Introduction

VAST.VoidFinder is a software package containing a Python 3 implementation of the VoidFinder algorithm (El-Ad & Piran, 1997) that is based on the algorithm’s Fortran implementation by Hoyle & Vogeley (2002). Motivated by the expectation that voids are spherical to first order, this algorithm defines voids as the unions of sets of spheres grown in the underdense regions of the large-scale structure.

The VoidFinder directory contains the package, which includes an efficient Multi-Process Cythonized version of VoidFinder. Options are available to identify voids in both observational galaxy surveys and periodic cosmological simulations. To import the main void-finding function of VoidFinder:

from vast.voidfinder import find_voids

VoidFinder begins by removing all isolated tracers from a catalog of objects. The remaining tracers are then placed on a grid, and spheres are grown from the centers of the empty cells until they are bounded by four tracers on their surfaces.

All spheres larger than a specified radius (typically around 10 Mpc/h) are considered possible maximal spheres – the largest sphere that can fit in a given void region. Filtering through these candidate maximal spheres by order of decreasing radius, no maximal sphere can overlap by more than 10% of its volume with any other previously identified (larger) maximal sphere. After the maximal spheres are identified, the remaining holes are combined with these maximal spheres to enhance the void structure if they overlap exactly one maximal sphere by at least 50% of its volume. The union of a set of spheres (one maximal and the remaining smaller holes) defines a void region.

Citation

Please cite Hoyle & Vogeley (2002) and El-Ad & Piran (1997) when using this algorithm.