orx/orx-hash-grid/README.md

orx-hash-grid

2D space partitioning for fast point queries.

Usage

orx-hash-grid provides the classes HashGrid and Cell, in most cases only HashGrid is used.

Create a hash grid for a given radius.

val grid = HashGrid(radius)

Check for a given query point if the grid is free, i.e. there is no point in the grid at distance less than radius away from the query point.

grid.isFree(query)

Add a point to the hash grid structure:

grid.insert(point)

Iterate over all points in the hash grid:

for (point in grid.points()) {
    // do something with point
}

Extensions to standard library

orx-hash-grid provides short-hand extension functions to List<Vector2>

---

fun List<Vector2>.filter(radius: Double) : List<Vector2>

filters the points in the list such that only points with an inter-distance of radius remain.

val points = (0 until 10_000).map { drawer.bounds.uniform() }
val filtered = points.filter(20.0)

---

fun List<Vector2>.hashGrid(radius: Double) : HashGrid

constructs a (mutable) HashGrid containing all points in the list.

val points = (0 until 10_000).map { drawer.bounds.uniform() }
val hashGrid = points.hashGrid(20.0)

---

References

• orx-noise uses HashGrid to generate Poisson distributed points. Link

Demos

DemoFilter01

The program performs the following steps:

• Generates 10,000 random points uniformly distributed within the drawable bounds.
• Filters the generated points to enforce a minimum distance of 20.0 units between them.
• Visualizes the filtered points as circles with a radius of 10.0 units on the canvas.

The filter method is provided by orx-hash-grid.

source code

DemoFilter3D01

Demonstrates how to use a 3D hash-grid filter operation to remove points from a random 3D point-collection that are too close to each other. The resulting points are displayed as small spheres.

The program performs the following key steps:

• Generates 10,000 random 3D points located between a minimum and maximum radius.
• Filters the points to ensure a minimum distance between any two points using a spatial hash grid.
• Creates a small sphere mesh that will be instanced for each filtered point.
• Sets up an orbital camera to allow viewing the 3D scene interactively.
• Renders the filtered points by translating the sphere mesh to each point's position and applying a shader that modifies the fragment color based on the view normal.

source code

DemoHashGrid01

This demo creates a HashGrid to manage points in a 2D space. Notice the desired cell size in the HashGrid constructor.

On every animation frame, it attempts to insert 100 random points into the HashGrid. When a HashGrid cell is free, a point is inserted.

The visual output includes:

• Rectangles representing the bounds of the occupied cells in the grid.
• Circles representing the generated random points.

source code