Would we need to add HISTORY keywords to AstroHeader?

[agioffe]

Would we consider it useful to add some HISTORY cards to AstroHeaders and also write them into the FITS output (probably, even in another DB column in some cases)? Keeping some processing parameters in the HISTORY records could be useful both within the regular pipeline processing (though it could be coded as pipeline version) but especially for some post-processing and non-standard tasks (e.g., adding objects, merging catalogs, etc.).

Do I take it right that we could just add HISTORY to ~/matlab/AstroPack/config/header.yml or ~/matlab/AstroPack/config/Header.Comments.Default.yml in order to be able to write such a field in an AstroHeader? One issue could be that in the FITS standard one can have many cards with the same HISTORY keyword, but I don't know if we can do the same with AstroHeader.

@EranOfek @EastEriq , what do you think?

[EastEriq]

No idea, on my part. What kind of information should this key contain? If it is not something related to the first step of creation of raw images, I'm not really into it.

[agioffe]

Usually such cards contain information on how the current image (or spectrum, or catalog) has been produced from some raw data. In particular, we discussed some time ago that we need to make a mark if the catalog of an AI object was augmented with another catalog of objects when their images were added ("injected") to the image contained in this AI.

P.S. Here is a short example of how such keywords are written to a FITS header by some Chandra tools:

HISTORY dmextract binning: File EVENTS Col pi Binning 0.5:1024.5:1 HISTORY TOOL :dmextract 2024-04-30T09:27:07 ASC00001 HISTORY PARM :infile=29384r/acisf29384_repro_evt2.fits[sky=region(SN24ASC00002 HISTORY CONT :ggi_ep2_0001_srcreg.fits)][bin PI] ASC00003 HISTORY PARM :outfile=SN24ggi_ep2_0001.pi ASC00004 HISTORY PARM :bkg= ASC00005 HISTORY PARM :error=gaussian ASC00006 HISTORY PARM :bkgerror=gaussian ASC00007 HISTORY PARM :bkgnorm=1 ASC00008 ... HISTORY TOOL :dmgroup 2024-04-30T09:27:14 ASC00037 HISTORY PARM :infile=SN24ggi_ep2_0001.pi[SPECTRUM] ASC00038 HISTORY PARM :outfile=SN24ggi_ep2_0001_grp.pi ASC00039 HISTORY PARM :grouptype=NUM_CTS ASC00040 HISTORY PARM :grouptypeval=15 ASC00041

[EastEriq]

I see. Since raw image taking is, in my view, a single pass operation, I think I should not be concerned with that in the code I mainly maintain, but please correct me if you see the need. OTOH yes, here and there there are procedure calls which allow some parameters (e.g. the steps of the focus loop), and we have empirically settled for some given value, but if we were to record all of them as HISTORY... Skeptical

[EranOfek]

no need at this stage. This is rarely useful.

On Thu, Jun 13, 2024 at 2:07 PM Alexander Krassilchtchikov < @.***> wrote:

Would we consider it useful to add some HISTORY cards to AstroHeaders and also write them into the FITS output (probably, even in another DB column in some cases)? Keeping some processing parameters in the HISTORY records could be useful both within the regular pipeline processing (though it could be coded as pipeline version) but especially for some post-processing and non-standard tasks (e.g., adding objects, merging catalogs, etc.).

Do I take it right that we could just add HISTORY to ~/matlab/AstroPack/config/header.yml or ~/matlab/AstroPack/config/Header.Comments.Default.yml in order to be able to write such a field in an AstroHeader? One issue could be that in the FITS standard one can have many cards with the same HISTORY keyword, but I don't know if we can do the same with AstroHeader.

@EranOfek https://github.com/EranOfek @EastEriq https://github.com/EastEriq , what do you think?

— Reply to this email directly, view it on GitHub https://github.com/EranOfek/AstroPack/issues/457, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABJUQ4JPK7DWKSHGYQZWJK3ZHF4PVAVCNFSM6AAAAABJIFGC3GVHI2DSMVQWIX3LMV43ASLTON2WKOZSGM2TAOBVG44TAMI . You are receiving this because you were mentioned.Message ID: @.***>