def main(args):
# Parse the command line
config = parseConfig(args)
filenames = config['args']
# Inspect the files to try and figure out what is what
metadataList = []
for filename in filenames:
uvw = LedaFits(verbose=False)
metadataList.append( (filename, uvw.inspectFile(filename)) )
# Group the files by start time and save the filenames and frequency ranges
groups = []
for filename,metadata in metadataList:
tStart = metadata['tstart']
chanBW = metadata['chanbw']
freqStart = metadata['reffreq'] + (1 - metadata['refpixel'])*chanBW
freqStop = metadata['reffreq'] + (metadata['nchan'] - metadata['refpixel'])*chanBW
## See if this file represents the start of a new group or not
new = True
for group in groups:
if tStart == group[0]:
new = False
group[1].append(filename)
group[2].append((freqStart,freqStop,chanBW))
break
## A new group has been found
if new:
group = [tStart, [filename,], [(freqStart,freqStop,chanBW),]]
groups.append(group)
# Report
print "Got %i files with groupings:" % len(filenames)
validity = []
for i,group in enumerate(groups):
## Sort the group by frequency
freqs = []
for start,stop,cbw in group[2]:
freqs.append(start)
freqOrder = [j[0] for j in sorted(enumerate(freqs), key=lambda x:x[1])]
group[1] = [group[1][j] for j in freqOrder]
group[2] = [group[2][j] for j in freqOrder]
## Report and validate
print " Group #%i" % (i+1,)
print " -> start time %s (%.2f)" % (datetime.utcfromtimestamp(group[0]), group[0])
valid = True
for j,(name,(start,stop,chanBW)) in enumerate(zip(group[1], group[2])):
### Check for frequency continuity
try:
freqDiff = start - oldStop
except NameError:
freqDiff = chanBW
oldStop = stop
if freqDiff != chanBW:
valid = False
### Report on this file
print " %i: %s from %.2f to %.2f MHz" % (j+1, os.path.basename(name), start/1e6, stop/1e6)
validity.append(valid)
print " -> valid set? %s" % valid
## Reset the validity between groups
del oldStop
print " "
# Combine
for i,(valid,group) in enumerate(zip(validity,groups)):
print "Combining group #%i..." % (i+1,)
## Jump over invalid groups
if not valid:
print " -> invalid, skipping"
continue
## Read in the files
uvws = []
for filename in group[1]:
uvws.append( LedaFits(filename, verbose=False) )
## Build the output name
obsDate = datetime.strptime(uvws[0].date_obs, "%Y-%m-%dT%H:%M:%S")
if len(group[1]) > 1:
outname = "%s_%s_%s_comb%i.FITS_" % (uvws[0].instrument, uvws[0].telescope, obsDate.strftime("%Y%m%d%H%M%S"), len(uvws))
else:
obsFreq = int((uvws[0].formatFreqs()).mean() / 1e6)
outname = "%s_%s_%s_%iMHz.FITS_" % (uvws[0].instrument, uvws[0].telescope, obsDate.strftime("%Y%m%d%H%M%S"), obsFreq)
print " -> group file basename will be '%s*'" % outname
## Make a note of lowest frequency value
freq_min = numpy.min(uvws[0].formatFreqs())
## Concatenate together the various FLUX sets
if len(uvws) > 1:
timeBL = uvws[0].d_uv_data["FLUX"].shape[0]
freqPolComp = uvws[0].d_uv_data["FLUX"].shape[1]
new_uv_data = numpy.zeros((timeBL, freqPolComp*len(group[1])), dtype=uvws[0].d_uv_data["FLUX"].dtype)
for i,uvw in enumerate(uvws):
new_uv_data[:,i*freqPolComp:(i+1)*freqPolComp] = 1.0*uvw.d_uv_data["FLUX"]
uvws[0].d_uv_data["FLUX"] = new_uv_data
## Overwrite frequency axis keywords so that we can export UV_DATA table correctly
uvws[0].h_common["REF_FREQ"] = freq_min
uvws[0].h_common["REF_PIXL"] = 1
uvws[0].h_common["NO_CHAN"] *= len(group[1])
uvws[0].h_params["NCHAN"] = uvws[0].h_common["NO_CHAN"]
uvws[0].d_frequency["TOTAL_BANDWIDTH"] *= len(group[1])
## Remove the other LedaFits instances since we only need the first one now
while len(uvws) > 1:
del uvws[-1]
## Add in the UVW coordinates
uvws[0].generateUVW(src='ZEN', use_stored=False, update_src=True)
if config['applyPhasing']:
## Apply the cable delays
uvws[0].apply_cable_delays()
## Phase to zenith
uvws[0].phase_to_src(src='ZEN')
else:
## Update the outname to reflect the fact that no phasing as been applied
outname = "%sNoPhasing_" % outname
## Save
if config['fullRes']:
### Extract all possible baselines
bls = getAllBaselines(uvws[0])
uvws[0].select_baselines(bls)
### Verify
uvws[0].verify()
### Save as FITS IDI
uvws[0].exportFitsidi(outname+'1')
### Cleanup the associated XML file
try:
xmlname = outname+'1.xml'
os.unlink(xmlname)
except OSError:
pass
if config['totalPower']:
### Extract the total power at full resolution
bls = getTotalPowerBaselines(uvws[0])
uvws[0].select_baselines(bls)
### Verify
uvws[0].verify()
### Save as FITS IDI
uvws[0].exportFitsidi(outname+'TP')
### Cleanup the associated XML file
try:
xmlname = outname+'TP.xml'
os.unlink(xmlname)
except OSError:
pass
if config['switching']:
### Extract the switching baselines at full resolution
bls = getSwitchingBaselines(uvws[0])
uvws[0].select_baselines(bls)
### Verify
uvws[0].verify()
### Save as FITS IDI
uvws[0].exportFitsidi(outname+'SW')
### Cleanup the associated XML file
try:
xmlname = outname+'SW.xml'
os.unlink(xmlname)
except OSError:
pass
if config['average']:
### Extract the static baselines
bls = getStaticBaselines(uvws[0])
uvws[0].select_baselines(bls)
### Decimate within a try...expect block to deal with bad decimation parameters
try:
uvws[0].average_time_frequency(config['tDecim'], config['sDecim'], mode='nearest')
except ValueError, e:
print "ERROR: %s, skipping" % str(e)
continue
### Verify
uvws[0].verify()
### Save as FITS IDI
uvws[0].exportFitsidi(outname+'AV')
### Cleanup the associated XML file
try:
xmlname = outname+'AV.xml'
os.unlink(xmlname)
except OSError:
pass
## Cleanup
del uvws[0]
gc.collect()
def main(args):
filenames = args
# Inspect the files to try and figure out what is what
metadataList = []
for filename in filenames:
uvw = LedaFits()
metadataList.append( (filename, uvw.inspectFile(filename)) )
# Group the files by start time and save the filenames and frequency ranges
groups = []
for filename,metadata in metadataList:
tStart = metadata['tstart']
chanBW = metadata['chanbw']
freqStart = metadata['reffreq'] + (1 - metadata['refpixel'])*chanBW
freqStop = metadata['reffreq'] + (metadata['nchan'] - metadata['refpixel'])*chanBW
## See if this file represents the start of a new group or not
new = True
for group in groups:
if tStart == group[0]:
new = False
group[1].append(filename)
group[2].append((freqStart,freqStop,chanBW))
break
## A new group has been found
if new:
group = [tStart, [filename,], [(freqStart,freqStop,chanBW),]]
groups.append(group)
# Report
print "Got %i files with groupings:" % len(filenames)
validity = []
for i,group in enumerate(groups):
## Sort the group by frequency
freqs = []
for start,stop,cbw in group[2]:
freqs.append(start)
freqOrder = [j[0] for j in sorted(enumerate(freqs), key=lambda x:x[1])]
group[1] = [group[1][j] for j in freqOrder]
group[2] = [group[2][j] for j in freqOrder]
## Report and validate
print " Group #%i" % (i+1,)
print " -> start time %s (%.2f)" % (datetime.utcfromtimestamp(group[0]), group[0])
valid = True
for j,(name,(start,stop,chanBW)) in enumerate(zip(group[1], group[2])):
### Check for frequency continuity
try:
freqDiff = start - oldStop
except NameError:
freqDiff = chanBW
oldStop = stop
if freqDiff != chanBW:
valid = False
### Report on this file
print " %i: %s from %.2f to %.2f MHz" % (j+1, os.path.basename(name), start/1e6, stop/1e6)
validity.append(valid)
print " -> valid set? %s" % valid
## Reset the validity between groups
del oldStop
print " "
# Combine
for i,(valid,group) in enumerate(zip(validity,groups)):
print "Combining group #%i..." % (i+1,)
## Jump over invalid groups
if not valid:
print " -> invalid, skipping"
continue
## Read in the files
uvws = []
for filename in group[1]:
uvws.append( LedaFits(filename) )
## Make a note of lowest frequency value
freq_min = numpy.min(uvws[0].formatFreqs())
## Concatenate together the various FLUX sets
if len(uvws) > 1:
timeBL = uvws[0].d_uv_data["FLUX"].shape[0]
freqPolComp = uvws[0].d_uv_data["FLUX"].shape[1]
new_uv_data = numpy.zeros((timeBL, freqPolComp*len(group[1])), dtype=uvws[0].d_uv_data["FLUX"].dtype)
for i,uvw in enumerate(uvws):
new_uv_data[:,i*freqPolComp:(i+1)*freqPolComp] = 1.0*uvw.d_uv_data["FLUX"]
uvws[0].d_uv_data["FLUX"] = new_uv_data
## Overwrite frequency axis keywords so that we can export UV_DATA table correctly
uvws[0].h_common["REF_FREQ"] = freq_min
uvws[0].h_common["REF_PIXL"] = 1
uvws[0].h_common["NO_CHAN"] *= len(group[1])
uvws[0].h_params["NCHAN"] = uvws[0].h_common["NO_CHAN"]
uvws[0].d_frequency["TOTAL_BANDWIDTH"] *= len(group[1])
## Remove the other LedaFits instances since we only need the first one now
while len(uvws) > 1:
del uvws[-1]
## Load in the current antenna mapping to associate antenna IDs with stand names
ant_ids = uvws[0].d_array_geometry['NOSTA']
ant_nms = [int(_annameRE.match(nm).group('id')) for nm in uvws[0].d_array_geometry['ANNAME']]
## Figure out which antenna ID each switching outrigger is
### Stand numbers for the switching outriggers
outriggers = [35, 257, 259]
ant_ids_outriggers = []
for outrigger in outriggers:
ant_ids_outriggers.append( ant_ids[ant_nms.index(outrigger)] )
## Extract
import pylab
pols = uvws[0].stokes_axis
freqs = uvws[0].formatFreqs()
for stand,ant_id in zip(outriggers, ant_ids_outriggers):
print " -> Extracting data for Stand #%i..." % stand
timestamps, data = uvws[0].extractTotalPower(ant_id, timestamps=True)
tRel = timestamps - timestamps[0]
tRel *= 24*3600 # Timestamps in seconds
data = data.real # Autocorrelations should be real
### Plot mean power over time
pylab.figure()
for i in xrange(data.shape[0]):
pylab.plot(tRel, data[i,:,:].mean(axis=-1), linestyle='-', marker='o', label='%i-%s' % (stand, pols[i]))
pylab.xlabel('Time [s]')
pylab.ylabel('Mean PSD [arb.]')
pylab.legend()
pylab.draw()
### Plot the spectra
pylab.figure()
for i in xrange(data.shape[0]):
pylab.subplot(2, 2, i+1)
for j in xrange(data.shape[1]):
pylab.plot(freqs/1e6, data[i,j,:])
pylab.title("%i-%s" % (stand, pols[i]))
pylab.xlabel('Frequency [MHz]')
pylab.ylabel('PSD [arb.]')
pylab.draw()
pylab.show()
