Poindexter
Snider LabsA Go library for spatial indexing - finding things that are near other things, quickly. Used for WASM KDTree lookups in UEPS networking NAT tables. Supports nearest-neighbour, k-nearest, and radius searches with two backends: a gonum-optimised version for speed, or a simpler linear backend. Compiles to WASM so browsers can make peer routing decisions without a server round-trip.
Features
Generic KDTree
KDTree[T] with parameterised payload - nearest-neighbour, k-nearest, radius search, insert/delete by ID
Dual Backend
Gonum backend (median-split with branch-and-bound pruning) or linear backend via WithBackend()
Distance Metrics
Euclidean (L2), Manhattan (L1), Chebyshev (L∞), Cosine, WeightedCosine - extensible via DistanceMetric interface
Point Construction
BuildND/2D/3D/4D helpers with axis inversion, per-axis weighting, ComputeNormStatsND()
Analytics Tracking
TreeAnalytics and PeerAnalytics for operation monitoring
Sorting & Search
SortInts/Strings/Float64s, generic SortBy(), binary search capabilities
Installation
go get github.com/Snider/PoindexterUsage
KDTree Spatial Search
pts := []poindexter.KDPoint[string]{
{ID: "A", Coords: []float64{0, 0}, Value: "alpha"},
{ID: "B", Coords: []float64{1, 1}, Value: "beta"},
{ID: "C", Coords: []float64{2, 2}, Value: "gamma"},
}
tree, _ := poindexter.NewKDTree(pts,
poindexter.WithMetric(poindexter.EuclideanDistance{}),
poindexter.WithBackend(poindexter.BackendGonum),
)
// Find nearest neighbour to query point
nearest := tree.NearestNeighbor([]float64{0.5, 0.5})
// K-nearest neighbours
kNearest := tree.KNearest([]float64{1, 1}, 2)
// Radius search
inRadius := tree.RadiusSearch([]float64{0, 0}, 1.5)Sorting Utilities
numbers := []int{3, 1, 4, 1, 5, 9}
poindexter.SortInts(numbers)
// Result: [1, 1, 3, 4, 5, 9]Quick Links
More from Snider Labs
Fancy helping out?
Spotted a bug? Got an idea? We'd love to hear from you.
Read the contributing guide →