logo
  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

use super::{swap_remove_to_first, trivial_hull};
use crate::algorithm::kernels::*;
use crate::{Coordinate, GeoNum, LineString};

/// The [Graham's scan] algorithm to compute the convex hull
/// of a collection of points. This algorithm is less
/// performant than the quick hull, but allows computing all
/// the points on the convex hull, as opposed to a strict
/// convex hull that does not include collinear points.
///
/// # References
///
/// Graham, R.L. (1972). ["An Efficient Algorithm for
/// Determining the Convex Hull of a Finite Planar
/// Set"](http://www.math.ucsd.edu/~ronspubs/72_10_convex_hull.pdf)
/// (PDF). \
/// Information Processing Letters. 1 (4): 132–133.
/// [doi:10.1016/0020-0190(72)90045-2](https://doi.org/10.1016%2F0020-0190%2872%2990045-2)
///
/// [Graham's scan]: //en.wikipedia.org/wiki/Graham_scan
pub fn graham_hull<T>(mut points: &mut [Coordinate<T>], include_on_hull: bool) -> LineString<T>
where
    T: GeoNum,
{
    if points.len() < 4 {
        // Nothing to build with fewer than four points.
        return trivial_hull(points, include_on_hull);
    }

    // Allocate output vector
    let mut output = Vec::with_capacity(points.len());

    // Find lexicographically least point and add to hull
    use crate::utils::least_index;
    use std::cmp::Ordering;
    let min_idx = least_index(points);
    let head = swap_remove_to_first(&mut points, min_idx);
    output.push(*head);

    // Sort rest of the points by angle it makes with head
    // point. If two points are collinear with head, we sort
    // by distance. We use kernel predicates here.
    let cmp = |q: &Coordinate<T>, r: &Coordinate<T>| match T::Ker::orient2d(*q, *head, *r) {
        Orientation::CounterClockwise => Ordering::Greater,
        Orientation::Clockwise => Ordering::Less,
        Orientation::Collinear => {
            let dist1 = T::Ker::square_euclidean_distance(*head, *q);
            let dist2 = T::Ker::square_euclidean_distance(*head, *r);
            dist1.partial_cmp(&dist2).unwrap()
        }
    };
    points.sort_unstable_by(cmp);

    for pt in points.iter() {
        while output.len() > 1 {
            let len = output.len();
            match T::Ker::orient2d(output[len - 2], output[len - 1], *pt) {
                Orientation::CounterClockwise => {
                    break;
                }
                Orientation::Clockwise => {
                    output.pop();
                }
                Orientation::Collinear => {
                    if include_on_hull {
                        break;
                    } else {
                        output.pop();
                    }
                }
            }
        }
        // Corner case: if the lex. least point added before
        // this loop is repeated, then we should not end up
        // adding it here (because output.len() == 1 in the
        // first iteration)
        if include_on_hull || pt != output.last().unwrap() {
            output.push(*pt);
        }
    }

    // Close and output the line string
    let mut output = LineString(output);
    output.close();
    output
}

#[cfg(test)]
mod test {
    use super::*;
    use crate::algorithm::is_convex::IsConvex;
    fn test_convexity<T: GeoNum>(initial: &[(T, T)]) {
        let mut v: Vec<_> = initial
            .iter()
            .map(|e| Coordinate::from((e.0, e.1)))
            .collect();
        let hull = graham_hull(&mut v, false);
        assert!(hull.is_strictly_ccw_convex());
        let hull = graham_hull(&mut v, true);
        assert!(hull.is_ccw_convex());
    }

    #[test]
    fn test_graham_hull_ccw() {
        let initial = vec![
            (1.0, 0.0),
            (2.0, 1.0),
            (1.75, 1.1),
            (1.0, 2.0),
            (0.0, 1.0),
            (1.0, 0.0),
        ];
        test_convexity(&initial);
    }

    #[test]
    fn graham_hull_test1() {
        let v: Vec<_> = vec![(0, 0), (4, 0), (4, 1), (1, 1), (1, 4), (0, 4), (0, 0)];
        test_convexity(&v);
    }

    #[test]
    fn graham_hull_test2() {
        let v = vec![
            (0, 10),
            (1, 1),
            (10, 0),
            (1, -1),
            (0, -10),
            (-1, -1),
            (-10, 0),
            (-1, 1),
            (0, 10),
        ];
        test_convexity(&v);
    }

    #[test]
    fn graham_test_complex() {
        let v = include!("../test_fixtures/poly1.rs");
        test_convexity(&v);
    }

    #[test]
    fn quick_hull_test_complex_2() {
        let coords = include!("../test_fixtures/poly2.rs");
        test_convexity(&coords);
    }
}