marvin
marvin copied to clipboard
Access SDSS photometry and single fiber spectroscopy via Marvin
Accessing SDSS Legacy Survey photometry and single fiber spectroscopy will be a common task. Can we make this doable in Marvin?
I have thought a bit about this one. It's not totally trivial to implement but doable. Here are some thoughts.
For the spectroscopy, that can be retrieved for each galaxy by downloading the spec file, which are created on a fibre-by-fibre basis.
- I think all the information needed to create the URL path is in the NSA entry but not in the drpall one, which is a bit of a complication, but not a bit one. I think we'd want the lite version of the file.
- The download can happen as a stream to a
StringIO
object so that we don't really need to download the file to the local SAS. I don't think we need to download this from theebosswork
version of the SAS. We can use the latest DR and save us a lot of trouble with authentication and stuff. - To contain the downloaded spectrum, we can create a thin wrapper around
astropy.io.fits.HDUList
. It would be basically that plus a method to convert to a Marvin Spectrum (one could make a Spectrum out of the flux or the model), and another one to save it to a file or to the local SAS. - There is information about the line fitting in the file but parsing it may be beyond what we want to do here.
For the imaging, I think we want to implement it as described in #401, i.e., as a link/mixin to the MaNGA preimaging and not to the SDSS Legacy imaging. The preimaging is created from the NSA imaging which, I think, does a better work for galaxies. Regarding photometry values, those are included in the NSA information we provide. Are there other values we want to include?
Given that the effort to accomplish this is not totally trivial, I'm leaning towards trying to implement this after the the DR15 freeze but, if possible, before we release the DR15 version. This is a small, well isolated change so I think it fall in the kind of development that we don't need to wait for another DR to release.
All that said, something that would be fairly easy to do is to add a link to the web pointing to the location of the spec file or, better, to the SAW.