Struct SpaceMatrix 

pub struct SpaceMatrix { /* private fields */ }

A general finite metric space with color classes.

• Distances are given by a symmetric distance matrix of size nxn,
• Colors are given by a list of size n,
• Implements the ColoredMetric trait.

Implementations

impl SpaceMatrix

pub fn new(distances: Vec<Vec<f32>>, colors: Vec<usize>) -> SpaceMatrix

Creates a new SpaceMatrix.

Input

• distances has to be a 2D-vector that represents a quadratic matrix satisfying the metric properties (symmetry, non-negativity, identity of indiscernibles and the triangle inequality),
• colors is a vector that stores the color of each point.

The running time of asserting the metric properties is O(n³).

Panics

• Panics, if distances is not a quadratic matrix.
• Panics, if the length of the color classes does not match the number of points given by distances.
• Panics, if distances does not satisfy the metric properties.

pub fn new_by_array<const N: PointCount>( distances: [[f32; N]; N], colors: [usize; N], ) -> SpaceMatrix

Creates a new SpaceMatrix as in new but the distances and color data are received by arrays.

Example

use ff_k_center::{Point,SpaceMatrix,ColoredMetric};
let space = SpaceMatrix::new_by_array(
                [[0.0, 2.0, 1.5],
                 [2.0, 0.0, 0.6],
                 [1.5, 0.6, 0.0]], [0, 0, 1]);
let points : Vec<&Point> = space.point_iter().collect();
assert!(space.is_metric());
assert_eq!(space.dist(points[1],points[2]),0.6);
assert_eq!(space.color(points[2]),1);
assert_eq!(space.n(),3);

Trait Implementations

impl ColoredMetric for SpaceMatrix

fn dist(&self, x1: &Point, x2: &Point) -> f32

Returns the distance between two points x1 and x2.

fn color(&self, x: &Point) -> usize

Returns the color of point x.

fn n(&self) -> PointCount

Return the number of points in the metric space.

fn get_point(&self, idx: usize) -> Option<&Point>

Return a reference (wrapped in Some) to the point with provided index. If there is not point with this index, return None.

fn point_iter(&self) -> Iter<'_, Point>

Provides an iterator of all points of the metric space. This is the only way to access the points.

fn gamma(&self) -> ColorCount

Returns the number of color classes that are present in the metric space. More precesely it return the highest color value + 1.

fn dist_set(&self, x: &Point, point_set: Vec<&Point>) -> f32

fn get_closest<'a>( &self, x: &Point, point_set: &Vec<&'a Point>, ) -> (f32, &'a Point)

fn is_metric(&self) -> bool

Checks whether the dist function satisfy the metric properties in which case true is returned. Care: this need O(n³) time.