com.wallstop-studios.unity-helpers

A Grab-Bag

Various Unity Helpers. Includes many deterministic, seedable random number generators.

Compatibility

Platform	Compatible
Unity 2021	Likely, but untested
Unity 2022	✓
Unity 2023	✓
Unity 6	✓
URP	✓
HDRP	✓

Installation

To Install as Unity Package

1. Open Unity Package Manager
2. (Optional) Enable Pre-release packages to get the latest, cutting-edge builds
3. Open the Advanced Package Settings
4. Add an entry for a new "Scoped Registry"
   • Name: NPM
   • URL: https://registry.npmjs.org
   • Scope(s): com.wallstop-studios.unity-helpers
5. Resolve the latest com.wallstop-studios.unity-helpers

From Source

Grab a copy of this repo (either git clone or download a zip of the source) and copy the contents to your project's Assets folder.

From Releases

Check out the latest Releases to grab the Unity Package and import to your project.

Package Contents

• Random Number Generators
• Spatial Trees
• Protobuf, Binary, and JSON formatting
• A resizable CyclicBuffer
• Simple single-threaded thread pool
• A LayeredImage for use with Unity's UI Toolkit
• Geometry Helpers
• Child/Parent/Sibling Attributes to automatically get components
• ReadOnly attribute to disable editing of serialized properties in the inspector
• An extensive collection of helpers
• Simple math functions including a generic Range
• Common buffers to reduce allocations
• A RuntimeSingleton implementation for automatic creation/accessing of singletons
• String helpers, like converting to PascalCase, like Unity does for variable names in the inspector
• A randomizable PerlinNoise implementation
• And more!

Random Number Generators

This package implements several high quality, seedable, and serializable random number generators. The best one (currently) is the PCG Random. This has been hidden behind the PRNG.Instance class, which is thread-safe. As the package evolves, the exact implementation of this may change.

All of these implement a custom IRandom interface with a significantly richer suite of methods than the standard .Net and Unity randoms offer.

To use:

IRandom random = PRNG.Instance;

random.NextFloat(); // Something between 0.0f and 1.0f
random.NextDouble(); // Something between 0.0d and 1.0d
random.Next(); // Something between 0 and int.MaxValue
random.NextUint(); // Something between 0 and uint.MaxValue

int [] values = {1, 2, 3};
random.NextOf(values); // 1, 2, or 3
random.NextOf(Enumerable.Range(0, 3)); // 1, 2, or 3
HashSet<int> setValues = new() {1, 2, 3};
random.NextOf(setValues); // 1, 2, or 3

random.NextGuid(); // A valid UUIDv4
random.NextGaussian(); // A value sampled from a gaussian curve around mean=0, stdDev=1 (configurable via parameters)
random.NextEnum<T>(); // A randomly selected enum of type T

int width = 100;
int height = 100;
random.NextNoiseMap(width, height); // A configurable noise map generated using random octave offsets

Implemented Random Number Generators

• PCG
• DotNet (uses the currently implemented .Net Random)
• RomoDuo
• SplitMix64
• Squirrel
• System (uses a port of the Windows .Net Random)
• Unity (uses Unity's random under the hood)
• Wy
• XorShift
• XorShiro

Spatial Trees

There are three implemented 2D immutable spatial trees that can store generic objects, as long as there is some resolution function that can convert them into Vector2 spatial positions.

• QuadTree (easiest to use)
• KDTree
• RTree

Spatial trees, after construction, allow for O(log(n)) spatial query time instead of O(n). They are extremely useful if you need repeated spatial queries, or if you have relatively static spatial data.

Usage

GameObject [] spatialStorage = { myCoolGameObject };
QuadTree<GameObject> quadTree = new(spatialStorage, go => go.transform.position);

// Might return your object, might not
GameObject [] inBounds = quadTree.GetElementsInBounds(new Bounds(0, 0, 100, 100));

// Uses a "good-enough" nearest-neighbor approximately for cheap neighbors
List<GameObject> nearestNeighbors = new();
quadTree.GetApproximateNearestNeighbors(myCoolGameObject.transform.position, 1, nearestNeighbors);
Assert.AreEqual(1, nearestNeighbors.Count);
Assert.AreEqual(myCoolGameObject, nearestNeighbors[0]);

Note

All spatial trees expect the positional data to be immutable. It is very important that the positions do not change. If they do, you will need to reconstruct the tree.