On Mon, 15 Sep 2014, Tim Lister wrote:
> On 09/12/2014 01:18 PM, Peter W. Draper wrote:
>> I think this is a merge bug (code missed out that initialised a
>> structure for output arrays). I've put that back so I'm afraid you need
>> to update the submodules again, so pull the main repository and
>> ./update-modules from the top, followed by a "make clean && make && make
>> install" in the applications/extractor directory....
>
> Hi Peter, sorry for the slow reply as I spent a frustrating Friday trying to
> make starjava compile on Fedora 19 so I could view the resulting FITS_LDAC
> catalog(ue)s.
>
> The good news is the latest patched version of Extractor now happily runs and
> generates a source catalog. Unfortunately the bad news is that resulting
> FITS_LDAC format catalog still causes SCAMP to barf:
> ----- SCAMP 2.0.4 started on 2014-09-15 at 14:57:16 with 2 threads
>
> ----- 1 inputs:
>
>> *Error*: Corrupted FITS header in lsc1m005-fl06-20140903-0077_cat.ldac
Seems scamp relies on a weird table with a single row that has all the
FITS headers of the input image in a single cell. Looked at that during
the merge and decided it was probably not useful. Anyway I've put that
functionality back so try again.
Cheers,
Peter.
> This seems to be a problem with the extractor-generated catalog from the NDF;
> if I convert the NDF to FITS and use the star/extractor/sex binary instead,
> then the resulting FITS_LDAC catalog is happily accepted by SCAMP and it
> solves.
>
> I used fitsverify -l to dump the headers of both versions of the FITS_LDAC
> catalog. The difference seems to be mostly in HDU2 (_v2.ldac came from the
> star/extractor/sex binary run on the FITS, the other one came from extractor
> run on the NDF):
> diff lsc1m005-fl06-20140903-0077_cat_v2.ldac.hdr
> lsc1m005-fl06-20140903-0077_cat.ldac.hdr
> 6c6
> < File: lsc1m005-fl06-20140903-0077_cat_v2.ldac
> ---
>> File: lsc1m005-fl06-20140903-0077_cat.ldac
> 32c32
> < 4 | NAXIS1 = 78240 / BYTES PER ROW
> ---
>> 4 | NAXIS1 = 3760 / BYTES PER ROW
> 39,40c39,40
> < 11 | TFORM1 = '78240A '
> < 12 | TDIM1 = '(80, 978)'
> ---
>> 11 | TFORM1 = '3760A '
>> 12 | TDIM1 = '(80, 47)'
> 48c48
> < 1 Field Header Card 78240A
> ---
>> 1 Field Header Card 3760A
>
> I'm not sure if something in the header is not propagating from the NDF to
> the FITS_LDAC and it's stopping early ?
>
> I've uploaded the catalogs and headers from both versions to the same LCOGT
> frames dropbox link as before.
>
>>
>> Cheers,
>>
>> Peter.
>>
>>
>
> Thanks,
> Tim
>
> --
> Dr Tim Lister
> Astronomer and Project Scientist
> Las Cumbres Observatory (LCOGT.net)
> +1 805 880 1989
>
--
Peter W. Draper, http://astro.dur.ac.uk/~pdraper
|