HealSparse Geometry¶

HealSparse has a basic geometry library that allows you to generate maps from circles and convex polygons, as supported by healpy. Each geometric object is associated with a single value. On construction, geometry objects only contain information about the shape, and they are only rendered onto a HEALPix grid when requested.

There are two methods to realize geometry objects. The first is that each object can be used to generate a HealSparseMap map, and the second, for integer-valued objects is the realize_geom() method which can be used to combine multiple objects by or-ing the integer values together.

HealSparse Geometry Shapes¶

The two shapes supported are Circle and Polygon. They share a base class, and while the instantiation is different, the operations are the same.

Circle

import healsparse

# All units are decimal degrees
circ = healsparse.Circle(ra=200.0, dec=0.0, radius=1.0, value=1)

Convext Polygon

# All units are decimal degrees
poly = healsparse.Polygon(ra=[200.0, 200.2, 200.3, 200.2, 200.1],
                          dec=[0.0, 0.1, 0.2, 0.25, 0.13],
                          value=8)

Making a Map¶

To make a map from a geometry object, use the get_map() method as such. The higher resolution you choose, the better the aliasing at the edges (given that these are pixelized approximations of the true shapes). You can also combine two maps using the general operations. Note that if the polygon is an integer value, the default sentinel when using get_map() is 0.

smap_poly = poly.get_map(nside_coverage=32, nside_sparse=32768, dtype=np.int16)
smap_circ = circ.get_map(nside_coverage=32, nside_sparse=32768, dtype=np.int16)

combo = healsparse.or_union([smap_poly, smap_circ])

Using `realize_geom()`¶

You can only use realize_geom() to create maps from combinations of polygons if you are using integer maps, and want to or them together. This method is more memory efficient than generating each individual individual map and combining them, as above.

realized_combo = healsparse.HealSparseMap.make_empty(32, 32768, np.int16, sentinel=0)
healsparse.realize_geom([poly, circ], realized_combo)

HealSparse Geometry¶