d803bfe2b1
git-subtree-dir: vendor/ruvector git-subtree-split: b64c21726f2bb37286d9ee36a7869fef60cc6900
80 lines
2.8 KiB
Rust
80 lines
2.8 KiB
Rust
//! Example: spatial indexing for fast nearest-neighbor search.
|
|
//!
|
|
//! Demonstrates:
|
|
//! - Creating a `SpatialIndex` from a point cloud
|
|
//! - kNN search for finding nearest neighbors
|
|
//! - Radius search for proximity queries
|
|
//! - Multiple distance metrics (Euclidean, Cosine, Manhattan)
|
|
|
|
use ruvector_robotics::bridge::{DistanceMetric, Point3D, PointCloud, SpatialIndex};
|
|
|
|
fn main() {
|
|
println!("=== Spatial Indexing Demo ===\n");
|
|
|
|
// Generate a synthetic warehouse layout.
|
|
let mut points = Vec::new();
|
|
// Shelves along the x-axis.
|
|
for row in 0..5 {
|
|
for col in 0..10 {
|
|
points.push(Point3D::new(
|
|
col as f32 * 2.0,
|
|
row as f32 * 3.0,
|
|
0.0,
|
|
));
|
|
}
|
|
}
|
|
// A few elevated points (items on shelves).
|
|
points.push(Point3D::new(4.0, 6.0, 2.5));
|
|
points.push(Point3D::new(4.1, 6.1, 2.6));
|
|
points.push(Point3D::new(8.0, 3.0, 1.8));
|
|
|
|
let cloud = PointCloud::new(points, 1_000_000);
|
|
println!("Point cloud: {} points", cloud.len());
|
|
|
|
// Build spatial index.
|
|
let mut index = SpatialIndex::new(3);
|
|
index.insert_point_cloud(&cloud);
|
|
println!("Index size: {} vectors\n", index.len());
|
|
|
|
// kNN search: find 5 nearest to robot position.
|
|
let robot_pos = [4.0, 6.0, 0.0];
|
|
println!("--- kNN Search (k=5, query={:?}) ---", robot_pos);
|
|
match index.search_nearest(&robot_pos, 5) {
|
|
Ok(neighbors) => {
|
|
for (i, (idx, dist)) in neighbors.iter().enumerate() {
|
|
println!(" #{}: index={}, distance={:.4}", i + 1, idx, dist);
|
|
}
|
|
}
|
|
Err(e) => println!(" Error: {}", e),
|
|
}
|
|
|
|
// Radius search: find all within 3.0 units.
|
|
println!("\n--- Radius Search (r=3.0, query={:?}) ---", robot_pos);
|
|
match index.search_radius(&robot_pos, 3.0) {
|
|
Ok(neighbors) => {
|
|
println!(" Found {} points within radius 3.0", neighbors.len());
|
|
for (idx, dist) in neighbors.iter().take(8) {
|
|
println!(" index={}, distance={:.4}", idx, dist);
|
|
}
|
|
if neighbors.len() > 8 {
|
|
println!(" ... and {} more", neighbors.len() - 8);
|
|
}
|
|
}
|
|
Err(e) => println!(" Error: {}", e),
|
|
}
|
|
|
|
// Cosine distance index.
|
|
println!("\n--- Cosine Distance Index ---");
|
|
let mut cos_index = SpatialIndex::with_metric(3, DistanceMetric::Cosine);
|
|
cos_index.insert_point_cloud(&cloud);
|
|
match cos_index.search_nearest(&[1.0, 1.0, 0.0], 3) {
|
|
Ok(neighbors) => {
|
|
println!(" Nearest 3 by cosine distance to [1, 1, 0]:");
|
|
for (idx, dist) in &neighbors {
|
|
println!(" index={}, cosine_dist={:.4}", idx, dist);
|
|
}
|
|
}
|
|
Err(e) => println!(" Error: {}", e),
|
|
}
|
|
}
|