Within radius incoherent results

[STPR]

Hello,
I would like to use it but it seems that some points aren't found while they obviously shall be.
I've made a simple example to explain my problem and I've the same result with the v0.2.5 and v2.0.0-beta.8.
This is the v2.0.0-beta.8 code:

use kiddo::{KdTree, distance::squared_euclidean};

fn main() -> Result<(), std::io::Error> {

let mut tree: KdTree<f64, 2> = KdTree::new();

tree.add(&[ 1.0,  2.0], 1);
tree.add(&[ 2.0,  3.0], 2);
tree.add(&[ 0.0, -1.0], 3);
tree.add(&[-2.0,  0.0], 4);
tree.add(&[ 2.0, -3.0], 5);
tree.add(&[-4.0,  1.0], 6);

println!("{:?}", tree);

let within = tree.within_unsorted(&[0.0, 0.0], 2.5f64, &squared_euclidean);
println!("{:?}", within);

    Ok(())
}

The ouput is:

KdTree { leaves: [LeafNode { content_points: [[1.0, 2.0], [2.0, 3.0], [0.0, -1.0], [-2.0, 0.0], [2.0, -3.0], [-4.0, 1.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0], [0.0, 0.0]], content_items: [1, 2, 3, 4, 5, 6, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], size: 6 }], stems: [], root_index: 2147483647, size: 6 }
[(1.0, 3)]

Which means it only found the point (0, -1) but 2 others points are within radius 2.5 at (0, 0), the (1, 2) and (-2, 0)

[STPR]

Every distances are squared 😅 so it works correctly only if I use the squared radius but it's not written in the documentation:

let within = tree.within_unsorted(&[0.0, 0.0], 6.25f64, &squared_euclidean);

[sdd]

Hi! Thanks for reporting - this was a regression that was introduced in 2.0.0.beta.7, I also spotted this today. I should have this fixed over the next day or two, but in the meantime you should be able to revert to 2.0.0.beta.6 to avoid the problem.

Thanks for reporting!

[sdd]

Actually sorry, I misread your report. Yes, for now the results are coming out squared, and that will be the same even in 2.0.0.beta.6. I'll update the documentation.

[STPR]

That the returned distances are squared must be in the documentation.
But I think the command must square the radius itself.
I've had to use 6.25 to search the points within radius 2.5.

[sdd]

I've updated the documentation to make the behaviour clearer, as well as renaming the parameter from radius to dist.

You'll need a slight refactor also as the results are now returned as a Neighbour struct rather than a tuple.

Released as https://crates.io/crates/kiddo/2.0.0-beta.9