search

SIMPLE-AstroDB / SIMPLE-db

BSD 3-Clause "New" or "Revised" License
11 stars 22 forks source link

Tweak PhotometryFilters table #293

Closed kelle closed 1 year ago

kelle commented 2 years ago

Right now, we require Telescope and Instrument in the PhotometryFilters Table but I think that's redudant since those fields should be associated with a particular observation. For example, lots of different instrument/telescope combos use the SDSS filter set. I think we should remove those columns from the PhotometryFilters table and only require band.

kelle commented 2 years ago

Also seems like if we're gonna keep UCDs, they should be in the Photometry Filters table and not in the Photometry table.

dr-rodriguez commented 2 years ago

I agree with these two changes. I think at some point we did consider PhotometryFilters to not have instrument/telescope but there were some arguments at the time against it. This change would greatly simplify the tables and the foreign key validation that the database tries to do.

kelle commented 2 years ago

Ok, so, the official filter names from the SVO usually include the specific instrument name so there's not a lot of duplicate use of filter names. This tells me that using the UCD (e.g., em.opt.J) is useful so that you can run a query to get all the J-band data without needing to know all the various names for J band filters. (e.g, 2MASS.J, WFCAM.J, VISTA.J, UKIDSS.J).

Having instrument and telescope on both the Photometry and PhotometryFilters table is still redundant. But I'm not sure which table to keep it on.

kelle commented 2 years ago

We're leaning towards removing Telescope and Instrument from PhotometryFilters.

kelle commented 1 year ago

But we want to keep UCD (e.g., em.J) in the PhotometryFilters table so that we can generate J-K plots, for example, which include lots of different J filters (MKO.J and 2MASS.J).

Also want to keep effective_wavelength, but it doesn't need to be perfectly accurate.

Will-Cooper commented 1 year ago

The website uses effective wavelength to construct colours to ensure all are blue - red

if photfilters.at['effective_wavelength', bandtrue] >= \
                    photfilters.at['effective_wavelength', nextbandtrue]:  # if not blue-red
                continue
Will-Cooper commented 1 year ago

Can connect an approximate central wavelength in the PhotometryFilters table to allow this grouping of bands

kelle commented 1 year ago
kelle commented 1 year ago