{"success":true,"data":{"query":"Spatial Computing","limit":10,"count":3,"sources":["spatial_computing.tah"],"synced":[],"results":[{"source":"spatial_computing.tah","text":"Spatial Math: Avoid Euler angles to prevent Gimbal Lock. Use THREE.Quaternion for all rotations. Interpolation: Use .slerp() for smooth orientation changes. Matrix Ops: Use matrix.decompose(pos, quat, scale) to extract transforms. Culling: Manual frustum culling via camera.frustum.containsPoint() for large-scale procedural data.","score":0.44,"links":[]},{"source":"spatial_computing.tah","text":"Spatial Indexing: BVH (Bounding Volume Hierarchy) is standard for complex raycasting/picking (use three-mesh-bvh). Octrees are best for sparse 3D point-clouds. Spatial Hashing: O(1) constant-time neighbor lookups for uniform dynamic objects (particles/boids).\n\n[SWARM_LINKS] 2b2bb4b414ee -> UNRESOLVED","score":0.44,"links":[]},{"source":"spatial_computing.tah","text":"Hybrid Spatial Pattern: Integrate Rotoscope SAM 2 masks as dynamic textures or sprites. Use Raycasting (BVH optimized) to project mouse/touch coordinates onto the 3D 'TacticalCloth' mesh. Apply physics forces programmatically using the mesh vertex data extracted via Matrix decomposition.\n\n[SWARM_LINKS] 7b0e4df6bcdd -> UNRESOLVED, e76e5f73b131 -> UNRESOLVED","score":0.44,"links":[]}]},"metadata":{},"timestamp":"2026-10-07T17:01:33.320Z"}