Overview

The hpmcm package collects a set of tools to do catalog cross matching using the following strategy.

1. The match region is mapped onto a single WCS and each sources is projected into pixel coordinates in that WCS. This is defined using a sub-class of hpmcm.match.Match

2. The match region is sub-divided into cells, which include some overlap region between them.

3. For each cell, a “counts map” of sources in that cell is produced, counting the number of sources in each pixel of the projected WCS.

4. A clustering algorithm is run to extract the “footprints” of sets of adjacent pixels that contain sources. Those are then used to create a set of clusters

5. Each cluster in then refined into objects, by repeating the counts map and footprint detection using progressive smaller pixels until each objects consists only of sources within the requested match radius. At each iteration the pixel scale is halved.

6. Information about the source-to-object and source-to-clusters associations, as well information about the resulting objects and clusters are collected into hpmcm.output_tables

The two sub-classes of Match are:

1. WcsMatch allows the user to define the WCS used for the projection and takes any number of input catalogs.

2. ShearMatch uses a predefined WCS that was used to create the meta-detection ShearObject catalogs, and expects exactly 5 input catalogs: a reference catalog and 4 counterfactual shear catalogs.