rohit |
04/03/2007 09:58PM (Read 9401 times)
|
|
|
Status: offline
Registered: 06/01/2006
Posts: 86
|
I have been working on low dispersion reduction. I have been using mktracer file to create a trace file from a flat file. I am not sure what I used last night, but other than the command flpr there was nothing that was changed. Since this morning when I do mktracer on a flat file I get the following error. lo> mktracer
Name of input image (24nos0063): 24nos0063
Name of output image (63trace):
ERROR on line 65: floating point overflow
mktracer ()Now this file 24nos0063 is the file I used last night and it created the trace file. When I started to work on it today it just wouldn't do it. It gives me this error. Please help!! NIRSPEC website has the mktracer file under Data reduction section. Thanks
|
|
|
|
fitz |
04/03/2007 09:58PM
|
|
|
Status: offline
Registered: 09/30/2005
Posts: 4040
|
I found the script online via Google but it isn't obvious where the error might be happening. The simplest thing is to type 'd_trace' before running the script to see which task is causing the problem. You might also do a quick IMSTAT on the input image and look for any strange numbers where a division or multiplication could be causing an overflow. Do the same with any of the leftover temp images that get created (e.g. the gradient flat image) Unlearning tasks, or logging in from a new MKIRAF in say /tmp will eliminate potential issues with some strange parameter setting.-Mike
|
|
|
|
rohit |
04/03/2007 09:58PM
|
|
|
Status: offline
Registered: 06/01/2006
Posts: 86
|
Thanks for your email. I did what you suggested and below is what I got. Another thing that I observed it, that every time I try to run the "mktracer" task it created a file mktracer4422b.fits, mktracer4422c.fits, mktracer4422e.fits...etc. This is not what it should create. If I get it a file names 24nos0063.fits and give it a output name as 63trace it should give me 63trace.imh. So, not only I am getting an error but it is creating these strange files. Please help!! In addition, the files are good as they work on another computer. The same files and the same task! But on my laptop it gives me this error. I have used it before and it worked. But all of a sudden it stopped working. The only new thing I ever did was to use "flpr" command. Nothing else.
cl> mktracer
----- task cl -----
0+5: call mktracer
5+2: exec
----- exec script mktracer (/root/Desktop/mktracer.cl) --
----- task mktracer -----
9+4: pushparam input
Name of input image (24nos0063):
13+4: assign i_input
17+4: pushparam output
Name of output image (63trace):
223+4: call imstat
227+5: pushparam gradflat
232+3: posargset 0
235+6: pushconst `max'
241+4: absargset field
245+4: swoff format
249+3: addpipe
252+2: redir
254+2: exec
----- exec external task imstatistics -----
imstatistics > $\nimages="mktracer4738e.fits"\nfields="max"\n
lower=15137.\nupper=16231.\nnclip=0\nlsigma=3.\nusigma=3.\n
binwidth=0.1\nformat=no\ncache=no\nmode="ql"\n$nargs=1\n
--------------------------------
----- task imstatistics -----
522+3: call bye
525+2: exec
----- exec builtin bye -----
256+4: call scan
260+9: pushconst `mktracer.maxval'
269+3: indirposset 0
272+4: getpipe
276+2: redirin
278+2: exec
----- exec builtin scan -----
280+3: rmpipes 1
283+4: call imarith
287+5: pushparam gradflat
292+3: posargset 0
295+6: pushconst `/'
301+3: posargset 1
304+4: pushparam maxval
308+3: posargset 2
311+5: pushparam gradflat
316+3: posargset 3
319+2: exec
----- exec external task imarith -----
imarith\noperand1="mktracer4738e.fits"\nop="/"\n
operand2="INDEF"\nresult="mktracer4738e.fits"\ntitle=""\n
divzero=0.\nhparams=""\npixtype=""\ncalctype=""\nverbose=no\n
noact=no\nmode="ql"\n$nargs=4\n
--------------------------------
----- task imarith -----
522+4: call error
526+5: pushconst 502
531+3: posargset 0
534+12: pushconst `floating point overflow'
546+3: posargset 1
549+2: exec
----- exec builtin error -----
ERROR on line 65: floating point overflow
mktracer ()
|
|
|
|
rohit |
04/03/2007 09:58PM
|
|
|
Status: offline
Registered: 06/01/2006
Posts: 86
|
Forgot to mention this: I also did the IMSTAT and it works well. imstat
----- task cl -----
0+4: call imstat
4+2: exec
----- exec external task imstatistics -----
imstatistics\nfields="image,npix,mean,stddev,min,max"\n
lower=INDEF\nupper=INDEF\nnclip=0\nlsigma=3.\nusigma=3.\n
binwidth=0.1\nformat=yes\ncache=no\nmode="ql"\n$nargs=0\n
--------------------------------
----- task imstatistics -----
8+4: inspect images
List of input images (scalib44): 24nos0063
12+2: end
# IMAGE NPIX MEAN STDDEV MIN MAX
24nos0063 1048576 312.6 994.1 -3623. 119323.
----- task imstatistics -----
8+3: call bye
11+2: exec
----- exec builtin bye -----
6+2: end
|
|
|
|
fitz |
04/03/2007 09:58PM
|
|
|
Status: offline
Registered: 09/30/2005
Posts: 4040
|
Trillian,Please send me the input image (fitz@iraf.net) and the IRAF and OS version you're using (if Mac, ppc or intel?). The problem is that in the IMARITH call the max value is an INDEF value but I can't tell exactly how it got that way. If you look at the script you can try reproducing the steps manually and check the intermediate images along the way.-Mike
|
|
|
|
fitz |
04/03/2007 09:58PM
|
|
|
Status: offline
Registered: 09/30/2005
Posts: 4040
|
Trillian,What platform and IRAF/OS version are you using? I don't see
any problems in the image or what's being done in the script (at
least the version on google). For example, on my linux system
I saved your file as trill.fits and was able to manually do what the
script did:ecl> imstat trill_s.fits
# IMAGE NPIX MEAN STDDEV MIN MAX
trill_s.fits 1048576 294.9 623.7 -97. 9490.
ecl>
ecl> gradient trill_s.fits trill_g.fits gradient="0" boundary=nearest
ecl> imstat trill_g.fits
# IMAGE NPIX MEAN STDDEV MIN MAX
trill_g.fits 1048576 0.0529 18.54 -1421. 1514.From your posted output the 'max' of the trill_g.fits file is being seen
as INDEF.You might trycl> unlearn boxcar gradient imstat imarith to reset the params of the affected tasks, otherwise are you able to reproduce
the above?-Mike
|
|
|
|
rohit |
04/03/2007 09:58PM
|
|
|
Status: offline
Registered: 06/01/2006
Posts: 86
|
Sorry, I am using RedHat Enterprise student edition. Forgot to mention that. It was bought in 2005. Did it create a .imh file for you when you ran mktracer? For example when I used the file 24nos0063.fits and ran the trace file, giving the output name 63trace, I got a file "63trace.imh". This does not happen when I run mktracer because of the errors. I will try what you suggested. Again, thanks for taking time and helping me! Appreciate that. Trillian
|
|
|
|
rohit |
04/03/2007 09:58PM
|
|
|
Status: offline
Registered: 06/01/2006
Posts: 86
|
This is the task I have been using. Just sending you the link so we are on the same page. http://www2.keck.hawaii.edu/inst/nirspec/wmkonspec/index.html
|
|
|
|
fitz |
04/03/2007 09:58PM
|
|
|
Status: offline
Registered: 09/30/2005
Posts: 4040
|
I was looking at the same script file. I did my test by hand and explicitly added the .fits extension (the script does as well), but even if I run the mktracer script, using an imtype of 'imh' or 'fits', it still works. Try ruling our parameter problems by doing a MKIRAF in /tmp and logging in from there and running it again. You might also copy the image to rule out any file corruption issues but from what I can tell it works just fine (at least by using the V2.13b2 release).
-Mike
|
|
|
|
rohit |
04/03/2007 09:58PM
|
|
|
Status: offline
Registered: 06/01/2006
Posts: 86
|
Hi Fitz, It worked..I ran the script and also reset IRAFThanks so much! Trillian
|
|
|
|