Luxar Viewer API Documentation - v2026.9.22
    Preparing search index...

    Canonical chunk spatial index shape consumed by SpatialQueryBuilder.

    All three geometry types (Points, Lines segments, GSplats) produce indices of this shape: a flattened (numChunks, ndim, 2) Float32Array of [min, max] bounds per chunk per dimension, plus the chunk count and dimensionality.

    Lines have a dual index (vertex + segment); only the segment side is used for the chunk query. The vertex side is loaded separately and consumed via a different code path (sorted-indices → contiguous ranges).

    interface ChunkSpatialIndex {
        chunkBounds: Float32Array;
        chunkCount: number;
        metadata: { ndim: number; chunk_size?: number };
    }
    Index
    chunkBounds: Float32Array

    Chunk bounding boxes: shape (numChunks, ndim, 2) flattened row-major.

    chunkCount: number

    Number of chunks in the index.

    metadata: { ndim: number; chunk_size?: number }

    Common metadata fields used by the query path.

    Type Declaration

    • ndim: number

      Full dimensionality of the dataset.

    • Optionalchunk_size?: number

      Elements per chunk (points/segments/splats).