Date - 2001.04.27 Tester - S.T. Myers (AOC) Platform - Linux (sandrock) Note: bugs are denoted by lines with prefix: >>>BUG: and queries are denoted by lines with prefix: >>>QUERY: while comments are prefixed by: >>>COMMENT: Some enhancement requests are given after >>>REQUEST: ------------------------------------------------------------------------------- GOAL: Check out aips++ performance for Debra's Q-Band 3C84 fast-switching expt 1. Date - 2001.04.27 Version - Weekly (1.5 build 340) Data - /home/sandrock/smyers/Testing/Apr01/q3c84.ms Interactive mode. ------------ ms.summary() ------------ MeasurementSet Name: /export/home/sandrock/smyers/Testing/Apr01/q3c84.ms MS Version 2 Observer: unavailable Project: DSTST Observation: VLA(27 antennas) Fields: 1 Name: 0319+415 RA: 03:19:48.16 Dec: +41.30.42.11 (J2000 ) Data records: 421848 Total integration time = 2830 seconds Observed from 22-Dec-2000/03:44:58 to 22-Dec-2000/04:32:08 Data descriptions: 2 (2 spectral windows and 1 polarization setups) ID Ref.Freq #Chans Resolution TotalBW Correlations 1 43314.9MHz 1 60500 kHz 50000 kHz RR LL RL LR 2 43364.9MHz 1 60500 kHz 50000 kHz RR LL RL LR Antennas: 27 ID= 1-5: 16=VLA:W48, 21=VLA:W40, 23=VLA:W16, 22=VLA:W8, 10=VLA:W24, ID= 6-10: 11=VLA:W32, 26=VLA:W64, 13=VLA:W72, 24=VLA:W56, 15=VLA:E48, ID= 11-15: 28=VLA:E40, 7=VLA:E16, 6=VLA:E8, 14=VLA:E24, 20=VLA:E32, ID= 16-20: 9=VLA:E64, 17=VLA:E72, 12=VLA:E56, 19=VLA:N48, 18=VLA:N40, ID= 21-25: 8=VLA:N8, 25=VLA:N32, 2=VLA:N16, 3=VLA:N24, 27=VLA:N64, ID= 26-27: 1=VLA:N72, 4=VLA:N56 Tables(rows): (-1 = table absent) MAIN(421848) ANTENNA(27) DATA_DESCRIPTION(2) DOPPLER(-1) FEED(27) FIELD(1) FLAG_CMD(0) FREQ_OFFSET(-1) HISTORY(0) OBSERVATION(1) POINTING(0) POLARIZATION(1) PROCESSOR(0) SOURCE(-1) SPECTRAL_WINDOW(2) STATE(0) SYSCAL(-1) WEATHER(-1) --------------------- Lets look with msplot --------------------- start with edit=true and flagfile=flagger1 *timing: flags copied in 103s plot amp vs. uvdist LL,RR *timing: selection 263210 rows = 41s *timing: plot 526240 pts = 22s data looks pretty good here, few outliers (previous editing good) plot amp vs. time iterate over ant1 iteration works! though still with pop-up stopsign box (maybe the same box should stay up until done or hit 'dont ask again') plot phase vs. time iterate over ant1 hey, its not asking - oh it remembered the "continue" from when I told it not to ask anymore. phases are in radians - how do I get degrees? (should be default) is see occasional short ramps in phase - probably due to funky interpolation from 5s solutions, could be shorter! >>>QUERY: Is there a way to change the units for the phase in the plots? go back with iteration set to pause try to edit out a few outliers just for fun hey - cant seem to edit while the stopsign is waiting hit STOP thinking this will halt and let me edit, but it dismisses plot and goes to multipage. start again but now iteration is off??? it says its on start again, now stop does stop on that plot and lets me edit, but of course iteration cannot proceed. Dismiss. start again, go to second plot and try to edit - AHA, when you stop on the second or higher plot it goes immediately to the multipage thing that wont let you edit. Go to Options and turn off mulipanel plots and see if this cures this start again - it only does first plot. Dismiss. start again - now it iterates (I did nothing in between!) stop on AN2 plot. it does stop, I can set a region, but there are no FLAG buttons on the plot window anymore! >>>BUG: This irrational behavior of the plot iteration must be fixed. The goal should be to step through AND BE ABLE TO EDIT! Done for now. The new "Keep new flags" question is much better! Thanks! Now I assume they get applied... The initial flags of this measurement set are saved in the flagger1 table. The msplotapplyflags function can be used to replace the flags in this measurement set with the ones you started with if you need to back out of your edits. ... seems clear enough! Except does this mean that the new flags are not in flagger1 (which is what I wanted) but are in the data somewhere? I guess this is ok but should probably be rationalized later on. -------------- Try calibrater -------------- >>>QUERY: Is there a way that the Tool Manager could remember which ms you are using, even across tools (it remembers in the same tool) - it seems a little silly that it can't guess you might want to run calibrater on the same ms you just did msplot on! (Actually running from GUI - commands just for reference) calib:=calibrater(filename="q3c84.ms" ); ok:=calib.setdata(mode="none" , nchan=1, start=1, step=1, mstart=[value=0.0, unit="km/s" ], mstep=[value=0.0, unit="km/s" ], uvrange=0, msselect=''); ok:=calib.setsolve(type="G" , t=60, preavg=0.0, phaseonly=F, refant=21, table="q3c84.calib" , append=F); ok:=calib.solve Looks like its going through slots at about 1/sec. Lots of them with "Sum of weights is zero" - can it know to skip these? Is it wasting time reporting each slot (Ive found I/O to terminal can really slow code)? *timing: 2646 slots in 18m31s and this was with no solutions with valid weights! what went wrong? 2. Date - 2001.04.29 Version - Weekly (1.5 build 354) Data - /home/sandrock/smyers/Testing/Apr01/q3c84.ms Interactive mode. Try to look at CAL_HISTORY table of q3c84.calib using Catalog GUI. Says "NO ROWS IN TABLE". Is there anything here? >>>QUERY: How does one look at the CAL_HISTORY table, or is just currently empty? What is it good for? It is not obvious how to go about diagnosing this problem. I would like to check out the current state of flags, for example. Where are these? OK, try VIEW from Catalog on q3c84.ms (not the subtables). Looks like there are lots of flags with [T, T, T, T] and with [F, F, F, F]. >>>QUERY: Where is there a one-stop description of what exactly is in a measurement set (like what all the tables are, and entries)?