Legacy Survey invvar?

Question

Hey all, 

The legacy survey has all their information about their files here: http://legacysurvey.org/dr7/files/

This page lists fits files such a chi2 fits, model fits, galdepth, and nexp, which do show up in the svc search in a jupyter notebook.

However some files, such as the invvar or the maskbits, do not show up (and additionally the tables here do not match up the ones in the schema, such as a galaxy table showing up on NOAO but not on their website). Does anyone know why this information isn't readily available?

Thank you,

Derrick

Answer 1

Hi Derrick, thank you for getting in touch, and for your patience. The "galaxy" and "star" tables are views of the tractor table for objects that respectively do not (or do) have type=PSF, with the rough assumption that PSF light profiles accurately select stars. In details, this is not a perfect sample separation so we will likely no longer generate these views. Regarding the first part of the question, we attempted to serve the data products for which we expected to be a demand by our users. We are open to suggestions and requests, which we consider as a team before ingesting additional datasets or data products. Could you please let us know which of the data products you would need for your work? It would be even more helpful for our assessment if you could possibly provide an example use case for the data products.

Comments

Hey,

So sorry for not getting back to you until much later, I got side-tracked by another project but am finally getting back to this. Your question helps a lot.

So I want to be able to take images to run in a light profile fitting software in order to get deconvolved radii and to better understand the structural parameters of over 2000 objects, so having an automated process is a necessity. My goal was to use the legacy survey available on datalabs to do so. With that said, I would need exactly 4 fits per object to do this work

1) The image fits
2) a segmentation map fits
3) a sigma fits
4) a PSF file (should just be of one isotropic source in the FOV)

With that said, I may have to make 2) myself using source extractor.

3) is a necessary file that doesn't seem to be readily available, which should also definitely be an output in some sort of reduction process. I think I might be able to derive it using the galdepth file, where each pixel is presumably a flux/sqrt(invvar), from which invvar = sigma^2. Does that seem like a safe way for me to get to what I need it to get to?

4) I could probably just make a fits of a PSF in the FOV myself, but would there be a simple way to do that in your jupyter notebooks by chance?

Hopefully this helps, since this question is old, if it is unlikely for me to get an answer could you let me know? Just so I am not waiting too long for something that most likely will not happen. Again, thanks S. Juneau