| Re: legend not working with symcat [message #68519] |
Thu, 05 November 2009 01:24 |
maurya
Messages: 16
Registered: February 2009
|
Junior Member |
|
|
On 5 Nov, 05:48, wlandsman <wlands...@gmail.com> wrote:
> On Nov 4, 6:19 pm, David Fanning <da...@dfanning.com> wrote:
>
>> Wayne Landsman writes:
>>> That would be true if all users had symcat.pro in their !path. But
>>> in case there are some inexperienced users who don't have the Coyote
>>> library installed yet, I've slightly modified your update to
>>> legend.pro to only call symcat.pro for non-standard PSYM values.
>
>>> http://idlastro.gsfc.nasa.gov/ftp/pro/plot/legend.pro
>
>> The problem now is that LEGEND won't compile *unless* you
>> have SYMCAT in your path. :-(
>
> Well, I don't think compilation is a problem, e.g. the program
>
> pro test,dum
> if keyword_set(dum) then a = some_program(3) else print,'AOK'
> return
> end
>
> will run fine so long as one doesn't try to access Some_program().
>
> IDL> test
> % Compiled module: TEST.
> AOK
> IDL> test,1
> % Variable is undefined: SOME_PROGRAM.
> % Execution halted at: TEST 2 /Volumes/Apps_and_Docs/
> landsman/codev/test.pro
>
> (If I had added a compile_opt idl2 at the beginning, it would also
> know that SOME_PROGRAM is an underined function/procedure rather than
> a variable.)
> Similarly, legend.pro should now run fine without symcat.pro in one's !
> path, so long as one doesn't try to use a value of PSYM>=11.
>
> But I agree the more general question is difficult. Currently, I
> keep JHUAPL routines I need in a separate directory, and tell users to
> delete the directory if they have the entire JHUAPL library
> installed.
>
> --Wayne
Thanks David and Wayne,
Actually, n=10 was not the symbol number in my previous post. It is
maximum number of increment in the for loop. The suggestions given by
David is most useful. Now, the "legend.pro" is working for all the
symbols defined in the "symcat.pro" except 10. You can see one of the
following test. Here, I am using new "legend.pro" program given by
Wayne (http://idlastro.gsfc.nasa.gov/ftp/pro/plot/legend.pro). I have
removed the symbol number 10.
-------------------------------------------------------
pro legend_test
set_plot,'win'
window,0,xs=600,ys=800
n=47
xx=findgen(200)
for i=0,n-1 do begin
ii=i+1
yy=sin(xx*!dtor*ii*0.2)
if (i eq 0) then plot,yy,yr=[0,1.1],psym=symcat(i),/yst $
else oplot,yy,psym=symcat(i),symsize=0.6,nsum=2
endfor
lines=indgen(n)
lines(where(lines eq 10)) = 1
item='p='+string(lines,format='(i2.2)')
print,lines
legend, item, psym=lines
stop
end
------------------------------------------------------------ -
Cheers
|
|
|
|
| Re: legend not working with symcat [message #68524 is a reply to message #68519] |
Wed, 04 November 2009 16:48  |
wlandsman
Messages: 743
Registered: June 2000
|
Senior Member |
|
|
On Nov 4, 6:19 pm, David Fanning <da...@dfanning.com> wrote:
> Wayne Landsman writes:
>> That would be true if all users had symcat.pro in their !path. But
>> in case there are some inexperienced users who don't have the Coyote
>> library installed yet, I've slightly modified your update to
>> legend.pro to only call symcat.pro for non-standard PSYM values.
>
>> http://idlastro.gsfc.nasa.gov/ftp/pro/plot/legend.pro
>
> The problem now is that LEGEND won't compile *unless* you
> have SYMCAT in your path. :-(
Well, I don't think compilation is a problem, e.g. the program
pro test,dum
if keyword_set(dum) then a = some_program(3) else print,'AOK'
return
end
will run fine so long as one doesn't try to access Some_program().
IDL> test
% Compiled module: TEST.
AOK
IDL> test,1
% Variable is undefined: SOME_PROGRAM.
% Execution halted at: TEST 2 /Volumes/Apps_and_Docs/
landsman/codev/test.pro
(If I had added a compile_opt idl2 at the beginning, it would also
know that SOME_PROGRAM is an underined function/procedure rather than
a variable.)
Similarly, legend.pro should now run fine without symcat.pro in one's !
path, so long as one doesn't try to use a value of PSYM>=11.
But I agree the more general question is difficult. Currently, I
keep JHUAPL routines I need in a separate directory, and tell users to
delete the directory if they have the entire JHUAPL library
installed.
--Wayne
|
|
|
|
| Re: legend not working with symcat [message #68526 is a reply to message #68524] |
Wed, 04 November 2009 15:19 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
Wayne Landsman writes:
> That would be true if all users had symcat.pro in their !path. But
> in case there are some inexperienced users who don't have the Coyote
> library installed yet, I've slightly modified your update to
> legend.pro to only call symcat.pro for non-standard PSYM values.
>
> http://idlastro.gsfc.nasa.gov/ftp/pro/plot/legend.pro
The problem now is that LEGEND won't compile *unless* you
have SYMCAT in your path. :-(
Of course, that might force users to download the Coyote Library,
a Very Good Thing, if you ask me, but...still.
I'm always in a state of confusion when I want to use a NASA or JHUAPL
routine in a Coyote Library program. What do I do? Include the routine
*with* my routines. (Don't like it, since this is a maintenance
nightmare.)
Add the routine at the front of my routine. (Well, at least I can
guarantee
it *works*.) Or, rename the routine and put it in my file. (This is
what
I usually do, along with the copyright notice, etc.)
In this case, I think calling the routine legend_symcat and sticking
it in front of legend makes sense. But then you can't draw the damn
plot with the SYMCAT symbols unless you compile legend first.
I don't know. That's why I suggested you should handle it. :-)
Cheers,
David
P.S. Why don't you write this as an object, like it should be,
and then we can just subclass off it. ;-)
|
|
|
|
| Re: legend not working with symcat [message #68528 is a reply to message #68526] |
Wed, 04 November 2009 13:59 |
wlandsman
Messages: 743
Registered: June 2000
|
Senior Member |
|
|
>
> Wayne would have to confirm that this is the only place a change
> would have to be made. But I think he could add this to legend
> and gain access to 35 more symbols then he has now and not be
> worried by backward compatibility at all. ;-)
That would be true if all users had symcat.pro in their !path. But
in case there are some inexperienced users who don't have the Coyote
library installed yet, I've slightly modified your update to
legend.pro to only call symcat.pro for non-standard PSYM values.
http://idlastro.gsfc.nasa.gov/ftp/pro/plot/legend.pro
--Wayne
|
|
|
|
| Re: legend not working with symcat [message #68531 is a reply to message #68528] |
Wed, 04 November 2009 11:28 |
wlandsman
Messages: 743
Registered: June 2000
|
Senior Member |
|
|
On Nov 4, 9:21 am, maurya <ramaury...@gmail.com> wrote:
> I changed the "legend.pro" as you have suggested. Now, it is working
> for n<=10 only. For larger values of n>10 it is giving following
> error:
>
> PLOTS: PSYM (plotting symbol) out of range
I suspect that it works for n>10 but is failing for n=10. That is
because the IDL definition of PSYM=10 (which SYMCAT() follows) does
not really define a symbol, but tells IDL to plot in histogram mode.
It *cannot* be used with PLOTS, e.g.
IDL> plot,indgen(10),indgen(10)
IDL> plots,5,5,psym=symcat(10)
% PLOTS: PSYM (plotting symbol) out of range
So it is probably best to not use PSYM=10. (What symbol should one
use in a legend to distinguish a line draw in histogram mode, versus
one connecting the points?)
--Wayne
|
|
|
|
| Re: legend not working with symcat [message #68540 is a reply to message #68531] |
Wed, 04 November 2009 06:35 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
maurya writes:
> I changed the "legend.pro" as you have suggested. Now, it is working
> for n<=10 only. For larger values of n>10 it is giving following
> error:
>
> PLOTS: PSYM (plotting symbol) out of range
>
> When I commentated the line 223 (on_error, 2), the program stop at the
> line 425-427 where the correction is made. Although, according to your
> "symcat" program it should work for 46 symbols. However, I need only
> for n<=12.
Humm. I'm not sure you made the change successfully. :-)
There was a parenthesis missing in my example. Did you fix that?
Can I see the line you fixed?
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Sepore ma de ni thui. ("Perhaps thou speakest truth.")
|
|
|
|
| Re: legend not working with symcat [message #68541 is a reply to message #68540] |
Wed, 04 November 2009 06:21 |
maurya
Messages: 16
Registered: February 2009
|
Junior Member |
|
|
On 4 Nov, 14:41, David Fanning <n...@dfanning.com> wrote:
> maurya writes:
>> The routine 'legend' is not working with the function 'symcat' for
>> the following test program.
>
>> ;;;;;;;;;===============================
>> pro legend_test
>
>> set_plot,'win'
>
>> window,0,xs=600,ys=500
>
>> n=12
>> xx=findgen(200)
>> for i=0,n-1 do begin
>> ii=i+1
>> yy=sin(xx*!dtor*ii*0.2)
>> if (i eq 0) then plot,yy,yr=[0,1.1],psym=symcat(i),/yst $
>> else oplot,yy,psym=symcat(i),symsize=0.6,nsum=2
>> endfor
>
>> lines=indgen(n)
>> item='p='+string(lines,format='(i2.2)')
>> legend,item,psym=symcat(lines)
>
>> stop
>> end
>> ;;;;;;;;;;;;================================================ ======
>
>> Can anyone help.
>
> Probably not. The problem is that SymCat is not "vectorized"
> in this way, and with good reason. Most of the symbols SymCat
> creates have to be built one-at-a-time with USERSYM. Then
> selected with the symbol index 8. So, if it was vectorized,
> you would certainly end up with a vector filled mostly with
> 8's.
>
> IDL> symbols = IntArr(10)
> IDL> FOR j=0,9 DO symbols[j] = SymCat(j + 10)
> IDL> Print, symbols
> 10 8 8 8 8 8 8 8 8 8
>
> In other words, the symbol created by USERSYM is "ephemeral"
> and can only be used for a short time. Well, until the *next*
> symbol is defined.
>
> The only way to fix this, as far as I can tell, would be
> to modify legend.pro to call SymCat whenever is creates a
> symbol.
>
> On lines 425-427, in which PLOTS is being used
> to draw the symbol in the legend with PSYM, I changed
> "psym=psym[i]" to "psym=SymCat(psym[i]" and I changed
> your example program call from this:
>
> legend, item, psym=SymCat(lines)
>
> To this:
>
> legend, item, psym=lines
>
> And all works as you expect.
>
> Wayne would have to confirm that this is the only place a change
> would have to be made. But I think he could add this to legend
> and gain access to 35 more symbols then he has now and not be
> worried by backward compatibility at all. ;-)
>
> Cheers,
>
> David
>
> --
> David Fanning, Ph.D.
> Fanning Software Consulting, Inc.
> Coyote's Guide to IDL Programming:http://www.dfanning.com/
> Sepore ma de ni thui. ("Perhaps thou speakest truth.")
Thanks David,
I changed the "legend.pro" as you have suggested. Now, it is working
for n<=10 only. For larger values of n>10 it is giving following
error:
PLOTS: PSYM (plotting symbol) out of range
When I commentated the line 223 (on_error, 2), the program stop at the
line 425-427 where the correction is made. Although, according to your
"symcat" program it should work for 46 symbols. However, I need only
for n<=12.
RAM
|
|
|
|
| Re: legend not working with symcat [message #68544 is a reply to message #68541] |
Wed, 04 November 2009 01:41 |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
maurya writes:
> The routine 'legend' is not working with the function 'symcat' for
> the following test program.
>
> ;;;;;;;;;===============================
> pro legend_test
>
> set_plot,'win'
>
> window,0,xs=600,ys=500
>
> n=12
> xx=findgen(200)
> for i=0,n-1 do begin
> ii=i+1
> yy=sin(xx*!dtor*ii*0.2)
> if (i eq 0) then plot,yy,yr=[0,1.1],psym=symcat(i),/yst $
> else oplot,yy,psym=symcat(i),symsize=0.6,nsum=2
> endfor
>
> lines=indgen(n)
> item='p='+string(lines,format='(i2.2)')
> legend,item,psym=symcat(lines)
>
>
> stop
> end
> ;;;;;;;;;;;;================================================ ======
>
> Can anyone help.
Probably not. The problem is that SymCat is not "vectorized"
in this way, and with good reason. Most of the symbols SymCat
creates have to be built one-at-a-time with USERSYM. Then
selected with the symbol index 8. So, if it was vectorized,
you would certainly end up with a vector filled mostly with
8's.
IDL> symbols = IntArr(10)
IDL> FOR j=0,9 DO symbols[j] = SymCat(j + 10)
IDL> Print, symbols
10 8 8 8 8 8 8 8 8 8
In other words, the symbol created by USERSYM is "ephemeral"
and can only be used for a short time. Well, until the *next*
symbol is defined.
The only way to fix this, as far as I can tell, would be
to modify legend.pro to call SymCat whenever is creates a
symbol.
On lines 425-427, in which PLOTS is being used
to draw the symbol in the legend with PSYM, I changed
"psym=psym[i]" to "psym=SymCat(psym[i]" and I changed
your example program call from this:
legend, item, psym=SymCat(lines)
To this:
legend, item, psym=lines
And all works as you expect.
Wayne would have to confirm that this is the only place a change
would have to be made. But I think he could add this to legend
and gain access to 35 more symbols then he has now and not be
worried by backward compatibility at all. ;-)
Cheers,
David
--
David Fanning, Ph.D.
Fanning Software Consulting, Inc.
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Sepore ma de ni thui. ("Perhaps thou speakest truth.")
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
