msccmatch problem

Josh Walawender wrote on Jun 02, 2006

Hi,

I'm running into a problem using msccmatch in mscred. I'm running it in non-interactive mode on a long list of object files. From time to time it stops with an error (see below). When restarted (after removing the sucessfully completed files from the input list), it usually continues happily for a while before stopping with an error again. In this way I am able to slowly work my way through the list of files, but it takes much longer with all the restarts. Do you know what might be causing this error?

I've pasted the output of two sucessive runs of msccmatch below. The first failed on obj402.fits, but the subsequent run suceeded.

thanks,
Josh

ms> msccmatch @list_obj2 coords="!mscgetcat $I $C magmax=20" nsearch=60 search=40 rsearch=1.0 nfit=200 rms=1.0 maxshift=5 verbose+ listcoords- reject=3 int-
MSCCMATCH:
  obj401.fits:
    2214 input coordinates
    1170/2214 coordinates out of bounds
    150/1044 coordinates masked
    search using up to 60 objects:
    search found offsets of (1, 1) pixels and rotation 0.01 degrees
    2214 input coordinates
    1170/2214 coordinates out of bounds
    95/1044 coordinates masked
    5/949 not centroided
    Fit coordinates:
      input number of coordinates = 944
      tangent point shift = (-0.01, 0.01) arcsec
      fractional scale change = (1.000, 1.000)
      axis rotation = (-0.001, 0.002) degrees
      rms = (0.621, 0.655) arcsec
    Coordinate system updated.
  obj402.fits:
ERROR: may not convert string to other types
    msccmatch (input=@list_obj2, coords=!mscgetcat $I $C magmax=20, verbose=yes, nsearch=60, search=40., rsearch=1., maxshift=5., listcoords=no, nfit=200, rms=1., reject=3., interactive=no)
    msccmatch (input=@list_obj2, coords=!mscgetcat $I $C magmax=20, verbose=yes, nsearch=60, search=40., rsearch=1., maxshift=5., listcoords=no, nfit=200, rms=1., reject=3., interactive=no)
ms> flpr
ms> flpr
ms> msccmatch @list_obj2 coords="!mscgetcat $I $C magmax=20" nsearch=60 search=40 rsearch=1.0 nfit=200 rms=1.0 maxshift=5 verbose+ listcoords- reject=3 int-
MSCCMATCH:
  obj402.fits:
    1895 input coordinates
    1002/1895 coordinates out of bounds
    123/893 coordinates masked
    search using up to 60 objects:
    search found offsets of (135, 3) pixels and rotation 0.14 degrees
    1895 input coordinates
    1013/1895 coordinates out of bounds
    151/882 coordinates masked
    search using up to 60 objects:
    search found offsets of (8, -1) pixels and rotation 0.12 degrees
    1895 input coordinates
    1010/1895 coordinates out of bounds
    119/885 coordinates masked
    7/766 not centroided
    Fit coordinates:
      input number of coordinates = 759
      tangent point shift = (0.28, 37.66) arcsec
      fractional scale change = (0.999, 0.999)
      axis rotation = (0.249, 0.276) degrees
      rms = (0.670, 0.703) arcsec
    Coordinate system updated.

---

Francisco Valdes wrote on Jun 02, 2006

Hi Josh,

I can't tell what is wrong without reproducing this. My guess is that something is happening when getting the coordinates by executing a command. What might work better for you is to turn you single command over a list of files into a script loop where in each loop one image is done at a time. In this loop you could call mscgetcat separately and pass the file to msccmatch or may just doing
what you are doing one exposure at a time with a flpr would be a convenient workaround. An example of what the loop would look like is

list = "list_obj2"
while (fscan (list, s1) != EOF) {
    msccmatch (s1, coords="!mscgetcat $I $C magmax=20", nsearch=60,
       search=40, rsearch=1.0, nfit=200, rms=1.0, maxshift=5, verbose+, 
       listcoords-, reject=3, int-)
    flpr
}
list = ""

Frank Valdes

---

Last post on Jun 02, 2006