The Alg Package

Oak provides oak/alg as a package to handle relatively miscellaneous math and selection algorithms. This package is a candidate for renaming in the future, or moving all of its contents to subpackages due to its miscellaneous nature. Right now alg itself contains several utilities and selection algorithms, while its two subpackages, intgeom and floatgeom respectively handle integer and floating point geometry data types.

Selection Algorithms

ChooseX and UniqueChooseX both operate on slices of float64 weights, returning a slice of indices of chosen elements from that weight set. ChooseX can choose the same index more than once, while UniqueChooseX cannot. In other words, ChooseX is with replacement and UniqueChooseX is without replacement.

There exists two smaller utilities, WeightedChooseOne and WeightedMapChoice, both of which operate on lists of cumulative weights. We call cumulative weights here a set of floats converted from their initial values so that each index represents the total sum of all indices before or after it in the original set, including that index. Both Weighted functions accept weights stylized as CumulativeWeights returns, with the latter operating on a map, if you happen to be using a map instead of a slice. $FUNC_Seeded variants allow for custom rng imeplmentations to be provided as well.

Math Helpers

F64eq and F64eqEps both equate floating points within input epsilon values, the former using a package local value. A user might find this useful if they're doing a lot of geometry with floating points.

RoundF64 rounds a float to an integer.

In addition, DegToRad and RadToDeg hold constant values that can be used to convert radians and degrees between one another.

Floatgeom and Intgeom

Both the alg/floatgeom and alg/intgeom packages provide Point and Rect types for working with floating point and integer geometry. Utilities on these types range from arithmetic to angle manipulation, to dot and cross products, and is continually expanding.