#!/usr/bin/env """ split, clean, and self-cal continuum and line data NOTE: this is intended to be an interactive, iterative process so this is more a log that should be run by cutting and pasting into casa rather than as an executable script search "CHANGEME" for variables to be changed 10/9/15 MCA """ # ======================== Setup =========================== # LupusIII_91 16:09:26.980 -38:36:27.601 # Class II, K7 field = 17 # CHANGEME file_ms = '../science_calibrated.ms' contspw = '2,3,4' contspw_w = [128,3840,1920] robust = 0.5 imsize = [640,640] cell = '0.03arcsec' imagermode = 'csclean' refant = 'DA52' # CHANGEME xc = 325 # CHANGEME yc = 331 # CHANGEME in_a = 80 out_a = 120 aper = 0.5 # CHANGEME boxwidth = 300. box = rg.box([xc-boxwidth,yc-boxwidth],[xc+boxwidth,yc+boxwidth]) # ======================= Split Off Continuum ======================== # split off field from full ms split(vis = file_ms, outputvis = 'f'+str(field)+'.vis', field = field, datacolumn = 'data') # split off continuum (take the large bw spw and average split(vis = 'f'+str(field)+'.vis', outputvis = 'f'+str(field)+'_cont.vis', spw = contspw, width = contspw_w, datacolumn = 'data') # plot uv-distance vs. amplitude plotms(vis='f'+str(field)+'_cont.vis', xaxis='uvdist',yaxis='amp', coloraxis='spw') # plotfile='f'+str(field)+'_ampuv_orig.png' # showgui=False, # highres=True, # overwrite=True) # source is unresolved # find antenna close to center of configuration # check pipeline log that this ant is OK plotants(vis='f'+str(field)+'_cont.vis') #, figfile='f'+str(field)+'_ants.png') # ================== Clean continuum before selfcal ================== # light clean (100 iterations) around BOTH sources # os.system('rm -rf f'+str(field)+'_cont_b4sc*') clean(vis = 'f'+str(field)+'_cont.vis', imagename = 'f'+str(field)+'_cont_b4sc', mode = 'mfs', psfmode = 'clark', niter = 100, threshold = '0.0mJy', interactive = True, mask = '', cell = cell, imsize = imsize, weighting = 'briggs', robust = robust, imagermode = imagermode) im_max = imstat(imagename = 'f'+str(field)+'_cont_b4sc.image')['max'][0] im_rms = imstat(imagename = 'f'+str(field)+'_cont_b4sc.image', region='annulus[['+str(xc)+'pix,'+str(yc)+'pix],['+str(in_a)+'pix,'+str(out_a)+'pix]]')['rms'][0] print 'Peak = {0:.2f} mJy, rms = {1:.2f} mJy, S/N = {2:.1f}'.format(1000*im_max, 1000*im_rms, im_max/im_rms) # Peak = 3.56 mJy, rms = 0.35 mJy, S/N = 10.2 ### STOP HERE. All self-calibraiton horrible, source too faint. # ======================== Self-Calibrate 1 ================== # first combine all the data by time (solint = inf) # i.e., phase self-cal over entire integration time gaincal(vis = 'f'+str(field)+'_cont.vis', caltable = 'f'+str(field)+'_cont_pcal1', refant = refant, solint = 'inf', combine = 'spw', gaintype = 'T', spw = '', calmode = 'p', minblperant = 4, minsnr = 3) # 30 of 40 solutions flagged due to SNR < 3 in spw=0 at 2015/06/14/04:33:28.0 # 30 of 40 solutions flagged due to SNR < 3 in spw=0 at 2015/06/14/05:07:10.7 # plot phase for each antenna plotcal(caltable = 'f'+str(field)+'_cont_pcal1', xaxis = 'time', yaxis = 'phase', spw = '', iteration = 'antenna', subplot = 421, plotrange = [0,0,-200,200]) # narrow yrange phases close to zero # no variation between integration times (expected since solin=inf) # note DV02 has no data (100% flagged in pipeline calibration) # apply calibration to data applycal(vis = 'f'+str(field)+'_cont.vis', spw = '', gaintable = ['f'+str(field)+'_cont_pcal1'], spwmap = [0,0,0], calwt = T, flagbackup = F) # clean self-calibrated data clean(vis = 'f'+str(field)+'_cont.vis', imagename = 'f'+str(field)+'_cont_pcal1_clean', mode = 'mfs', psfmode = 'clark', niter = 100, threshold = '0.0mJy', interactive = False, mask = 'f'+str(field)+'_cont_b4sc.mask', cell = cell, imsize = imsize, weighting = 'briggs', robust = robust, imagermode = imagermode) im_max = imstat(imagename = 'f'+str(field)+'_cont_pcal1_clean.image')['max'][0] im_rms = imstat(imagename = 'f'+str(field)+'_cont_pcal1_clean.image', region='annulus[['+str(xc)+'pix,'+str(yc)+'pix],['+str(in_a)+'pix,'+str(out_a)+'pix]]')['rms'][0] print 'Peak = {0:.2f} mJy, rms = {1:.2f} mJy, S/N = {2:.1f}'.format(1000*im_max, 1000*im_rms, im_max/im_rms) # Peak = 12.63 mJy, rms = 2.16 mJy, S/N = 5.8 (significantly larger noise, due to issues with gaincal?) # inspect images imview(raster=[{'file':'f'+str(field)+'_cont_b4sc.image'}, {'file':'f'+str(field)+'_cont_pcal1_clean.image'}]) # calibration looks absolutely horrible # ======================== Self-Calibrate 2 ================== # decrease phase self-cal solution interval to a few times integration time # int = 6s (from X125.log) gaincal(vis = 'f'+str(field)+'_cont.vis', caltable = 'f'+str(field)+'_cont_pcal2', refant = refant, solint = '20s', # CHANGEME combine = 'spw', gaintype = 'T', spw = '', calmode = 'p', minblperant = 4, minsnr = 3) # 7 of 10 solutions flagged due to SNR < 3 in spw=0 at 2015/06/14/04:33:23.0 # 6 of 10 solutions flagged due to SNR < 3 in spw=0 at 2015/06/14/05:07:15.3 plotcal(caltable = 'f'+str(field)+'_cont_pcal2', xaxis = 'time', yaxis = 'phase', spw = '', iteration = 'antenna', subplot = 421, plotrange = [0,0,-200,200]) applycal(vis = 'f'+str(field)+'_cont.vis', spw = '', gaintable = ['f'+str(field)+'_cont_pcal2'], spwmap = [0,0,0], calwt = T, flagbackup = F) clean(vis = 'f'+str(field)+'_cont.vis', imagename = 'f'+str(field)+'_cont_pcal2_clean', mode = 'mfs', psfmode = 'clark', niter = 100, threshold = '0.0mJy', interactive = False, mask = 'f'+str(field)+'_cont_b4sc.mask', cell = cell, imsize = imsize, weighting = 'briggs', robust = robust, imagermode = imagermode) im_max = imstat(imagename = 'f'+str(field)+'_cont_pcal2_clean.image')['max'][0] im_rms = imstat(imagename = 'f'+str(field)+'_cont_pcal2_clean.image', region='annulus[['+str(xc)+'pix,'+str(yc)+'pix],['+str(in_a)+'pix,'+str(out_a)+'pix]]')['rms'][0] print 'Peak = {0:.2f} mJy, rms = {1:.2f} mJy, S/N = {2:.1f}'.format(1000*im_max, 1000*im_rms, im_max/im_rms) # Peak = 25.95 mJy, rms = 8.81 mJy, S/N = 2.9 (even worse noise) # inspection of the image shows no change from pcal1 imview(raster=[{'file':'f'+str(field)+'_cont_b4sc.image'}, {'file':'f'+str(field)+'_cont_pcal1_clean.image'}, {'file':'f'+str(field)+'_cont_pcal2_clean.image'}]) # wow. no. just no. # ======================== Best Continuum Map ================== # do not apply self-cal. all were horrible, due to faint source? # deep clean, trying different robust weights # os.system('rm -rf f'+str(field)+'_cont_best*') clean(vis = 'f'+str(field)+'_cont.vis', imagename = 'f'+str(field)+'_cont_best', mode = 'mfs', psfmode = 'clark', niter = 2000, threshold = '0.0mJy', interactive = True, mask = '', cell = cell, imsize = imsize, weighting = 'briggs', robust = 0.5, # CHANGEME imagermode = imagermode) # placed mask around source conservatively (just red parts) # stopped after 200 iterations once the inside became green im_max = imstat(imagename = 'f'+str(field)+'_cont_best.image')['max'][0] im_rms = imstat(imagename = 'f'+str(field)+'_cont_best.image', region='annulus[['+str(xc)+'pix,'+str(yc)+'pix],['+str(in_a)+'pix,'+str(out_a)+'pix]]')['rms'][0] bmaj = imhead(imagename = 'f'+str(field)+'_cont_best.image', mode="get", hdkey="beammajor") bmin = imhead(imagename = 'f'+str(field)+'_cont_best.image', mode="get", hdkey="beamminor") print 'Peak = {0:.2f} mJy, rms = {1:.2f} mJy, S/N = {2:.1f}'.format(1000*im_max, 1000*im_rms, im_max/im_rms) print 'Beam = {0:.2f} x {1:.2f} arcsec'.format(bmaj.get('value'),bmin.get('value')) # robust = 0.5 # Peak = 3.55 mJy, rms = 0.35 mJy, S/N = 10.2 # Beam = 0.34 x 0.30 arcsec # save this to a fits file exportfits(imagename='f'+str(field)+'_cont_best.image', fitsimage='f'+str(field)+'_cont.fits') # measure flux # imview(raster=[{'file':'f'+str(field)+'_cont_best.image'}]) im_rms = imstat(imagename = 'f'+str(field)+'_cont_best.image', region='annulus[['+str(xc)+'pix,'+str(yc)+'pix],['+str(in_a)+'pix,'+str(out_a)+'pix]]')['rms'][0] im_flux = imstat(imagename = 'f'+str(field)+'_cont_best.image', region='circle[['+str(xc)+'pix,'+str(yc)+'pix],'+str(aper)+'arcsec]')['flux'][0] print 'Flux = {0:.2f} mJy, rms = {1:.2f} mJy, S/N = {2:.1f}'.format(1000*im_flux, 1000*im_rms, im_flux/im_rms) # Flux = 4.96 mJy, rms = 0.35 mJy, S/N = 14.2 # re-center image on source and use get_flux.py to get COG flux ia.fromimage(outfile = 'f'+str(field)+'_cont_cropped.image', infile = 'f'+str(field)+'_cont.fits', region = box ) ia.close() exportfits(imagename = 'f'+str(field)+'_cont_cropped.image', fitsimage = 'f'+str(field)+'_cont_cropped.fits') # ======================== Measure flux with UVMODELFIT ================== # calculate offset from phase center in arcsec pixscale = 0.03 # must match 'cell' dx = pixscale*(320.0-xc) # offset to east (left) dy = pixscale*(yc-320.0) # offset to north (up) # measure flux as gaussian uvmodelfit(vis = 'f'+str(field)+'_cont.vis', comptype = 'G', sourcepar = [im_flux,dx,dy,0.5,1,0], varypar = [T,T,T,T,T,T], niter = 10) ''' reduced chi2=1.39447 I = 0.00468205 +/- 0.000454796 x = -0.13654 +/- 0.0127005 arcsec y = 0.318432 +/- 0.0117113 arcsec a = 0.156388 +/- 0.0495202 arcsec r = 1 +/- 0.489232 p = 0 +/- 57.2958 deg errors on FWHM too large, using point source ''' # measure flux as point source uvmodelfit(vis = 'f'+str(field)+'_cont.vis', comptype = 'P', sourcepar = [im_flux,dx,dy], varypar = [T,T,T], niter = 10) ''' reduced chi2=1.39451 I = 0.00389431 +/- 0.000273126 x = -0.137737 +/- 0.0111461 arcsec y = 0.31714 +/- 0.00993015 arcsec slightly lower than aperture method gaussian gave same chi2 but higher flux '''