-- DirkPetry - 2012-05-03

casapy-stable-34.0.19261-001-64b, 16-19 April, Anita Richards, Linux

1. clean on multi-core machine. Multiscale MFS runs very slowly unless setenv OMP_NUM_THREADS 2 . However, is also rather unstable (diverged or crashed), but it did behave sensibly some of the time. The only difference was how many minor iterations and cycles.
    • Also, attempting to use pbcorr=T with nterms > 1 gives an error message which is OK (although it would be good to be able to use it, I do not see why it is scale dependent). Using mask=0.2 crashes - presumably for a similar reason, again, ideally it would work, but if not it should give a message.

2. Hereafter on my linux laptop: Working through NGC3256 CASAGuide: gencal complains 'Inconsistent number of parameters specified.' for the Guide inputs, which work in 3.3
      gencal(vis=name+'.ms', caltable='cal-'+name+'_del.K', caltype='sbd', antenna='DV07', pol='X,Y', spw='1,3,5,7', parameter=[1.00, 1.10, -3.0, -3.0, -3.05, -3.05, -3.05, -3.05])
      error 'Inconsistent number of parameters specified.' (works OK in 3.3)

TICKET CAS-4011 - fixed by George in r19478, confirmed by Anita

3. Tried using 'K' mode in gaincal instead, corrections did not behave as expected when applied, may be my error, will investigate further.

4. Applycal including 'K' table gave error Table column SPECTRAL_WINDOW_ID is unknown

5. tflagdata(vis='ngc3256_line.ms', mode='manual',flagbackup=T, spw='*:0~16;125~127') does not work, have to do the two selections separately, i.e. tflagdata(vis='ngc3256_line.ms', mode='manual',flagbackup=T, spw='*:0~16') and ...spw='*:125~127'...

casapy-stable-34.0.19261-001-64b , 16-19 April, Jeremie Boissier, Linux (italian arc cluster)

  1. PCLEAN of calibrator (mode continuum) --> very long and crash at some point with: ** Error ** coercing to Unicode: need string or buffer, list found I tried twice: On the first try it has been very early, the second time it ran for more than 30min (I got the evolution 10--> 100% several time) before the crash.

casapy-stable-34.0.19261-001-64b, 19-?April, Roberto Galvan-Madrid, Linux (arca4 @ ESO-Garching)

* April 19 2012
-casaviewer
The 'Tools' -> 'Shape Manager' option to load markers that was present in 3.3 is not there anymore.
I tried to load my .reg file with 'Data' -> 'Open' -> 'Load Regions', but viewer crashes giving message:
CASA <4>: 2012-04-19 12:05:15 WARN CASA::Viewer currently only supports rectangle, ellipse, symbol (somewhat), and polygon region shapes

If I use the regions menu at the bottom it also crashes without error message.

Although the original .reg file used circles, I tried a creating one with a single rectangle (and tested that worked on CASA 3.3),
and viewer still crashes giving the same error message.

Suggestion: It would be nice to have the Shape Manager back.

-casaviewer
It is odd that now, if I open an image and click on it, the regions menu is automatically loaded at the bottom
of the GUI, making the fraction of the screen allocated to the map itself very small.

There is a solution to this. A setup file in .casa (.casa/viewer/rc) has to be deleted. Requested that this be added to the known issues.

casapy-stable-34.0.19261-001-64b, 17 April 2012, Rosita Paladino, Linux (italian arc cluster)


# CLEAN

- Issues
On CenA data I had problems with the task clean in interactive mode,
using stable versions r18527 and r17353.

The r19261 gives the same error message.

These are the clean input:
clean(vis='CenA.cal.line.ms.contsub',
     imagename='CenA.CO2_1Line_TEST.clean',
     spw='0',field='',phasecenter='0',
     mode='velocity',outframe='LSRK',restfreq='230.538GHz',
     nchan=35,start='200km/s',width='20km/s',
     imagermode='mosaic',
     imsize=1500,cell='0.2arcsec',minpb=0.2,
     interactive=T,
     weighting='briggs',robust=0.5,
     niter=10000, threshold='0.01mJy')

It is faster than the previous releases.

But after two cycles it gives this error message:
terminate called after throwing an instance of 'DBus::Error'
  what():  Did not receive a reply. Possible causes include: the remote
application did not send a reply, the message bus security policy blocked the
reply, the reply timeout expired, or the network connection was broken.
Abort

It is the same using:
"setenv OMP_NUM_THREADS 2" before starting CASA

(see also report by Boissier and Paladino further below)

TICKET CAS-4031 still under investigation by Darrell and Urvashi as of 2 May 12


#CLEAN with NTERMS >1
     It works!


#VIEWER

- Suggestions
save and load regions
It would be useful to be able to delete all the regions
from the panel. I did'nt find the way, if it is already implemented.
In the spectral profile tool. Is it possible to save gaussian properties
(Cent., FWHM, Equ Width)?
They appear in the bottom of the panel but cannot be saved.

- Issues
I saved a region file with boxes and poligonal regions.
Trying to load it, I got this error:
pure virtual method called
terminate called without an active exception

and the viewer crashes.

After two attempts trying to open viewer again:
---------> viewer()
**** Error ****  viewer() failed to get the current working directory [<type
'exceptions.RuntimeError'>: DBus Viewer service has exited....]

Need to exit CASA to be able to open the viewer again.
TICKET CAS-4012 potentially fixed in r19461, waiting for confirmation from Rosita Paladino

# SPLIT & CLEAN
Earlier, I got a problem with the clean after a split (width=7, wich was not a divisor of the original number of channels, 3840).
The new version of SPLIT gives these warning:

WARN    split::SubMS::fillDDTables()    The last output channel of spw 0 has only 4 input channels.
WARN    split::SubMS::fillDDTables()    You will not be
able to export an MS where the width varies by channel to UVFITS!

and produces an ms file with 549 channels, 427.25 kHz.

The user should also be warn about the clean.
In this case Clean makes an image with 960 channels, 244.14 KHz wide.
Which is not what one would expect.

TICKET CAS-4013 documentation was improved in r19475, a more general solution in split is in preparation for CASA 3.5

casapy-stable-34.0.19406-001.tar.gz, 24 April, Ivan Marti-Vidal, Linux (Xubuntu 11.10. Laptop)

  • While running the script for the M100 calibration, I get many errors from plotcal: "SEVERE Exception Reported: TablePlot? : Error in Inputs:" There is no additional information on the errors in the log. However, the plots seem to be well generated.

see Boissier and Paladino below

When running 'help' on any task, I get: "less: Symbol `ospeed' has different size in shared object, consider re-linking WARNING: terminal is not fully functional". However, I get the helpt texts afterwards.

TICKET CAS-4027 may be a freak feature of Xubuntu, waiting for comment from Wes

  • It would be nice to have a list of contour values embeded in the lower part of the contour plots, as well as the value of the peak intensity (i.e., a la AIPS). If that would be possible, all the necessary information to interprete the contour plots would be provided in the figures. That would also simplify the content of the Figure captions in publications (since it wouldn't be necessary to write, for instance, the values of the contours by hand).

Improvement TICKET CAS-4028 for CASA 3.5 or later

CASA 3.4.0 (r19406), 24-26 April, Jeremie Boissier & Rosita Paladino, Linux (italian arc cluster)

Error in plotcal Tsys:
spw='9'
Inputs:
* Error * TablePlot? : Error in Inputs: 

TICKET CAS-4029 (may be a duplicate of CAS-4024) fixed in r19504, confirmed

* after the application of the preliminary tables (tsys, wvr, delay) it becomes impossible to visualize the data with plotms using an averaging in time (it is OK with a channel averaging) ERROR: Error during cache loading: Big problem in finalizeAverage!

Existing TICKET CAS-3952 supposedly fixed in r19534, needs confirmation

* Applycal: In the input the gaintable name does not appear in red if the table does not exist! The error is reported only after the initialization of the corrected column.

The same happens when launching applycal with NO gaintable! A check on the input would be useful before the initialization of the dataset, which sometimes takes long time.

TICKET CAS-4030 fixed in r19558

* Clean

For previous releases we reported about a problem with interactive clean ("DBus error" ....).

We checked also this release and the error is still there:

terminate called after throwing an instance of 'DBus::Error'
what(): Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.
Abort 

TICKET created for first report, see above

casapy-stable-34.0.19406-001-64b, 19-26 April, Roberto Galvan-Madrid, RHEL Linux (arca4 @ ESO-Garching)

While trying to make a large cube containing an SMA mosaic of 7 pointings, 512x512 spatial pixels, 128x24 channels (24 spws of 128 channels each) clean crashes.

No error message in log, last messages were:

2012-04-26 15:15:43 INFO Imager::imagecoordinates() Image spectral coordinate:1721 channels, starting at radio velocity Radialvelocity: -654.187 stepped by Radialvelocity: 1500
2012-04-26 15:15:43 INFO Imager::imagecoordinates() + Reference Frequency = 220.76 GHz, spectral increment = -0.00110456 GHz, frequency frame = LSRK
2012-04-26 15:15:43 INFO Imager::imagecoordinates() + Rest frequency is 220.76 GHz
2012-04-26 15:16:12 INFO imager::clean() Using multifield Clark clean
2012-04-26 15:16:12 INFO imager::createFTMachine() Multiple fields or facets: transforms will be padded by a factor 1.2
2012-04-26 15:16:12 INFO imager::createFTMachine() Performing mosaic gridding
2012-04-26 15:16:12 INFO Imager::setmosaicftmachine Using defaults for primary beams used in gridding
2012-04-26 15:16:12 INFO VPSkyJones::VPSkyJones PB used SMA
2012-04-26 15:16:14 INFO imager::clean() Clean gain = 0.1, Niter = 0, Threshold = 0
2012-04-26 15:16:14 INFO imager::clean() Starting deconvolution
2012-04-26 15:17:49 INFO SimplePBConv::findConvFunction Convolution support = 16 pixels in Fourier plane
2012-04-26 15:26:50 INFO CubeSkyEquation::makeSimplePSF Maximum of approximate PSF for field 0 = 1.62762 : renormalizing to unity
2012-04-26 15:27:03 INFO ImageSkyModel::makeApproxPSFs bmaj: 5.18076", bmin: 2.72012", bpa: 176.301 deg

and in terminal meesages are:

2aaab011a000-2aaab0319000 ---p 00005000 08:03 24415187 /mnt/arch01/data1/opsw/alma/casapy-stable-34.0.19406-001-64b/lib64/python2.6/site-packages/scipy/linalg/clapack.so
2aaab0319000-2aaab031a000 rw-p 00004000 08:03 24415187 /mnt/arch01/data1/opsw/alma/casapy-stable-34.0.19406-001-64b/lib64/python2.6/site-packages/scipy/linalg/clapack.so
2aaab031a000-2aaab0324000 r-xp 00000000 08:03 24415204 /mnt/arch01/data1/opsw/alma/casapy-stable-34.0.19406-001-64b/lib64/python2.6/site-packages/scipy/linalg/calc_lwork.so
2aaab0324000-2aaab0523000 ---p 0000a000 08:03 24415204 /mnt/arch01/data1/opsw/alma/casapy-stable-34.0.19406-001-64b/lib64/python2.6/site-packages/scipy/linalg/calc_lwork.so
2aaab0523000-2aaab0526000 rw-p 00009000 08:03 24415204 /mnt/arch01/data1/opsw/alma/casapy-stable-34.0.19406-001-64b/lib64/python2.6/site-packages/scipy/linalg/calc_lwork.so
2aaab0526Aborted

The same happens in version 3.3 but error messages are different: 'index out of list', ' Exception: std::bad_alloc', and

2012-04-24 17:08:57 SEVERE image::detached Image is detached - cannot perform operation.
2012-04-24 17:08:57 SEVERE image::detached+ Call image.open('filename') to reattach.

My practical solution was to image spw by spw, but I wonder if CASA is now able to make mosaics of thousands of channels without problem.

TICKET CAS-4032 may be due to insufficient SWAP space, the machine in question had 12 GB RAM and only 8 GB swap, requested upgrade

CASA 3.4.0 (r19148), 12-25 April, Gaelle Dumas, Scientific Linux

I reduced and imaged two SV dataset (N3256 and IRAS16293), comparing casa3.3 with casa3.4 for the first one

problems encountered:
1. tflagdata with mode='list' unable to read input file

#create input file:
flagmodes="mode = 'manual' intent='*POINTING*'\n"\
"mode = 'shadow'\n"\
"mode = 'manual' antenna='*&&&'" #autocorrelation
#pas de atmosphere calib ici voire listobs....

flagfile = 'IRAS16293_1stflags.txt'

if os.path.exists(flagfile):
os.system('rm '+flagfile)

with open(flagfile, 'w') as f:
f.write(flagmodes)


f.close()

for name in data:
#tflagdata(vis=name+'.ms', mode = 'unflag', flagbackup = F)
tflagdata(vis=name, mode = 'list', inpfile=flagfile, flagbackup = F)
flagmanager(vis = name, mode = 'save', versionname = 'Apriori')
#2012-04-19 13:22:38 SEVERE tflagdata::::casa Error reading the input file IRAS16293_1stflags.txt
#**** Error ** Error reading the input file IRAS16293_1stflags.txt


the input file is:
> more IRAS16293_1stflags.txt:

mode = 'manual' intent='*POINTING*'
mode = 'shadow'
mode = 'manual' antenna='*&&&'



TICKET CAS-4033 resolved as won't fix - was a problem of having spaces before and after the "=", see ticket

2. gaincal for self-calibration (imaging part): segmentation fault
gaincal(vis='I16293_corrected.ms',
caltable='I16293_corrected.ms.selfcal.pcal1',
field='0,1',refant='DV03', gaintype='T',
spw='0:410~430',
calmode='p',solint='inf',combine='',
minsnr=2.0,minblperant=4)


> segmentation fault

but no warning/error and the table is not created:
2012-04-18 08:28:49 INFO gaincal::::casa
2012-04-18 08:28:49 INFO gaincal::::casa ##########################################
2012-04-18 08:28:49 INFO gaincal::::casa ##### Begin Task: gaincal #####
2012-04-18 08:28:49 INFO gaincal::::casa
2012-04-18 08:28:49 INFO gaincal::calibrater::open Opening MS: I16293_corrected.ms for calibration.
2012-04-18 08:28:50 INFO gaincal::Calibrater:: Initializing nominal selection to the whole MS.
2012-04-18 08:28:50 INFO gaincal::calibrater::setdata Beginning selectvis--(MSSelection version)-------
2012-04-18 08:28:50 INFO gaincal::calibrater::reset Reseting solve/apply state
2012-04-18 08:28:50 INFO gaincal::Calibrater::selectvis Performing selection on MeasurementSet?
2012-04-18 08:28:50 INFO gaincal::Calibrater::selectvis+ Selecting on spw: '0:410~430'
2012-04-18 08:28:50 INFO gaincal::Calibrater::selectvis+ Selecting on field: '0,1'
2012-04-18 08:28:50 INFO gaincal::Calibrater::selectvis Selection did not drop any rows
2012-04-18 08:28:50 INFO gaincal::Calibrater::selectvis Frequency selection:
2012-04-18 08:28:50 INFO gaincal::Calibrater::selectvis . Spw 0:410~430 (21 channels, step by 1)
2012-04-18 08:28:50 INFO gaincal::calibrater::setsolve Beginning setsolve--(MSSelection version)-------
2012-04-18 08:28:50 INFO gaincal::Calibrater::setsolve Arranging to SOLVE:
2012-04-18 08:28:50 INFO gaincal::Calibrater::setsolve . T Jones: table=I16293_corrected.ms.selfcal.pcal
1 append=false solint=inf refant='DV03' minsnr=2 apmode=P solnorm=false
2012-04-18 08:28:50 INFO gaincal::calibrater::solve Beginning solve-----------------------------
2012-04-18 08:28:50 INFO gaincal::Calibrater::solve The following calibration terms are arranged for ap
ply:
2012-04-18 08:28:50 INFO gaincal::Calibrater::solve . (None)
2012-04-18 08:28:50 INFO gaincal::Calibrater::solve The following calibration term is arranged for solv
e:
2012-04-18 08:28:50 INFO gaincal::Calibrater::solve . T Jones: table=I16293_corrected.ms.selfcal.pcal
1 append=false solint=inf refant='DV03' minsnr=2 apmode=P solnorm=false
2012-04-18 08:28:50 INFO gaincal::Calibrater::solve Solving for T Jones
2012-04-18 08:28:50 INFO gaincal:::: For solint = inf, found 36 solution intervals.

exiting CASA and entering again didn't solve the pb

TICKET CAS-4034 may be fixed in stable r19502, confirmed by Dirk

3. viewer:

viewer -> open dirty image
-> spectral profile
-> draw box

viewer crashed:
unhandled exception...
from "" from event 2
exiting ...

no message in the logger
I exited CASA then enter again and redid it with no error, so I don't know what happened...

when drawing a region in an image, the region menu appears and then the image gets smaller.
Would it be possible to keep the size of the image (with a scroller for example....) 
TICKET CAS-4035 resolved, could not reproduce

casa stable 19261, Maria Diaz Trigo, MacOS? X 10.6.8

1. tflagdata: I tested this command extensively, and it seemed to work fine. Contrary to results from Gaelle, in the MacOS? , reading from a 'list' works as expected (see below full command, tested on SgrA? _B6 SV data):

flagcomd= "mode='manual' antenna='DV05'\n"\
"mode='manual' intent='*POINTING*'\n"\
"mode='manual' intent='*ATMOSPHERE*'\n"\
"mode='shadow'\n"\
"mode='manual' antenna='*&&&'" # autocorrelations

flagfile = "DV05_auto_point_atm_sh.txt"
with open(flagfile, 'w') as f:
f.write(flagcomd)
f.close()

for i in range(len(svis)):
tflagdata(vis=svis[i],mode='list',inpfile=flagfile,flagbackup=False)

2. wvrgcal: I tested also this command, and it worked fine for the 'r19261' version (it did not work for previous releases). The command I used was:

wvrgcal(vis="Xb59.ms",caltable='cal-wvr_Xb59_bis', toffset=-1, wvrflag=['DV05'], segsource=True, statsource="nrao530")

I used SgrA? _B6 SV data again.

3. I tested Tsys tables generation, conversion of old Tsys tables to new format, and applycal (with 'linear' interpolation for the tsys). I could generate Tsys tables with gencal and convert old tables with caltabconvert. However,trying to apply such tables did not work. I got errors like:


2012-04-18 12:58:48 SEVERE applycal Table column SPECTRAL_WINDOW_ID is unknown Check inputs and try again.
2012-04-18 12:58:48 SEVERE applycal Exception Reported: Error in Calibrater::setapply.

I tried it both on new and converted tsys tables (with SgrA? _B6 and BR1202_B7 SV data). I have seen that Miro used a workaround to get through the problem above, but I guess this should still be reported as a bug.

4. viewer: I tested the new features 'very superficially', but it seemed OK. In particular, fitting and collapsing in the spectral profiler and export of images worked as expected.

casapy-stable-34.0.19406-001-64b, A. Rossetti, Italian ARC node

(confirmed findings from above using complete M100 ALMA SV dataset)

VIEWER: I could not change the box shape interactively (this is possible in 3.3.0 for all type of boxes).

TICKET CAS-4036 * fixed in r19549, needs confirmation*
Topic revision: r6 - 2012-05-03, DirkPetry
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback