1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
// Copyright 2018 the Kurbo Authors
// SPDX-License-Identifier: Apache-2.0 OR MIT

//! A trait for curves parametrized by a scalar.

use core::ops::Range;

use arrayvec::ArrayVec;

use crate::{common, Point, Rect};

#[cfg(not(feature = "std"))]
use crate::common::FloatFuncs;

/// A default value for methods that take an 'accuracy' argument.
///
/// This value is intended to be suitable for general-purpose use, such as
/// 2d graphics.
pub const DEFAULT_ACCURACY: f64 = 1e-6;

/// A curve parametrized by a scalar.
///
/// If the result is interpreted as a point, this represents a curve.
/// But the result can be interpreted as a vector as well.
pub trait ParamCurve: Sized {
    /// Evaluate the curve at parameter `t`.
    ///
    /// Generally `t` is in the range [0..1].
    fn eval(&self, t: f64) -> Point;

    /// Get a subsegment of the curve for the given parameter range.
    fn subsegment(&self, range: Range<f64>) -> Self;

    /// Subdivide into (roughly) halves.
    #[inline]
    fn subdivide(&self) -> (Self, Self) {
        (self.subsegment(0.0..0.5), self.subsegment(0.5..1.0))
    }

    /// The start point.
    fn start(&self) -> Point {
        self.eval(0.0)
    }

    /// The end point.
    fn end(&self) -> Point {
        self.eval(1.0)
    }
}

// TODO: I might not want to have separate traits for all these.

/// A differentiable parametrized curve.
pub trait ParamCurveDeriv {
    /// The parametric curve obtained by taking the derivative of this one.
    type DerivResult: ParamCurve;

    /// The derivative of the curve.
    ///
    /// Note that the type of the return value is somewhat inaccurate, as
    /// the derivative of a curve (mapping of param to point) is a mapping
    /// of param to vector. We choose to accept this rather than have a
    /// more complex type scheme.
    fn deriv(&self) -> Self::DerivResult;

    /// Estimate arclength using Gaussian quadrature.
    ///
    /// The coefficients are assumed to cover the range (-1..1), which is
    /// traditional.
    #[inline]
    fn gauss_arclen(&self, coeffs: &[(f64, f64)]) -> f64 {
        let d = self.deriv();
        coeffs
            .iter()
            .map(|(wi, xi)| wi * d.eval(0.5 * (xi + 1.0)).to_vec2().hypot())
            .sum::<f64>()
            * 0.5
    }
}

/// A parametrized curve that can have its arc length measured.
pub trait ParamCurveArclen: ParamCurve {
    /// The arc length of the curve.
    ///
    /// The result is accurate to the given accuracy (subject to
    /// roundoff errors for ridiculously low values). Compute time
    /// may vary with accuracy, if the curve needs to be subdivided.
    fn arclen(&self, accuracy: f64) -> f64;

    /// Solve for the parameter that has the given arc length from the start.
    ///
    /// This implementation uses the IPT method, as provided by
    /// [`common::solve_itp`]. This is as robust as bisection but
    /// typically converges faster. In addition, the method takes
    /// care to compute arc lengths of increasingly smaller segments
    /// of the curve, as that is likely faster than repeatedly
    /// computing the arc length of the segment starting at t=0.
    fn inv_arclen(&self, arclen: f64, accuracy: f64) -> f64 {
        if arclen <= 0.0 {
            return 0.0;
        }
        let total_arclen = self.arclen(accuracy);
        if arclen >= total_arclen {
            return 1.0;
        }
        let mut t_last = 0.0;
        let mut arclen_last = 0.0;
        let epsilon = accuracy / total_arclen;
        let n = 1.0 - epsilon.log2().ceil().min(0.0);
        let inner_accuracy = accuracy / n;
        let f = |t: f64| {
            let (range, dir) = if t > t_last {
                (t_last..t, 1.0)
            } else {
                (t..t_last, -1.0)
            };
            let arc = self.subsegment(range).arclen(inner_accuracy);
            arclen_last += arc * dir;
            t_last = t;
            arclen_last - arclen
        };
        common::solve_itp(f, 0.0, 1.0, epsilon, 1, 0.2, -arclen, total_arclen - arclen)
    }
}

/// A parametrized curve that can have its signed area measured.
pub trait ParamCurveArea {
    /// Compute the signed area under the curve.
    ///
    /// For a closed path, the signed area of the path is the sum of signed
    /// areas of the segments. This is a variant of the "shoelace formula."
    /// See:
    /// <https://github.com/Pomax/bezierinfo/issues/44> and
    /// <http://ich.deanmcnamee.com/graphics/2016/03/30/CurveArea.html>
    ///
    /// This can be computed exactly for Béziers thanks to Green's theorem,
    /// and also for simple curves such as circular arcs. For more exotic
    /// curves, it's probably best to subdivide to cubics. We leave that
    /// to the caller, which is why we don't give an accuracy param here.
    fn signed_area(&self) -> f64;
}

/// The nearest position on a curve to some point.
///
/// This is returned by [`ParamCurveNearest::nearest`]
#[derive(Debug, Clone, Copy)]
pub struct Nearest {
    /// The square of the distance from the nearest position on the curve
    /// to the given point.
    pub distance_sq: f64,
    /// The position on the curve of the nearest point, as a parameter.
    ///
    /// To resolve this to a [`Point`], use [`ParamCurve::eval`].
    pub t: f64,
}

/// A parametrized curve that reports the nearest point.
pub trait ParamCurveNearest {
    /// Find the position on the curve that is nearest to the given point.
    ///
    /// This returns a [`Nearest`] struct that contains information about
    /// the position.
    fn nearest(&self, p: Point, accuracy: f64) -> Nearest;
}

/// A parametrized curve that reports its curvature.
pub trait ParamCurveCurvature: ParamCurveDeriv
where
    Self::DerivResult: ParamCurveDeriv,
{
    /// Compute the signed curvature at parameter `t`.
    #[inline]
    fn curvature(&self, t: f64) -> f64 {
        let deriv = self.deriv();
        let deriv2 = deriv.deriv();
        let d = deriv.eval(t).to_vec2();
        let d2 = deriv2.eval(t).to_vec2();
        // TODO: What's the convention for sign? I think it should match signed
        // area - a positive area curve should have positive curvature.
        d2.cross(d) * d.hypot2().powf(-1.5)
    }
}

/// The maximum number of extrema that can be reported in the `ParamCurveExtrema` trait.
///
/// This is 4 to support cubic Béziers. If other curves are used, they should be
/// subdivided to limit the number of extrema.
pub const MAX_EXTREMA: usize = 4;

/// A parametrized curve that reports its extrema.
pub trait ParamCurveExtrema: ParamCurve {
    /// Compute the extrema of the curve.
    ///
    /// Only extrema within the interior of the curve count.
    /// At most four extrema can be reported, which is sufficient for
    /// cubic Béziers.
    ///
    /// The extrema should be reported in increasing parameter order.
    fn extrema(&self) -> ArrayVec<f64, MAX_EXTREMA>;

    /// Return parameter ranges, each of which is monotonic within the range.
    fn extrema_ranges(&self) -> ArrayVec<Range<f64>, { MAX_EXTREMA + 1 }> {
        let mut result = ArrayVec::new();
        let mut t0 = 0.0;
        for t in self.extrema() {
            result.push(t0..t);
            t0 = t;
        }
        result.push(t0..1.0);
        result
    }

    /// The smallest rectangle that encloses the curve in the range (0..1).
    fn bounding_box(&self) -> Rect {
        let mut bbox = Rect::from_points(self.start(), self.end());
        for t in self.extrema() {
            bbox = bbox.union_pt(self.eval(t));
        }
        bbox
    }
}