsp = # of spiral templates'
print 'nsb = # of starburst templates'
print 'These must add up to the number of templates in the SPECTRA list'
print 'Quitting BPZ.'
sys.exit()
if ninterp:
nti=nt+(nt-1)*ninterp
buffer=zeros((nz,nti,nf))*1.
tipos=arange(0.,float(nti),float(ninterp)+1.)
xtipos=arange(float(nti))
for iz in arange(nz):
for jf in range(nf):
buffer[iz,:,jf]=match_resol(tipos,f_mod[iz,:,jf],xtipos)
nt=nti
f_mod=buffer
#for j in range(nf):
# plot=FramedPlot()
# for i in range(nt): plot.add(Curve(z,log(f_mod[:,i,j]+1e-40)))
# plot.show()
# ask('More?')
#Load all the parameters in the columns file to a dictionary
col_pars=params()
col_pars.fromfile(col_file)
# Read which filters are in which columns
flux_cols=[]
eflux_cols=[]
cals=[]
zp_errors=[]
zp_offsets=[]
for filter in filters:
datos=col_pars.d[filter]
flux_cols.append(int(datos[0])-1)
eflux_cols.append(int(datos[1])-1)
cals.append(datos[2])
zp_errors.append(datos[3])
zp_offsets.append(datos[4])
zp_offsets=array(map(float,zp_offsets))
if pars.d['ZP_OFFSETS']:
zp_offsets+=array(map(float,pars.d['ZP_OFFSETS']))
flux_cols=tuple(flux_cols)
eflux_cols=tuple(eflux_cols)
#READ the flux and errors from obs_file
f_obs=get_2Darray(obs_file,flux_cols)
ef_obs=get_2Darray(obs_file,eflux_cols)
#Convert them to arbitrary fluxes if they are in magnitudes
if pars.d['MAG']=='yes':
seen=greater(f_obs,0.)*less(f_obs,undet)
no_seen=equal(f_obs,undet)
no_observed=equal(f_obs,unobs)
todo=seen+no_seen+no_observed
#The minimum photometric error is 0.01
#ef_obs=ef_obs+seen*equal(ef_obs,0.)*0.001
ef_obs=where(greater_equal(ef_obs,0.),clip(ef_obs,pars.d['MIN_MAGERR'],1e10),ef_obs)
if add.reduce(add.reduce(todo))<>todo.shape[0]*todo.shape[1]:
print 'Objects with unexpected magnitudes!'
print """Allowed values for magnitudes are
0<m<"""+`undet`+" m="+`undet`+"(non detection), m="+`unobs`+"(not observed)"
for i in range(len(todo)):
if not alltrue(todo[i,:]):
print i+1,f_obs[i,:],ef_obs[i,:]
sys.exit()
#Detected objects
try:
f_obs=where(seen,10.**(-.4*f_obs),f_obs)
except OverflowError:
print 'Some of the input magnitudes have values which are >700 or <-700'
print 'Purge the input photometric catalog'
print 'Minimum value',min(f_obs)
print 'Maximum value',max(f_obs)
print 'Indexes for minimum values',argmin(f_obs,0.)
print 'Indexes for maximum values',argmax(f_obs,0.)
print 'Bye.'
sys.exit()
try:
ef_obs=where(seen,(10.**(.4*ef_obs)-1.)*f_obs,ef_obs)
except OverflowError:
print 'Some of the input magnitude errors have values which are >700 or <-700'
print 'Purge the input photometric catalog'
print 'Minimum value',min(ef_obs)
print 'Maximum value',max(ef_obs)
print 'Indexes for minimum values',argmin(ef_obs,0.)
print 'Indexes for maximum values',argmax(ef_obs,0.)
print 'Bye.'
sys.exit()
#print 'ef', ef_obs[0,:nf]
#print 'f', f_obs[1,:nf]
#print 'ef', ef_obs[1,:nf]
#Looked at, but not detected objects (mag=99.)
#We take the flux equal to zero, and the error in the flux equal to the 1-sigma detection error.
#If m=99, the corresponding error magnitude column in supposed to be dm=m_1sigma, to avoid errors
#with the sign we take the absolute value of dm
f_obs=where(no_seen,0.,f_obs)
ef_obs=where(no_seen,10.**(-.4*abs(ef_obs)),ef_obs)
#Objects not looked at (mag=-99.)
f_obs=where(no_observed,0.,f_obs)
ef_obs=where(no_observed,0.,ef_obs)
#Flux codes:
# If f>0 and ef>0 : normal objects
# If f==0 and ef>0 :object not detected
# If f==0 and ef==0: object not observed
#Everything else will crash the program
#Check that the observed error fluxes are reasonable
#if sometrue(less(ef_obs,0.)): raise 'Negative input flux errors'
if less(ef_obs,0.).any(): raise 'Negative input flux errors'
f_obs=where(less(f_obs,0.),0.,f_obs) #Put non-detections to 0
ef_obs=where(less(f_obs,0.),maximum(1e-100,f_obs+ef_obs),ef_obs) # Error equivalent to 1 sigma upper limit
#if sometrue(less(f_obs,0.)) : raise 'Negative input fluxes'
seen=greater(f_obs,0.)*greater(ef_obs,0.)
no_seen=equal(f_obs,0.)*greater(ef_obs,0.)
no_observed=equal(f_obs,0.)*equal(ef_obs,0.)
todo=seen+no_seen+no_observed
if add.reduce(add.reduce(todo))<>todo.shape[0]*todo.shape[1]:
print 'Objects with unexpected fluxes/errors'
#Convert (internally) objects with zero flux and zero error(non observed)
#to objects with almost infinite (~1e108) error and still zero flux
#This will yield reasonable likelihoods (flat ones) for these objects
ef_obs=where(no_observed,1e108,ef_obs)
#Include the zero point errors
zp_errors=array(map(float,zp_errors))
zp_frac=e_mag2frac(zp_errors)
#zp_frac=10.**(.4*zp_errors)-1.
ef_obs=where(seen,sqrt(ef_obs*ef_obs+(zp_frac*f_obs)**2),ef_obs)
ef_obs=where(no_seen,sqrt(ef_obs*ef_obs+(zp_frac*(ef_obs/2.))**2),ef_obs)
#Add the zero-points offset
#The offsets are defined as m_new-m_old
zp_offsets=array(map(float,zp_offsets))
zp_offsets=where(not_equal(zp_offsets,0.),10.**(-.4*zp_offsets),1.)
f_obs=f_obs*zp_offsets
ef_obs=ef_obs*zp_offsets
#Convert fluxes to AB if needed
for i in range(f_obs.shape[1]):
if cals[i]=='Vega':
const=mag2flux(VegatoAB(0.,filters[i]))
f_obs[:,i]=f_obs[:,i]*const
ef_obs[:,i]=ef_obs[:,i]*const
elif cals[i]=='AB':continue
else:
print 'AB or Vega?. Check '+col_file+' file'
sys.exit()
#Get m_0 (if present)
if col_pars.d.has_key('M_0'):
m_0_col=int(col_pars.d['M_0'])-1
m_0=get_data(obs_file,m_0_col)
m_0+=pars.d['DELTA_M_0']
#Get the objects ID (as a string)
if col_pars.d.has_key('ID'):
# print col_pars.d['ID']
id_col=int(col_pars.d['ID'])-1
id=get_str(obs_file,id_col)
else:
id=map(str,range(1,len(f_obs[:,0])+1))
#Get spectroscopic redshifts (if present)
if col_pars.d.has_key('Z_S'):
z_s_col=int(col_pars.d['Z_S'])-1
z_s=get_data(obs_file,z_s_col)
#Get the X,Y coordinates
if col_pars.d.has_key('X'):
datos = col_pars.d['X']
if len(datos) == 1: # OTHERWISE IT'S A FILTER!
x_col=int(col_pars.d['X'])-1
x=get_data(obs_file,x_col)
if col_pars.d.has_key('Y'):
datos = col_pars.d['Y']
if len(datos) == 1: # OTHERWISE IT'S A FILTER!
y_col=int(datos)-1
y=get_data(obs_file,y_col)
#If 'check' on, initialize some variables
check=pars.d['CHECK']
# This generates a file with m,z,T and observed/expected colors
#if check=='yes': pars.d['FLUX_COMPARISON']=root+'.flux_comparison'
checkSED = check<>'no'
ng=f_obs.shape[0]
if checkSED:
# PHOTOMETRIC CALIBRATION CHECK
#r=zeros((ng,nf),float)+1.
#dm=zeros((ng,nf),float)+1.
#w=r*0.
# Defaults: r=1, dm=1, w=0
frat = ones((ng,nf), float)
dmag = ones((ng,nf), float)
fw = zeros((ng,nf), float)
#Visualize the colors of the galaxies and the templates
#When there are spectroscopic redshifts available
if interactive and col_pars.d.has_key('Z_S') and plots and checkSED and ask('Plot colors vs spectroscopic redshifts?'):
color_m=zeros((nz,nt,nf-1))*1.
if plots == 'pylab':
figure(1)
nrows=2
ncols=(nf-1)/nrows
if (nf-1)%nrows: ncols+=1
for i in range(nf-1):
##plot=FramedPlot()
# Check for overflows
fmu=f_obs[:,i+1]
fml=f_obs[:,i]
good=greater(fml,1e-100)*greater(fmu,1e-100)
zz,fmu,fml=multicompress(good,(z_s,fmu,fml))
colour=fmu/fml
colour=clip(colour,1e-5,1e5)
colour=2.5*log10(colour)
if plots == 'pylab':
subplot(nrows,ncols,i+1)
plot(zz,colour,"bo")
elif plots == 'biggles':
d=Points(zz,colour,color='blue')
plot.add(d)
for it in range(nt):
#Prevent overflows
fmu=f_mod[:,it,i+1]
fml=f_mod[:,it,i]
good=greater(fml,1e-100)
zz,fmu,fml=multicompress(good,(z,fmu,fml))
colour=fmu/fml
colour=clip(colour,1e-5,1e5)
colour=2.5*log10(colour)
if plots == 'pylab':
plot(zz,colour,"r")
elif plots == 'biggles':
d=Curve(zz,colour,color='red')
plot.add(d)
if plots == 'pylab':
xlabel(r'$z$')
ylabel('%s - %s' %(filters[i],filters[i+1]))
elif plots == 'biggles':
plot.xlabel=r'$z$'
plot.ylabel='%s - %s' %(filters[i],filters[i+1])
plot.save_as_eps('%s-%s.eps'%(filters[i],filters[i+1]))
plot.show()
if plots == 'pylab':
show()
inp = raw_input('Hit Enter to continue.')
#Get other information which will go in the output file (as strings)
if col_pars.d.has_key('OTHER'):
if col_pars.d['OTHER']<>'all':
other_cols=col_pars.d['OTHER']
if type(other_cols)==type((2,)):
other_cols=tuple(map(int,other_cols))
else:
other_cols=(int(other_cols),)
other_cols=map(lambda x: x-1,other_cols)
n_other=len(other_cols)
else:
n_other=get_2Darray(obs_file,cols='all',nrows=1).shape[1]
other_cols=range(n_other)
others=get_str(obs_file,other_cols)
if len(other_cols)>1:
other=[]
for j in range(len(others[0])):
lista=[]
for i in range(len(others)):
lista.append(others[i][j])
other.append(join(lista))
else:
other=others
if pars.d['GET_Z']=='no': get_z=0
else: get_z=1
#Prepare the output file
out_name=pars.d['OUTPUT']
if get_z:
if os.path.exists(out_name):
os.system('cp %s %s.bak' % (out_name,out_name))
print "File %s exists. Copying it to %s.bak" % (out_name,out_name)
output=open(out_name,'w')
if pars.d['PROBS_LITE']=='no': save_probs=0
else: save_probs=1
if pars.d['PROBS']=='no': save_full_probs=0
else: save_full_probs=1
if pars.d['PROBS2']=='no': save_probs2=0
else: save_probs2=1
#Include some header information
# File name and the date...
time_stamp=time.ctime(time.time())
if get_z: output.write('## File '+out_name+' '+time_stamp+'\n')
#and also the parameters used to run bpz...
if get_z:output.write("""##
##Parameters used to run BPZ:
##
""")
claves=pars.d.keys()
claves.sort()
for key in claves:
if type(pars.d[key])==type((1,)):
cosa=join(list(pars.d[key]),',')
else:
cosa=str(pars.d[key])
if get_z: output.write('##'+upper(key)+'='+cosa+'\n')
if save_full_probs:
#Shelve some info on the run
full_probs=shelve.open(pars.d['PROBS'])
full_probs['TIME']=time_stamp
full_probs['PARS']=pars.d
if save_probs:
probs=open(pars.d['PROBS_LITE'],'w')
probs.write('# ID p_bayes(z) where z=arange(%.4f,%.4f,%.4f) \n' % (zmin,zmax+dz,dz))
if save_probs2:
probs2=open(pars.d['PROBS2'],'w')
probs2.write('# id t z1 P(z1) P(z1+dz) P(z1+2*dz) ... where dz = %.4f\n' % dz)
#probs2.write('# ID\n')
#probs2.write('# t z1 P(z1) P(z1+dz) P(z1+2*dz) ... where dz = %.4f\n' % dz)
#Use a empirical prior?
tipo_prior=pars.d['PRIOR']
useprior=0
if col_pars.d.has_key('M_0'): has_mags=1
else: has_mags=0
if has_mags and tipo_prior<>'none' and tipo_prior<>'flat': useprior=1
#Add cluster 'spikes' to the prior?
cluster_prior=0.
if pars.d['ZC'] :
cluster_prior=1
if type(pars.d['ZC'])==type(""): zc=array([float(pars.d['ZC'])])
else: zc=array(map(float,pars.d['ZC']))
if type(pars.d['FC'])==type(""): fc=array([float(pars.d['FC'])])
else: fc=array(map(float,pars.d['FC']))
fcc=add.reduce(fc)
if fcc>1. :
print ftc
raise 'Too many galaxies in clusters!'
pi_c=zeros((nz,nt))*1.
#Go over the different cluster spikes
for i in range(len(zc)):
#We define the cluster within dz=0.01 limits
cluster_range=less_equal(abs(z-zc[i]),.01)*1.
#Clip values to avoid overflow
exponente=clip(-(z-zc[i])**2/2./(0.00333)**2,-700.,0.)
#Outside the cluster range g is 0
g=exp(exponente)*cluster_range
norm=add.reduce(g)
pi_c[:,0]=pi_c[:,0]+g/norm*fc[i]
#Go over the different types
print 'We only apply the cluster prior to the early type galaxies'
for i in range(1,3+2*ninterp):
pi_c[:,i]=pi_c[:,i]+pi_c[:,0]
#Output format
format='%'+`maximum(5,len(id[0]))`+'s' #ID format
format=format+pars.d['N_PEAKS']*' %.3f %.3f %.3f %.3f %.5f'+' %.3f %.3f %10.3f'
#Add header with variable names to the output file
sxhdr="""##
##Column information
##
# 1 ID"""
k=1
if pars.d['N_PEAKS']>1:
for j in range(pars.d['N_PEAKS']):
sxhdr+="""
# %i Z_B_%i
# %i Z_B_MIN_%i
# %i Z_B_MAX_%i
# %i T_B_%i
# %i ODDS_%i""" % (k+1,j+1,k+2,j+1,k+3,j+1,k+4,j+1,k+5,j+1)
k+=5
else:
sxhdr+="""
# %i Z_B
# %i Z_B_MIN
# %i Z_B_MAX
# %i T_B
# %i ODDS""" % (k+1,k+2,k+3,k+4,k+5)
k+=5
sxhdr+="""
# %i Z_ML
# %i T_ML
# %i CHI-SQUARED\n""" % (k+1,k+2,k+3)
nh=k+4
if col_pars.d.has_key('Z_S'):
sxhdr=sxhdr+'# %i Z_S\n' % nh
format=format+' %.3f'
nh+=1
if has_mags:
format=format+' %.3f'
sxhdr=sxhdr+'# %i M_0\n' % nh
nh+=1
if col_pars.d.has_key('OTHER'):
sxhdr=sxhdr+'# %i OTHER\n' % nh
format=format+' %s'
nh+=n_other
#print sxhdr
if get_z: output.write(sxhdr+'##\n')
odds_i=float(pars.d['ODDS'])
oi=inv_gauss_int(odds_i)
print odds_i,oi
#Proceed to redshift estimation
if checkSED: buffer_flux_comparison=""
if pars.d['CONVOLVE_P']=='yes':
# Will Convolve with a dz=0.03 gaussian to make probabilities smoother
# This is necessary; if not there are too many close peaks
sigma_g=0.03
x=arange(-3.*sigma_g, 3.*sigma_g + dz/10., dz) # made symmetric --DC
gaus=exp(-(x/sigma_g)**2)
if pars.d["NMAX"]<>None: ng=int(pars.d["NMAX"])
for ig in range(ng):
#Don't run BPZ on galaxies with have z_s > z_max
#if col_pars.d.has_key('Z_S'):
# if z_s[ig]<9.9 and z_s[ig]>zmax : continue
if not get_z: continue
if pars.d['COLOR']=='yes': likelihood=p_c_z_t_color(f_obs[ig,:nf],ef_obs[ig,:nf],f_mod[:nz,:nt,:nf])
else: likelihood=p_c_z_t(f_obs[ig,:nf],ef_obs[ig,:nf],f_mod[:nz,:nt,:nf])
if 0:
print f_obs[ig,:nf]
print ef_obs[ig,:nf]
iz_ml=likelihood.i_z_ml
t_ml=likelihood.i_t_ml
red_chi2=likelihood.min_chi2/float(nf-1.)
#p=likelihood.Bayes_likelihood
#likelihood.various_plots()
#print 'FULL BAYESAIN LIKELIHOOD'
p=likelihood.likelihood
if not ig:
print 'ML * prior -- NOT QUITE BAYESIAN'
#plo=FramedPlot()
#for i in range(p.shape[1]):
# plo.add(Curve(z,likelihood.likelihood[:nz,i]/sum(sum(likelihood.likelihood[:nz,:]))))
# plo.add(Curve(z,likelihood.bayes_likelihood[:nz,i]/sum(sum(likelihood.bayes_likelihood[:nz,:])),color='red'))
# #plo.add(Curve(z,p[:nz,i]/sum(sum(p[:nz,:])),color='red'))
#plo.show()
#ask('More?')
if pars.d['ONLY_TYPE']=='yes': #Use only the redshift information, no priors
p_i=zeros((nz,nt))*1.
j=searchsorted(z,z_s[ig])
#print j,nt,z_s[ig]
p_i[j,:]=1./float(nt)
else:
if useprior:
if pars.d['PRIOR']=='lensing':
p_i=prior(z,m_0[ig],tipo_prior,nt0,ninterp,x[ig],y[ig])
else:
p_i=prior(z,m_0[ig],tipo_prior,nt0,ninterp)
else:
p_i=ones((nz,nt),float)/float(nz*nt)
if cluster_prior:p_i=(1.-fcc)*p_i+pi_c
if save_full_probs: full_probs[id[ig]]=[z,p_i[:nz,:nt],p[:nz,:nt],red_chi2]
#Multiply the prior by the likelihood to find the final probability
pb=p_i[:nz,:nt]*p[:nz,:nt]
#plo=FramedPlot()
#for i in range(p.shape[1]):
# plo.add(Curve(z,p_i[:nz,i]/sum(sum(p_i[:nz,:]))))
#for i in range(p.shape[1]):
# plo.add(Curve(z,p[:nz,i]/sum(sum(p[:nz,:])),color='red'))
#plo.add(Curve(z,pb[:nz,-1]/sum(pb[:nz,-1]),color='blue'))
#plo.show()
#ask('More?')
#Convolve with a gaussian of width \sigma(1+z) to take into
#accout the intrinsic scatter in the redshift estimation 0.06*(1+z)
#(to be done)
#Estimate the bayesian quantities
p_bayes=add.reduce(pb[:nz,:nt],-1)
#print p_bayes.shape
#print argmax(p_bayes)
#print p_bayes[300:310]
#Convolve with a gaussian
if pars.d['CONVOLVE_P']=='yes' and pars.d['ONLY_TYPE']=='no':
#print 'GAUSS CONV'
p_bayes=convolve(p_bayes,gaus,1)
#print 'gaus', gaus
#print p_bayes.shape
#print argmax(p_bayes)
#print p_bayes[300:310]
# Eliminate all low level features in the prob. distribution
pmax=max(p_bayes)
p_bayes=where(greater(p_bayes,pmax*float(pars.d['P_MIN'])),p_bayes,0.)
norm=add.reduce(p_bayes)
p_bayes=p_bayes/norm
if specprob:
p_spec[ig,:]=match_resol(z,p_bayes,z_spec[ig,:])*p_spec[ig,:]
norma=add.reduce(p_spec[ig,:])
if norma==0.: norma=1.
p_spec[ig,:]/=norma
#vyjod=tuple([id[ig]]+list(z_spec[ig,:])+list(p_spec[ig,:])+[z_s[ig],
# int(float(other[ig]))])
vyjod=tuple([id[ig]]+list(z_spec[ig,:])+list(p_spec[ig,:]))
formato="%s "+5*" %.4f"
formato+=5*" %.3f"
#formato+=" %4f %i"
formato+="\n"
print formato % vyjod
specout.write(formato % vyjod)
if pars.d['N_PEAKS']>1:
# Identify maxima and minima in the final probability
g_max=less(p_bayes[2:],p_bayes[1:-1])*less(p_bayes[:-2],p_bayes[1:-1])
g_min=greater(p_bayes[2:],p_bayes[1:-1])*greater(p_bayes[:-2],p_bayes[1:-1])
g_min+=equal(p_bayes[1:-1],0.)*greater(p_bayes[2:],0.)
g_min+=equal(p_bayes[1:-1],0.)*greater(p_bayes[:-2],0.)
i_max=compress(g_max,arange(nz-2))+1
i_min=compress(g_min,arange(nz-2))+1
# Check that the first point and the last one are not minima or maxima,
# if they are, add them to the index arrays
if p_bayes[0]>p_bayes[1]:
i_max=concatenate([[0],i_max])
i_min=concatenate([[0],i_min])
if p_bayes[-1]>p_bayes[-2]:
i_max=concatenate([i_max,[nz-1]])
i_min=concatenate([i_min,[nz-1]])
if p_bayes[0]<p_bayes[1]:
i_min=concatenate([[0],i_min])
if p_bayes[-1]<p_bayes[-2]:
i_min=concatenate([i_min,[nz-1]])
p_max=take(p_bayes,i_max)
#p_min=take(p_bayes,i_min)
p_tot=[]
z_peaks=[]
t_peaks=[]
# Sort them by probability values
p_max,i_max=multisort(1./p_max,(p_max,i_max))
# For each maximum, define the minima which sandwich it
# Assign minima to each maximum
jm=searchsorted(i_min,i_max)
p_max=list(p_max)
for i in range(len(i_max)):
z_peaks.append([z[i_max[i]],z[i_min[jm[i]-1]],z[i_min[jm[i]]]])
t_peaks.append(argmax(pb[i_max[i],:nt]))
p_tot.append(sum(p_bayes[i_min[jm[i]-1]:i_min[jm[i]]]))
# print z_peaks[-1][0],f_mod[i_max[i],t_peaks[-1]-1,:nf]
if ninterp:
t_peaks=list(array(t_peaks)/(1.+ninterp))
if pars.d['MERGE_PEAKS']=='yes':
# Merge peaks which are very close 0.03(1+z)
merged=[]
for k in range(len(z_peaks)):
for j in range(len(z_peaks)):
if j>k and k not in merged and j not in merged:
if abs(z_peaks[k][0]-z_peaks[j][0])<0.06*(1.+z_peaks[j][0]):
# Modify the element which receives the accretion
z_peaks[k][1]=minimum(z_peaks[k][1],z_peaks[j][1])
z_peaks[k][2]=maximum(z_peaks[k][2],z_peaks[j][2])
p_tot[k]+=p_tot[j]
# Put the merged element in the list
merged.append(j)
#print merged
# Clean up
copia=p_tot[:]
for j in merged:
p_tot.remove(copia[j])
copia=z_peaks[:]
for j in merged:
z_peaks.remove(copia[j])