A simple nearest neighbors implementation in rust.
A rust example, this same example can be run with the following command.
cargo run --example colors
use std::error::Error;
use closest::{Data, KDTree, Point, SquaredEuclideanDistance};
fn main() -> Result<(), Box<dyn Error>> {
// RGB color coordinates
let colors = vec![
Data::new("blue", vec![0., 0., 255.]),
Data::new("red", vec![255., 0., 0.]),
Data::new("navy", vec![17., 4., 89.]),
Data::new("purple", vec![171., 3., 255.]),
Data::new("light-blue", vec![61., 118., 224.]),
Data::new("pink", vec![255., 3., 213.]),
Data::new("yellow", vec![255., 234., 0.]),
Data::new("green", vec![16., 145., 25.]),
Data::new("orange", vec![255., 106., 0.]),
];
// Construct the tree from the vector of data points.
let tree = KDTree::from_vec(colors, 1).unwrap();
let point = Point::new(vec![237., 139., 69.]); // Light Orange
let closest_colors =
tree.get_nearest_neighbors(&point, 2, &SquaredEuclideanDistance::default());
println!("The nearest colors to light orange.");
for color in closest_colors {
println!("color: {}, squared euclidean distance: {}", color.data, color.distance);
}
Ok(())
}
// The nearest colors to light orange.
// color: yellow, squared euclidean distance: 14110
// color: orange, squared euclidean distance: 6174
And the equivalent python example.
from nearest import KDTree
colors = [
("blue", [0., 0., 255.]),
("red", [255., 0., 0.]),
("navy", [17., 4., 89.]),
("purple", [171., 3., 255.]),
("light-blue", [61., 118., 224.]),
("pink", [255., 3., 213.]),
("yellow", [255., 234., 0.]),
("green", [16., 145., 25.]),
("orange", [255., 106., 0.]),
];
tree = KDTree(colors)
light_orange = [237., 139., 69.]
print(tree.get_nearest_neighbors(light_orange, 2))
#> [(14110.0, 'yellow'), (6174.0, 'orange')]