#bezier_easing

[!IMPORTANT] This package has been renamed from bezier-easing to bezier_easing. The old package is now deprecated.

#Features

• Numerical strategy consistent with the JS version: Sample Table + Newton-Raphson + Binary Subdivision.
• Balanced performance and precision (11 samples, 4 Newton iterations).
• Exact endpoints guaranteed for x == 0 and x == 1.
• Fails immediately on invalid control point parameters (x1 or x2 not in [0,1]).

#Install

moon add cg-zhou/bezier_easing

#Quick Start

let easing = @bezier_easing.bezier(0.25, 0.1, 0.25, 1.0)
let y0 = easing(0.0) // 0.0
let y1 = easing(0.5)
let y2 = easing(1.0) // 1.0

#API

• bezier(x1 : Double, y1 : Double, x2 : Double, y2 : Double) -> (Double) -> Double

• bezier is the primary entry point, returning an easing function.

#Behavior & Constraints

• Requires control point X-coordinates to satisfy: 0 <= x1 <= 1 and 0 <= x2 <= 1.
• If requirements are not met, the function will fail immediately (aligned with JS throw behavior).
• No additional clamping is performed on the easing input x, however x == 0 and x == 1 return exact endpoints.
• Returns an identity function for linear curves (x1 == y1 && x2 == y2).

#Precision Strategy

1. Precompute 11 sample points for initial guess estimation.
2. Use Newton-Raphson for fast convergence when slope is sufficient.
3. Fall back to binary subdivision for stability when slope is too small.

• Newton iterations: 4
• Newton min slope: 0.001
• Subdivision precision: 1e-7
• Subdivision max iterations: 10

#Development

#Tests

moon test

#Benchmarks

moon bench

#Compare with JS

cd scripts && npm install && npm test

#License