Category: IDL hints

In our galactic magnetic field project,  color plots seem to fail sporadically, leaving plots with blank places where there should be colored symbols or lines.  Very frustrating!  I am finding this with AO_ATNF_Pulsars_RM_Sag_Region1a.pro

So I tried manually to play around with this from the IDL command line, basically trying to duplicate it.  Note that the above program uses cgcolor multiple times to set up definitions of say red to then be used in plot procedure (without quotes, e.g. plot,x,y,red ). Note comments in above
say we needed to use cgcolor because regular idl colors
were not working. Note all this is with plot procedure, and
newer plot function may not have any of these probs! New programs should use the latter.

From command line, if I do the following:

x=findgen(10)
y=findgen(10)
set_plot,’ps’  (I think needs it only once)
device,filename=’help.ps’ (puts the file in home directory)
Red =cgColor(“Red”, !D.Table_Size-4)  (stolen from pgm, and cgcolor is a coyote routine [there are such calls for every color used])
plot, x, y, color=red (makes every bit of plot red)
device,/close

The above made a red plot in the file ~help.ps .

And then FWIW when I reran the AO_ATNF . . . Sag..1a.pro,
the two plots’ colors were there. I’m not saying cause and
effect, but at least in this case a rerun after above jimmying
brought the colors back.  I am guessing that somehow I
am not properly totally initializing colors when I run the
program, so if something color-related  is left in a funny
state by a previous run, then the next program run does
not automatically fix it.

Coyote has some suggestions at
http://www.idlcoyote.com/color_tips/redps.html   ,
including:

    Set_Plot, 'PS'    
Device, DECOMPOSED=1, COLOR=1, BITS_PER_PIXEL=8, filename='help.ps' 
(Joel added the filename part.)

and he claims that all the above are needed for color to
work, so maybe the program’s lack of such a full Device command is the problem. However the above advice is old and I’m not
sure if today’s idl might need say 24 bits per pixel???? But
the color=1 command sounds like it is turning on color!  Decomposed=1 turns on decomposed colors which I think means R,G,B are expressed separately. That is documented here and there but I’m not sure if our cgcolor routines expect that value, so, be careful!

Sorry I don’t have the definitive fix but hopefully the user can create a workaround with some subset of above, if she suddenly loses color on plots that were colored in previous program runs!

–Joel

using breakpoints on idlde over vncviewer (as opposed to chicken-of-the-vnc) was fraught.  The usual key combinations would not place breakpoints anywhere.  But I eventually learned that:

DOUBLE CLICKING ON GRAY BORDER AT VERY LEFT OF DESIRED LINE TOGGLES BREAKPOINT

Here is a harder way that I learned before above red easy way:

1. select desired line
2. shift control b toggles a breakpoint on that line.

Here is an even harder way to do it from keyboard:

help,/breakpoint  will give a list of existing breakpoints

cd, ‘full path to desired directory’ will change idl’s current directory from users home directory (the default) to another. remember apostrophes CASE SENSITIVE

printd   prints the current directory

IDL> printd
Current Directory: /data/psrdata/arecibopuppi/B1913+16/2019Aug
Directory stack is empty.

breakpoint,/set, ‘filename inside of current idl directory path [see printd above to discover current path value] ‘, 150   (use apostrophes, note final value is linenumber) CASE SENSITIVE

i think i did not see the red circle on line 150 of idlde until i recompiled.

check out help,/breakpoint to see a list of current breakpoints, including their “index” which can be used in some other breakpoint commands like breakpoint,/clear,1 (index 1)

IDL> help,/breakpoint
Index Line Attributes                                                 File
1         150      Pro=PLOTADAYSRMTOT       /data/psrdata/arecibopuppi/B1913+16/2019Aug/plotAdaysRMtot.pro

I have noticed slight object-oriented differences in plotting on mirzam (v8.6)  vs

on deneb or canopus (v8.3)

–joel

Carl Thomas got running DIALOG_PICKFILE in IDL, which enables one to point and click one’s way to selecting a desired file, say for processing.  I (Joel ) embellished it a little to get it to allow choice of both file AND subdir.  See spectraoverlay.pro in /data/psrdata/arecibowappHI/overlayLAB-EBHIS/

idl 8.3 is on deneb and canopus.  I have no maintenance contract on them any more, so they are frozen at that version.  They reside on local disks on each machine at /usr/local/exelis. . .

i dropped maint on mirzam for about 10 months in 2016(?). When I picked it up again, we loaded idl 8.6 into /usr/local/harris/ . . .

–Joel

Here’s how to figure out the contents of a save set.

1. This is the classic way:
restore,’filename.sav’,/verbose  will list the variable names in the saveset.

help, variable_name will then give details on any such variable.

2. This  is the modern (object-oriented) way. Note that it does not require a restore:

[After http://northstar-www.dartmouth.edu/doc/idl/html_6.2/IDL_Savefile.html#wp1034474 ] :

Use the IDL_Savefile object to query the contents of a SAVE file containing data and selectively restore variables.

So first, create a savefile object out of of the .sav file:   sObj = obj_new(‘IDL_Savefile’, ‘p2206.B1913+16.wapp4.53958.0004_shifted.sav’)

Typically, IDL_Savefile::Contents is the first method called on a savefile object; the information retrieved indicates the number of different types of items contained in the file (variables, routines, etc.).

sContents = sObj->Contents()

one type of query of sContents is:  print, sContents.N_var
20

Next, a call to IDL_Savefile::Names is used to obtain the names of items of a specific type.

sNames = sObj ->Names()
print,snames
ACTUAL_NRECS BINSHIFT BW DELFREQ IDLFILE LSRK NBIN NCHAN OBSJULDATE OBSTIME PROFILE_VERSION ROUNDBINSHIFT SCANNO SHIFTONOFFA SHIFTONOFFB
SHIFTTOTA SHIFTTOTB SRC VEL WAPPHOST

Then,  IDL_Savefile::Size can be used to determine the size of a given (named) variable in the file:

size_of_shiftonoffa = sObj->Size(‘shiftonoffA’)
print, size_of_shiftonoffA
1    512   4    512,  which means:

If no keywords are set, SIZE returns a vector of integer type. The first element is equal to the number of dimensions of Expression. This value is zero if Expression is scalar or undefined. The next elements contain the size of each dimension, one element per dimension (none if Expression is scalar or undefined). After the dimension sizes, the last two elements contain the type code (zero if undefined) and the number of elements in Expression, respectively. The type codes are listed at https://www.harrisgeospatial.com/docs/size.html.  (the four here means floating point).

–Joel