Struct space::LinearKnn

pub struct LinearKnn<I>(pub I);

Performs a linear knn search by iterating over everything in the space and performing a binary search on running set of neighbors.

Example

use space::{Knn, LinearKnn, MetricPoint, Neighbor};

#[derive(PartialEq)]
struct Hamming(u8);

impl MetricPoint for Hamming {
    type Metric = u8;

    fn distance(&self, other: &Self) -> Self::Metric {
        (self.0 ^ other.0).count_ones() as u8
    }
}

let data = [
    Hamming(0b1010_1010),
    Hamming(0b1111_1111),
    Hamming(0b0000_0000),
    Hamming(0b1111_0000),
    Hamming(0b0000_1111),
];

let search = LinearKnn(data.iter());
assert_eq!(
    &search.knn(&Hamming(0b0101_0000), 3),
    &[
        Neighbor { index: 2, distance: 2 },
        Neighbor { index: 3, distance: 2 },
        Neighbor { index: 0, distance: 6 },
    ]
);

Trait Implementations

impl<'a, P: 'a, I> Knn<P> for LinearKnn<I> where
    P: MetricPoint,
    I: Clone + Iterator<Item = &'a P>, 

type KnnIter = Vec<Neighbor<P::Metric>>

fn knn(&self, query: &P, num: usize) -> Self::KnnIter

Get num nearest neighbors of target.

fn nn(&self, query: &P) -> Option<Neighbor<P::Metric>>

Get the nearest neighbor of target.