I have a 3D spatiotemporal ($x, y, t$) index. I would really like to be able to collapse one or more dimensions so I can separately factor the spatial ($x, y$) and temporal ($t$) dimensions. Is this something that rtree supports?
Context: TorchGeo stores all geospatial files in a 3D spatiotemporal index. Historically, we would randomly sample files at random. However, we would now like to have separate samplers for space and time so that they can be arbitrarily combined. The spatial samplers would only sample over the spatial axis, while the temporal samplers would only sampler over the temporal axis.
I have a 3D spatiotemporal ($x, y, t$ ) index. I would really like to be able to collapse one or more dimensions so I can separately factor the spatial ($x, y$ ) and temporal ($t$ ) dimensions. Is this something that rtree supports?
Context: TorchGeo stores all geospatial files in a 3D spatiotemporal index. Historically, we would randomly sample files at random. However, we would now like to have separate samplers for space and time so that they can be arbitrarily combined. The spatial samplers would only sample over the spatial axis, while the temporal samplers would only sampler over the temporal axis.