def split_frames(cubefile, range_pairs, target_dir):
dirname, filename = os.path.split(cubefile)
filebase = filename.rsplit('.', 1)[0]
if len(filebase) > 8:
warn("IRAF doesn't like long filenames. "
"Consider shortening the cube filename ({0})".format(filebase))
outfiles = []
for fromidx, toidx in range_pairs:
for i in range(fromidx, toidx+1):
infile = cubefile + "[*,*,{0}]".format(i)
outfile = '{0}/frame_{1:04}.fit'.format(target_dir, i)
debug("imcopy", infile, outfile)
iraf.imcopy( # easier to use imcopy and preserve headers than to use pyfits I think
input=infile,
output=outfile
)
outfiles.append(outfile)
# f = open(inlst, 'w')
# f.writelines(infiles)
# f.write('\n')
# f.close()
# outfile = '{0}/{1}_{2}-{3}.fit'.format(target_dir, filebase, fromidx, toidx)
# debug("imcombine input={input} output={output} combine=sum reject=none".format(
# input="@{0}".format(inlst), #','.join(infiles),
# output=outfile,
# ))
# outfiles.append(outfile)
# iraf.imcombine(
# input=','.join(infiles),
# output=outfile,
# combine="sum",
# reject="none",
# # project='no', # IRAF wants bools specified / default is nonboolean?
# # mclip='no',
# )
return outfiles
def daofind_brightest(filename, fwhmpsf=2.5, threshold=20.0):
debug("finding brightest in", filename)
data = pyfits.getdata(filename)
sigma = np.std(data) * 0.75 # background stddev... approximated because we don't know where the star IS yet
_dao_setup(fwhmpsf, threshold, sigma)
tmp_target_dir = tempfile.mkdtemp()
outfile = os.path.join(tmp_target_dir, 'daofind.coo')
iraf.noao.digiphot.apphot.daofind.run(
image=filename,
output=outfile,
interactive=False,
verify=False,
)
found_stars = parse_coo_file(outfile)
if len(found_stars) == 0:
warn("HAX: halving fwhmpsf to try and get a detection")
# FIXME: kludge to get AAS 2013 data
_dao_setup(fwhmpsf, threshold / 2.0, sigma)
outfile = os.path.join(tmp_target_dir, 'daofind.coo.2')
iraf.noao.digiphot.apphot.daofind.run(
image=filename,
output=outfile,
interactive=False,
verify=False,
)
found_stars = parse_coo_file(outfile)
debug(found_stars)
found_stars.sort(order=['MAG'])
brightest = found_stars[0]
debug("brightest found @", brightest['XCENTER'], ',', brightest['YCENTER'], 'with mag', brightest['MAG'])
shutil.rmtree(tmp_target_dir)
return brightest
def aoavgcube(cubefile, outfile, newexposure, fromidx=1, toidx=None):
"""
Combine (without rejection) many short exposures in a data cube
to approximate a "seeing disk" from a longer integration time.
cubefile - input data cube path
outfile - output data cube path
fromidx - 1-based index of start of combining range (default: 1)
toidx - 1-based index of end of range (default: None, end of cube)
"""
newexposure = float(newexposure)
ffile = pyfits.open(cubefile)
header = ffile[0].header
data = ffile[0].data
if 'EXPOSURE' in header.keys():
oldexposure = float(header['EXPOSURE'])
elif 'EXPTIME' in header.keys():
oldexposure = float(header['EXPTIME'])
else:
raise Exception("No exposure time value found in the datacube header!")
assert newexposure > oldexposure, ("Can't get a shorter exposure time by combining frames! "
"oldexposure {0} > newexposure {1}".format(oldexposure, newexposure))
frames_per_combined = int(newexposure / oldexposure)
epsilon = (newexposure / oldexposure) - int(newexposure / oldexposure)
if epsilon > 0.01:
warn("New exposure is not an integer multiple of old exposure, rounding to", frames_per_combined)
assert frames_per_combined > 1, "Only one old frame per combined frame; this is probably not what you want"
if toidx <= fromidx:
# use entire cube
toidx = data.shape[0]
else:
assert toidx <= data.shape[0], "toidx ({0}) > number of frames ({1})".format(toidx, data.shape[0])
assert data.shape[0] > 1, "Only one frame found! Is this a data cube?"
total_frames = toidx - fromidx
debug("toidx=",toidx,"fromidx=",fromidx)
if total_frames % frames_per_combined != 0:
warn("Total frames in range (toidx - fromidx =", toidx - fromidx, ") is not an integer multiple "
"of the number of frames per combined exposure (", frames_per_combined,
") so", total_frames % frames_per_combined, "frames from the end of "
"the range will be left off.")
toidx = toidx - (total_frames % frames_per_combined)
total_frames = toidx - fromidx
total_combined_frames = int(total_frames / frames_per_combined)
info("Output data cube will have", total_combined_frames, "total frames of", newexposure, "sec exposure")
info("Processing input data cube frames", fromidx, "to", toidx)
target_dir = tempfile.mkdtemp()
info("Created working directory {0} for intermediate data".format(target_dir))
try:
range_pairs = list((
fromidx + n * frames_per_combined,
fromidx + (n+1) * frames_per_combined - 1
) for n in range(0, total_combined_frames))
frame_paths = combine_cube_frames(cubefile, range_pairs, target_dir)
# frames_to_cube(frame_paths, outfile)
finally:
# Don't leave clutter if the task fails
# shutil.rmtree(target_dir)
info("Removed working directory {0}".format(target_dir))