(feature/fVDB) [Question/Feature] How to add two VDBTensors with different topologies

[OasisArtisan]

Thanks for the great work.

I had two questions:

1- Is it possible to add/aggregate VDB Tensors such that voxels that coincide are averaged or aggregated using an arbitrary function that the user can specify ?

2- Is fVDB a fixed sized voxel representation? I understand that OpenVDB allows storing voxels at different resolutions with the pruneGrid method for higher efficiency. Does fVDB still have this capability ?

Thank you

[OasisArtisan]

I see from

openvdb/fvdb/fvdb/nn/vdbtensor.py

Lines 193 to 197 in f1d9e2b

	def _binop(self, other, op):
	if isinstance(other, VDBTensor):
	return VDBTensor(self.grid, op(self.data, other.data), self.kmap)
	else:
	return VDBTensor(self.grid, op(self.data, other), self.kmap)

That it is possible to add VDBTensors as long as they have the same Grid/Topology. Is there a way to aggregate them if they have different Grids ?

The application I am interested in is robot mapping where multiple frames typically come sequentially each with its own point cloud and I would like to aggregate them with their features in a VDBTensor representation.