Cutting pixelized sky masks in a pipeline way

Skykatana is a pacakge to create and maniputate boolean spatial masks on the celestial sphere, by combining healsparse pixel maps accounting for various effects such as cutting out regions around bright stars, low depth, bad seeing, extended sources, among others. We call these partial maps stages, which are then combined into a final mask.

For each stage you can generate random points, quickly visualize masks, do plots overlaying bright stars, and apply the mask to an arbitrary catalog to select sources located inside.

It has been designed to produce masks for large 8-meter surveys such as the upcoming half-sky dataset of the Vera Rubin Observatory and the HSC-SSP survey. It can handle multi-billion pixel masks with very limited memory resources and is flexible to accomodate custom recipes for masking different objects.

Main Class

• SkyMaskPipe() Main class for assembling and handling pixelized masks

Main Methods

• build_footprint_mask(), build_circ_mask(), buld_propmap_mask(), build_star_mask_online(), etc --> Generate maps for each stage from discrete sources, geometric shapes or other healsparse maps
• combine() --> Merge the maps created above to generate a new mask
• plot() --> Visualize a mask stage by plotting randoms. Options to zoom, oveplot stars, etc.
• plot_moc() --> Visualize a mask stage by plotting its MOC (multiorder coverage map).
• makerans() --> Generate randoms over a mask stage
• apply() --> Cut out sources outside of a given mask stage

Dependencies

• lsdb, healsparse, tqdm, healpy, fitsio, ipyaladin, pillow

Install

There are two ways to get skykatana:

• pip install skykatana

or

• Clone the repo, switch to the pacakge directory and do pip install .  This has the advantage that you will get the latest version and example notebooks.

Example Dataset

There a small dataset of ~8 million HSC sources to start using the package. Get it here (170 MB) and decompress it. Then, adjust the folder location in the provided notebooks and just run them.

Documentation

• A quick tutorial notebook with HSC data is available here
• A tutorial notebook for building Rubin masks can be found here
• The full documentation and API is available here

Gallery

Credits

• Main author: Emilio Donoso
• Contributors: Mariano Dominguez, Claudio Lopez, Konstantin Malanchev

Acknowledgements

This software was partially developed with the generous support of the LINCC Frameworks Incubator Program using LINCC resources. The healsparse code was written by Eli Rykoff and Javier Sanchez. mocpy is a fantastic library developed by the mocpy team, and ipyaladin is a great tool to enable interactive sky visualizations.