/
postprocess_maps.py
262 lines (243 loc) · 9.22 KB
/
postprocess_maps.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
from __future__ import print_function, division
import argparse
parser = argparse.ArgumentParser()
parser.add_argument("idir")
parser.add_argument("odir")
parser.add_argument("-i", "--iteration", type=str, default="individual")
parser.add_argument("-a", "--allow-nonstandard", action="store_true")
parser.add_argument("-c", "--cont", action="store_true")
parser.add_argument("-d", "--dry", action="store_true")
parser.add_argument("-O", "--output", type=str, default="map,ivar,sens,xlink,hits,totmap,totsens,totxlink,tothits")
parser.add_argument("--exclude", type=str, default=None)
parser.add_argument("--only", type=str, default=None)
parser.add_argument("--repixwin", type=str, default=None)
parser.add_argument("--suffix", type=str, default="")
args = parser.parse_args()
import numpy as np, glob, re, os, shutil, sys
from enlib import enmap, utils, retile, bunch, mpi
comm = mpi.COMM_WORLD
outputs = set(args.output.split(","))
verbose = True
dtype = np.float32
def get_ref(ivar):
v = ivar[ivar>0]
if len(v > 1000): return np.median(v[::1+utils.floor(len(v)/1000)])
elif len(v) > 1: return np.max(v)
else: return 1
# Optional pixel window correction
def apply_fourier(map, ivar, op, tol=0.01):
# To make fourier operations safe, we want to whiten areas with
# very high hitcount variance, which will be areas where the hitcount
# is low.
ref = get_ref(ivar)*tol
low = (ivar>0)&(ivar<ref)
whiten = (ivar[low]/ref)**0.5
# Apply edge whitening
map = map.copy()
map[:,low] *= whiten
# Actually apply our fourier operation
fmap = enmap.fft(map)
del map
fmap = op(fmap)
map = enmap.ifft(fmap).real
del fmap
# Undo edge whitening
map[:,low] /= whiten
# Remask
map[:,ivar<=0] = 0
return map
if args.repixwin:
iwin, owin = utils.parse_ints(args.repixwin)
def mapfix(map, ivar):
return apply_fourier(map, ivar,
lambda fmap: enmap.apply_window(enmap.unapply_window(fmap, order=iwin, nofft=True), order=owin, nofft=True)
)
else:
def mapfix(map, ivar): return map
# Look for map files in the input directory
datasets = {}
for fname in utils.glob(args.idir + "/*map????.fits"):
fname = os.path.basename(fname)
if args.exclude and re.search(args.exclude, fname): continue
if args.only and not re.search(args.only, fname): continue
m = re.search(r"^(.*)_(\d+)way_(\d+)_sky_map(\d\d\d\d).fits", fname)
if not m:
if comm.rank == 0: print("Skipping unrecognized map name: " + fname)
continue
prefix = m.group(1)
nway = int(m.group(2))
sub = int(m.group(3))
it = int(m.group(4))
if prefix not in datasets:
datasets[prefix] = [bunch.Bunch(it=0, maxit=0, its=set(), name="") for i in range(nway)]
datasets[prefix][sub].maxit = max(datasets[prefix][sub].maxit, it)
datasets[prefix][sub].its.add(it)
datasets[prefix][sub].name = "_".join(fname.split("_")[:-2])
nchar = max([len(key) for key in datasets])
# Find the iteration count to use for each dataset
for key in sorted(datasets.keys()):
d = datasets[key]
minit = min([sub.it for sub in d])
if args.iteration == "individual":
for i in range(len(d)): d[i].it = d[i].maxit
elif args.iteration == "min":
for i in range(len(d)): d[i].it = minit
else:
try:
for i in range(len(d)):
it = int(args.iteration)
if it in d[i].its:
d[i].it = int(args.iteration)
except ValueError:
raise ValueError(args.iteration)
# Do we have any zeros at this point? If so, warn and skip the dataset
nzero = sum([sub.it == 0 for sub in d])
if nzero > 0:
if comm.rank == 0:
desc = " %4d"*len(d) % tuple([sub.it for sub in d])
print("%-*s with %s has only %d/%d splits. Skipping" % (nchar, key, desc, len(d)-nzero, len(d)))
del datasets[key]
elif comm.rank == 0:
print("%-*s using" % (nchar, key) + " %4d"*len(d) % tuple([sub.it for sub in d]))
# Check if we follow the standard format
onames = {}
for key in list(datasets.keys()):
m = re.match(r"s\d\d_\w+_pa\d_f\d\d\d_(no)?hwp_(day|night|daynight)\b.*", key)
if not m:
if args.allow_nonstandard: onames[key] = key
else: del datasets[key]
else:
toks = key.split("_")
if toks[1] == "cmb": toks[1] = "advact"
onames[key] = "_".join(toks[:6])
def read_map(ifile, slice=None):
if not os.path.isdir(ifile):
map = enmap.read_map(ifile)
if slice: map = eval("map"+slice)
else:
map = retile.read_monolithic(ifile, slice=slice, verbose=False)
map = map.astype(dtype, copy=False)
return map
def copy_mono(ifile, ofile, slice=None):
if args.cont and os.path.exists(ofile): return
if verbose: print("%3d copy_mono %s" % (comm.rank, ofile))
if args.dry: return
tfile = ofile + ".tmp"
map = read_map(ifile, slice=slice)
enmap.write_map(tfile, map)
shutil.move(tfile, ofile)
def add_mono(ifiles, ofile, slice=None, factors=None):
if args.cont and os.path.exists(ofile): return
if verbose: print("%3d add_mono %s" % (comm.rank, ofile))
if args.dry: return
tfile = ofile + ".tmp"
if factors is None:
omap = read_map(ifiles[0], slice=slice)
for ifile in ifiles[1:]:
omap += read_map(ifile, slice=slice)
else:
omap = read_map(ifiles[0], slice=slice)*factors[0]
for i in range(1, len(ifiles)):
omap += read_map(ifiles[i], slice=slice)*factors[i]
enmap.write_map(tfile, omap)
shutil.move(tfile, ofile)
def coadd_mono(imapfiles, idivfiles, omapfile, odivfile=None, op=lambda map,ivar:map):
if args.cont and os.path.exists(omapfile) and (odivfile is None or os.path.exists(odivfile)): return
if verbose: print("%3d coadd_mono %s" % (comm.rank, omapfile))
if args.dry: return
omap, odiv = 0, 0
tmapfile = omapfile + ".tmp"
if odivfile: tdivfile = odivfile + ".tmp"
if len(imapfiles) != 1:
for imapfile, idivfile in zip(imapfiles, idivfiles):
imap = read_map(imapfile)
idiv = read_map(idivfile, slice=".preflat[0]")
omap += imap*idiv
odiv += idiv
del imap, idiv
mask = odiv > 0
omap[...,mask] /= odiv[mask]
else:
omap = read_map(imapfiles[0])
odiv = read_map(idivfiles[0], slice=".preflat[0]")
omap = op(omap, odiv)
enmap.write_map(tmapfile, omap)
if odivfile: enmap.write_map(tdivfile, odiv)
shutil.move(tmapfile, omapfile)
if odivfile: shutil.move(tdivfile, odivfile)
def copy_plain(ifile, ofile):
if args.cont and os.path.exists(ofile): return
if verbose: print("%3d copy_plain %s" % (comm.rank, ofile))
if args.dry: return
tfile = ofile + ".tmp"
shutil.copyfile(ifile, tfile)
shutil.move(tfile, ofile)
def cat_files(ifiles, ofile):
if args.cont and os.path.exists(ofile): return
if verbose: print("%3d cat %s" % (comm.rank, ofile))
if args.dry: return
with open(ofile, "w") as ofh:
for ifile in ifiles:
with open(ifile, "r") as ifh:
shutil.copyfileobj(ifh, ofh)
def link(ifile, ofile):
if args.cont and os.path.exists(ofile): return
if verbose: print("%d link %s" % (comm.rank, ofile))
os.symlink(ifile, ofile)
queue = []
def schedule(func, *fargs, **kwargs):
queue.append([func, fargs, kwargs])
utils.mkdir(args.odir)
# We can now process each dataset
all_files = [os.path.basename(p) for p in glob.glob(args.idir + "/*")]
for iname in sorted(datasets.keys()):
d = datasets[iname]
obase = args.odir + "/" + onames[iname]
# Per split stuff
for si, sub in enumerate(d):
ipre = args.idir + "/" + sub.name + "_"
opre = obase + "_%dway_set%d_" % (len(d),si) + args.suffix
if "map" in outputs:
schedule(coadd_mono, [ipre + "sky_map%04d.fits" % sub.it], [ipre + "sky_div.fits"], opre + "map.fits", opre + "ivar.fits", op=mapfix)
if "ivar" in outputs and "map" not in outputs:
schedule(copy_mono, ipre + "sky_div.fits", opre + "ivar.fits", slice=".preflat[0]")
if "div" in outputs:
schedule(copy_mono, ipre + "sky_div.fits", opre + "div.fits")
if "hits" in outputs:
schedule(copy_mono, ipre + "sky_hits.fits", opre + "hits.fits")
if "xlink" in outputs:
schedule(copy_mono, ipre + "sky_crosslink.fits", opre + "xlink.fits")
if "sens" in outputs:
schedule(copy_plain, ipre + "noise.txt", opre + "sens.txt")
# Coadds
opre = obase + "_%dway_coadd_" % len(d) + args.suffix
if "totmap" in outputs:
imaps = [args.idir + "/" + sub.name + "_sky_map%04d.fits" % sub.it for sub in d]
idivs = [args.idir + "/" + sub.name + "_sky_div.fits" for sub in d]
schedule(coadd_mono, imaps, idivs, opre + "map.fits", opre + "ivar.fits", op=mapfix)
if "tothits" in outputs:
imaps = [args.idir + "/" + sub.name + "_sky_hits.fits" for sub in d]
schedule(add_mono, imaps, opre + "hits.fits")
if "totxlink" in outputs:
imaps = [args.idir + "/" + sub.name + "_sky_crosslink.fits" for sub in d]
schedule(add_mono, imaps, opre + "xlink.fits")
if "toticov" in outputs:
imaps = [args.idir + "/" + sub.name + "_sky_icov.fits" for sub in d]
schedule(add_mono, imaps, opre + "icov.fits", slice=".preflat[0]")
schedule(copy_plain, args.idir + "/" + d[0].name + "_sky_icov_pix.txt", opre + "icov_pix.txt")
if "totsens" in outputs:
ifiles = [args.idir + "/" + sub.name + "_noise.txt" for sub in d]
schedule(cat_files, ifiles, opre + "sens.txt")
# Process the scheduled items
np.random.seed(0)
inds = np.random.permutation(len(queue))
for i in range(comm.rank, len(queue), comm.size):
ind = inds[i]
func, fargs, kwargs = queue[ind]
try:
func(*fargs, **kwargs)
except Exception as e:
sys.stderr.write("Exception in %s\n" % func.__name__)
sys.stderr.write(str(fargs) + "\n")
sys.stderr.write(str(kwargs) + "\n")
raise