-
Notifications
You must be signed in to change notification settings - Fork 0
/
decor.py
275 lines (224 loc) · 8.35 KB
/
decor.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
import sys
import os
import numpy as np
import time
import glob
import pickle,pyfits
#from numpy import nanstd, nanmedian #For CASA 5.1
from scipy.stats import nanstd,nanmedian
import datetime as dt
import numpy.ma as ma
from numpy import matrix
import matplotlib.pyplot as plt
import selfcal_input
reload(selfcal_input)
del selfcal_input
from selfcal_input import *
def apply_acorr(dat, ants1, ants2, idx_ant, elen, tile, uvw, BW, fout):
"""
Amplitude correction for the beamformer-to-receiver cable length
differences (""decorrelation"")
"""
c_speed_of_light = 299792458. # m/s
tauw_sign = +1.
iau = np.where(ants1 == ants2)[0] # Pointers into ants where a1 == a2
uant = ants1[iau] # Unique antennas in the same order as they appear
nant = len(uant) # Number of unique antennas
antmax = ants1.max() # Max antenna numeric *name*
#
# Fill in the array of antenna indices idx_ant
#
iuant = 0
for ia in uant:
idx_ant[ia] = iuant
iuant = iuant + 1
#
# Baseline u,v,w coordinates and lengths
#
u = uvw[0,:]
v = uvw[1,:]
w = uvw[2,:]
blen = np.sqrt(u**2 + v**2)
#
# Amplitude-corrected data block
#
cdat = np.zeros_like(dat)
ndat = len(dat[0,0,:]) # Number of correlations per a pol and a chan
#
# Loop over all the baselines
#
fout.write('# a1 tl1 elen1 tau1 a2 tl2 ' \
'elen2 tau2 blen w tauw acor\n')
for idat in xrange(ndat):
a1 = ants1[idat] # First antenna 'name'
a2 = ants2[idat] # Second antenna 'name'
ia1 = idx_ant[a1] # First antenna index
ia2 = idx_ant[a2] # Second antenna index
#
# Here we calculate the cable delays for each pair of tiles,
# tau1 and tau2, from their electrical cable lengths, elen.
# Also, from the wave number (in meters), w, we get the
# geometric delay between the tiles, tauw.
# The delays are combined in tau:
#
tau1 = elen[ia1]/c_speed_of_light
tau2 = elen[ia2]/c_speed_of_light
# Let us not take this into account for a time being
tauw = w[idat]/c_speed_of_light
tau = tau2 - tau1 + tauw_sign*tauw
#
# The correction factor, acor, is the result of integration:
#
# acor = {1/BW}\int_{-BW/2}^{+BW/2} cos(2\pi f tau) df = sinc(BW tau).
#
acor = np.sinc(BW*tau)
line = '%4d %4d %9.2f %12.5e %4d %4d %9.2f %12.5e ' \
'%8.2f %8.2f %12.5e %9.6f\n' % \
(a1, tile[ia1], elen[ia1], tau1, \
a2, tile[ia2], elen[ia2], tau2, \
blen[idat], w[idat], tauw, acor)
fout.write(line)
#print 'tau2=%g, tau1=%g, tauw=%g, w=%g, tau=%g, acor=%g' % \
# (tau2, tau1, tauw, w[idat], tau, acor)
cdat[:,:,idat] = dat[:,:,idat]/acor
return cdat
def decor(msname,metafits,n_tblk,single_time): # n_tblk: Size of a read/write block: how many individual times in it
c_speed_of_light = 299792458. # m/s
#
# The sign of geometric delay term
#
tauw_sign = +1.
#tauw_sign = -1.
#
# Read the cable delays (electrical cable lengths, elen)
# from the metafits file.
#
ms_dir = msname
hl = pyfits.open(metafits)
#
# The second HDU is a fits table with 256 (nPol*nTile) rows.
# The order of the rows is a little strange, but they can be
# ordered in the same order as casa as follows:
# Find the CASA order of the tiles and elen:
#
h1 = hl[1]
tile_order = np.argsort(h1.data['Tile'])
#
# The lengths are prefixed by "EL_" which indicates that the
# lengths given (in metres) are electrical lengths.
# Cut off the 'EL_' prefixes and reorder the e-lengths:
#
tile = [til for til in h1.data['Tile'][tile_order][::2]]
tile = np.array(tile, dtype=int)
elen = [float(Len[3:]) for Len in h1.data['Length'][tile_order][::2]]
elen = np.array(elen, dtype=np.float64)
hl.close()
#raise SystemExit
sqrt_2 = np.sqrt(2.)
r2d = 180./np.pi
before = time.time()
fout = open('decor_'+msname.split('/')[-1]+'.txt', 'w')
ms.open(ms_dir, nomodify=False)
#
# Find exact number of individual times in MS
#
md = ms.metadata()
if single_time:
n_times = 1
else:
tims = md.timesforfield(0)
n_times = len(tims)
# Convert MS times from metadata to yyyymmdd hh:mm:ss
ymds = []
for tim in tims:
q = qa.quantity(tim, 's')
ymd = qa.time(q, form='ymd', prec=8) # 8 means 12:34:56.75 format
ymds.append(ymd)
cwids = md.chanwidths(0) # Hz, 64 channel widths (40000.Hz = 4 kHz)
cwid = cwids[0]
#cwid0 = 40000. # Hz, channel width (40000.Hz = 4 kHz)
IntgTime = md.exposuretime(scan=1, spwid=0, polid=0)
tintg = IntgTime['value'] # Integration time. IntgTime['unit'] = 's'
an_samp = cwid*tintg # = 20000, cplx samples in 0.5s of 40kHz channel
#an_samp = cwid0*tintg # = 20000, cplx samples in 0.5s of 40kHz channel
n_samp = int(an_samp)
ms.selectinit(datadescid=0) # Untranslatable CASA dirty swearword
if single_time:
# Select the whole MS dataset
rec = ms.getdata(['data'])
dat = np.copy(rec['data'])
ants1 = ms.getdata(['antenna1'])['antenna1']
ants2 = ms.getdata(['antenna2'])['antenna2']
tims = ms.getdata(['time'])['time']
uvw = ms.getdata(['uvw'])['uvw']
idx_ant = 999999*np.ones(512, dtype=int) # Fill with 999999 where empty
cdat = apply_acorr(dat, ants1, ants2, idx_ant, elen, tile, \
uvw, cwid, fout)
rec['data'][:,:,:] = cdat
print '%5d, time = %.2f' % (0, tims[0]) # Only one single time
#
# Put the corrected data back to the MS database
#
ms.putdata(rec)
else:
n_io = n_times//n_tblk
if n_times % n_tblk <> 0: n_io = n_io + 1
tblk = np.arange(n_io + 1)*n_tblk
tblk[-1] = n_times
for iblk in xrange(n_io):
it0 = tblk[iblk] # Block start (points at the first)
it1 = tblk[iblk+1] # Block end (points at the next after the last)
tim0 = tims[it0]
tim1 = tims[it1-1]
ms.selectinit(reset=True) # Reset the cumulative selection!!!
ms.select({'time':[tim0, tim1]})
rec = ms.getdata(['data'])
blkdat = np.copy(rec['data'])
blkants1 = ms.getdata(['antenna1'])['antenna1']
blkants2 = ms.getdata(['antenna2'])['antenna2']
blktims = ms.getdata(['time'])['time']
blkuvw = ms.getdata(['uvw'])['uvw']
blkcdat = np.zeros_like(blkdat) # Corrected data to be written to MS
#
# Start/end indices of same times into blktims[]
#
ibt = np.where(np.diff(blktims) <> 0)[0] + 1
ibt = np.hstack((0, ibt, len(blkants1)))
idx_ant = 999999*np.ones(512, dtype=int) # Fill with 999999 where empty
#raise SystemExit
#sys.exit(0)
#
# Apply amplitude corrections to the (large) piece of correlation data
# loaded to RAM. For each single time 'tim' and full set of
# baselines (ants1, ants2), a corresponding data block 'dat' is
# extracted
# from 'blkdata', corrected, and stored in 'vdat'.
#
for itim in xrange(it0,it1):
tim = tims[itim]
ibt0 = ibt[itim-it0]
ibt1 = ibt[itim-it0+1]
dat = blkdat[:,:,ibt0:ibt1]
ants1 = blkants1[ibt0:ibt1]
ants2 = blkants2[ibt0:ibt1]
uvw = blkuvw[:,ibt0:ibt1]
cdat = apply_acorr(dat, ants1, ants2, idx_ant, elen, tile, \
uvw, cwid, fout)
blkcdat[:,:,ibt0:ibt1] = cdat
tymd = qa.time(qa.quantity(tim, 's'), form='ymd', prec=8)[0]
print '%5d, time = %.2f or %s' % (itim, tim, tymd)
print 'Block %d-%d done.' % (it0,it1-1)
#
# Store the corrected correlations in the measurement set
#
rec['data'][:,:,:] = blkcdat
#
# Put the corrected data back to the MS database
#
ms.putdata(rec)
ms.close()
fout.close()
return
if os.path.isfile('.decor_applied.check')==False:
decor(msname,metafits,10,False)
os.system("touch .decor_applied.check")