Testing log - 2007.06.27
Steven T. Myers

-----------

Test using Jupiter 6cm Synth2000 tutorial data:
/home/sandrock2/smyers/NAUG2/Data/VLA_CONT/FLUX99-6CM.CBAND
or
http://www.aoc.nrao.edu/~smyers/naug/Data/tut2000/VLA_CONT/FLUX99-6CM.CBAND

Machine: sandrock
/home/sandrock2/smyers/Testing2/Jun07/

-----------

Note: this is daily build 2666 because the RPM was broken

*** = serious problem
**  = serious problem
*   = important issue

importvla :

   better to use '' as defaults for start and stop times (as in other
   timeranges)  Should we use timerange instead?

   if you give it a list of files, e.g. archivefiles =
   ['AP314_A950519.xp1','AP314_A950519.xp2','AP314_A950519.xp3']
   then it will show as red (but still work)

plotxy :

   should add the xaxis,yaxis options: re, im, wt
   (if keeping different xaxis,yaxis options re, im are both and wt is y)

gaincal :

   should provide more feedback, such as SNR of solutions and number
   of successful/failed solutions (use AIPS CALIB as guide)
   could have a verbose=True mode to have more detailed reporting
   
plotcal :

   *** using MarkRegion and Locate causes coredump

   should be as similar to plotxy as possible

   should use Next button for iteration rather than hot keys n and s

   but, keys should work in plot window not shell, and should not require
   <return> (should really be hot-keys)

   *** using multiplot=True (which should iterate over antenna) as you
   iterate it keeps adding antennas (even if overplot=False) to plot,
   needs to clear

   ** need legend to explain colors (e.g. for spw and pol), or at least
   send info to shell/logger, otherwise will have to use Locate to
   figure out (but see above)

   window name should be 'CASA Plotcal 1' instead of generic 
   'CASA Figure 1'

   no way to quit other than killing window, seems to stay active even
   after done plotting and returned to CASA prompt. Add Quit button.

   sometimes you get 2 rows of buttons, this seems to be after
   plotxy has used the same plot window

   ----------------
   need better multiplot/multipanel control, e.g.

      multiplot  = 'none'   # none, antenna, poln, channel
      multipanel = 'none'   # none, antenna, poln, channel

      e.g. multiplot='antenna', multipanel='poln'  would do a 2 panel
      plot (R and L) with iteration over 'antenna'

      should make multiplot/iteration consistent with plotxy
   ----------------

   spews lots of "Reading data..." and "Time to extract Flags" lines
   to shell - unnecessary

accum :

   unclear to me for interpolation whether to use tablein or incrtable

   if using for interpolation (with incrtable) how to control which
   fields are interpolated ONTO (seems calfield is the FROM).

   should have smoothing options (as in smoothcal or AIPS CLCAL)
   see smooth...

smooth :

   should be smoothcal

   can this be just incorporated into accum?
   and combined with correct (see below)?

correct :

   should be named applycal

   not clear how solutions are interpolated onto target sources
   nearest? 2pt? doesnt seem to be any control of this
   Joe says its currently hardwired to 'nearest'

   I dont understand what blbased does

   should we have the interpolation in accum (where it is now)
   or in correct? or both?

   ----------------
   NOTE: I am inclined to combine accum + smooth(cal) + correct into the
   single task "applycal" which will have params:

   vis 

   field
   spw
   selectdata (and its sub-params) # as in correct

   # Note: the selections field/spw/selectdata will apply to the
   # tablein for mode=accum, and to data and tablein for mode=correct
   # I dont think we need gainselect (but if so put in mode=correct)

   interp # add the option 'boxcar' which adds the sub-pars
      smoothtype = 'median'  # median, mean
      smoothtime = 60.0      # boxcar length (sec)

   tablein = ''        # Input calibration table

   mode = ''           # correct, accum, interp

   mode = 'correct'    # apply tablein to DATA to make CORRECTED_DATA column
      calwt = True     # correct the weights also

   mode = 'accum'      # accumulate tablead onto tablein make caltable
      tablead = ''     # name of incremental cal table to add
      fieldad = ''     # fields from incrtable to add
      spwad = ''       # spw:chan from incrtable to add
      caltable = ''    # output caltable name
      accumtime = -1.0 # cumulative table timescale when creating

   mode = 'interp'
      caltable = ''    # output caltable name
      calfield = ''    # which fields are interpolated onto?
      accumtime = -1.0 # cumulative table timescale when creating
      
   spwmap              # map of spw to apply/form (<0 apply to self only)

   I'm not sure what to do with opacity and gaincurve, I would maybe
   add gaincurve and tau as mode=correct options

   ----------------

split : 

   seems AFU in this version, cannot specify field (seems to want id)
   wait for tomorrows fix