ihar-shalouski/swift-vector-and-matrix

Features

• Vector2<T> — a generic 2D vector:

- Arithmetic (+, -, multiply by scalar, dot product via dot, 2D pseudo‑cross via * returning a scalar). - Length / squared length, normalization, rotation, perpendicular vector perp, angle, projection/rejection, lerp. - Convenient subscript (v[0], v[1]), swapped, withX/withY, ExpressibleByArrayLiteral. - Interop with CGPoint (CoreGraphics) and SIMD2 (simd) when those modules are available.

• Matrix2<T> — a compact, dynamically‑sized matrix (historical name):

- Initializers from (rows, cols, elements) and from rows [[T]]. - Safe indexing, extracting rows/cols, submatrices and minors, transpose/in‑place transpose. - Matrix arithmetic (+, -, scalar multiply, matrix×matrix multiply). - Determinant (fast paths for 1×1 and 2×2; otherwise expansion by first row). - Identity matrix factory and isIdentity (for Numeric & Equatable).

• Modern protocols: Codable, Hashable, Sendable, Equatable.
• Safety first: strict precondition checks for dimensions and indices.
• Performance: row‑major storage, reserved capacity, @inlinable in hot paths.

✅ The current codebase is fully covered by unit tests.

Installation (Swift Package Manager)

Xcode

File → Add Packages… and paste this repo URL.

Import in code:

import SwiftVectors

Quick start

Vector2

// Doubles
var a = Vector2<Double>(x: 3, y: 4)
let b: Vector2<Double> = [1, 2]   // ExpressibleByArrayLiteral

let sum   = a + b                  // Vector2(4, 6)
let diff  = a - b                  // Vector2(2, 2)
let dot   = a.dot(b)               // 11
let cross = a * b                  // 2D pseudo-cross: 3*2 - 4*1 = 2

let len   = a.length               // 5
let unit  = a.normalized()         // Vector2(0.6, 0.8)

a.rotate(by: .pi / 2)              // +90° (radians)

let proj = a.project(on: b)        // projection of a onto b
let rej  = a.reject(from: b)       // orthogonal component

Matrix2 (dynamically sized)

let A = Matrix2(rows: 2, cols: 3, elements: [1, 2, 3,
                                             4, 5, 6])

let B = Matrix2([[7, 8],
                 [9, 10],
                 [11, 12]])        // init from rows

let C = A * B                       // 2×3 · 3×2 = 2×2

let I = Matrix2<Int>.identity(3)    // 3×3
let AT = A.transposed()             // transpose

let D = Matrix2([[1.0, 2.0],
                 [3.0, 4.0]])
let det = D.determinant             // -2.0
let minor = D.minor(ofRow: 0, col: 1)

Type constraints & notes

• Most operations require T: Numeric. Division, normalization, rotations, etc. require T: BinaryFloatingPoint.
• Matrix2.identity(_:) and isIdentity require T: Numeric & Equatable.
• Determinant for n > 2 uses expansion by first row (fine for small n; for larger matrices consider LU in the roadmap).

Testing

Run tests with:

swift test

All public APIs and edge cases are covered: arithmetic, indexing bounds, determinants, transposition, projections, etc.

Roadmap

• Vector3, Vector4, Matrix3x3, Matrix4x4, and a general MatrixN (fixed and dynamic sizes).
• Inverse, trace, rank; LU/QR/Cholesky; faster determinant via LU.
• Linear system solvers (Gaussian elimination, LU), pseudo‑inverse (SVD), eigenvalues/eigenvectors.
• Performance: SIMD‑accelerated kernels, custom memory layouts, micro‑benchmarks (swift-benchmark).
• Protocols (AdditiveArithmetic, VectorArithmetic) and utilities (row/column slices).
• DocC documentation, richer examples, playgrounds; CI (GitHub Actions); coverage reports.
• SwiftUI demos and small visual inspectors.

Contributing

PRs and issues are welcome! Please include tests and a short motivation for changes.

License

MIT — see LICENSE.

Package Metadata

Repository: ihar-shalouski/swift-vector-and-matrix

Default branch: main

README: README.md