Struct linfa_nn::BallTreeIndex

pub struct BallTreeIndex<'a, F: Float, D: Distance<F>> { /* private fields */ }

Spatial indexing structure created by BallTree

Implementations

impl<'a, F: Float, D: Distance<F>> BallTreeIndex<'a, F, D>

pub fn new<DT: Data<Elem = F>>(
    batch: &'a ArrayBase<DT, Ix2>,
    leaf_size: usize,
    dist_fn: D
) -> Result<Self, BuildError>

Creates a BallTreeIndex using the K-D construction algorithm

Trait Implementations

impl<'a, F: Debug + Float, D: Debug + Distance<F>> Debug for BallTreeIndex<'a, F, D>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'a, F: Float, D: Distance<F>> NearestNeighbourIndex<F> for BallTreeIndex<'a, F, D>

fn k_nearest<'b>(
    &self,
    point: ArrayView1<'b, F>,
    k: usize
) -> Result<Vec<(ArrayView1<'_, F>, usize)>, NnError>

Returns the k points in the index that are the closest to the provided point, along with their positions in the original dataset. Points are returned in ascending order of the distance away from the provided points, and less than k points will be returned if the index contains fewer than k.

fn within_range<'b>(
    &self,
    point: ArrayView1<'b, F>,
    range: F
) -> Result<Vec<(ArrayView1<'_, F>, usize)>, NnError>

Returns all the points in the index that are within the specified distance to the provided point, along with their positions in the original dataset. The points are not guaranteed to be in any order, though many algorithms return the points in order of distance.