passing multi-dimensional array

[uturuncoglu]

I wonder if there is any example that demonstrates sending multi-dimensional array back and forth between Fortran and Python.

[cyrush]

@uturuncoglu for round trip with the multi-d context, you will have to provide and manage additional shape information yourself.

We have an ndarray spec, where we use shape, along with optional offset and stride. These entries are stored as additional conduit nodes.

values - actuall data
shape - dims
stride - optional strides per dim
offset - optional additional indirection at start of each dim

Here is an example of a C++ object that helps with indexing in these cases:

https://github.com/LLNL/conduit/blob/554373f4bd4ffdfbd5bb9c43ac6152ed4c856f57/src/libs/blueprint/conduit_blueprint_ndarray_index.hpp#L66

For numpy cases and fortran, the shape, offset, and stride could be used to construct a native nd representation - but we don't have direct support for that in conduit.

[uturuncoglu]

@cyrush Thanks for the clarification. It makes sense. I was wondering about direct support for multi-dimensional array s but this also works for me.

[uturuncoglu]

@cyrush Sorry for opening this issue again but do I need to define shape, stride and offset under field or it would be defined in topologies/mesh section. It is not clear in the conduit_blueprint_ndarray_index.hpp. Is there any simple node output for it? So, I could reference it.

[cyrush]

@uturuncoglu sorry I missed this issue. we have tests for the ndarray index here:

https://github.com/LLNL/conduit/blob/554373f4bd4ffdfbd5bb9c43ac6152ed4c856f57/src/tests/blueprint/t_blueprint_ndarray_index.cpp

the NDIndex accepts a node that has shape offset and stride

Shape is required, offset and stride are optional.

As far as how strides and ND shape fits into a blueprint mesh, please see the "strided structured" example mesh, here is an example:

Note: the shape is always derived from the topology, this example shows different ways to provide offsets and strides.

import conduit
n = conduit.Node()
conduit.blueprint.mesh.examples.generate("strided_structured",n)
print(n.to_yaml())

state: 
 time: 3.1415
 cycle: 100
coordsets: 
 coords: 
   type: "explicit"
   values: 
     x: [-10.0, -6.66666666666667, -3.33333333333333, 4.44089209850063e-16, 3.33333333333333, 6.66666666666667, 10.0, -10.0, -6.66666666666667, -3.33333333333333, 4.44089209850063e-16, 3.33333333333333, 6.66666666666667, 10.0, -10.0, -6.66666666666667, -3.33333333333333, 4.44089209850063e-16, 3.33333333333333, 6.66666666666667, 10.0, -10.0, -6.66666666666667, -3.33333333333333, 4.44089209850063e-16, 3.33333333333333, 6.66666666666667, 10.0, -10.0, -6.66666666666667, -3.33333333333333, 4.44089209850063e-16, 3.33333333333333, 6.66666666666667, 10.0, -10.0, -6.66666666666667, -3.33333333333333, 4.44089209850063e-16, 3.33333333333333, 6.66666666666667, 10.0]
     y: [-10.0, -10.0, -10.0, -10.0, -10.0, -10.0, -10.0, -6.0, -6.0, -6.0, -6.0, -6.0, -6.0, -6.0, -2.0, -2.0, -2.0, -2.0, -2.0, -2.0, -2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 2.0, 6.0, 6.0, 6.0, 6.0, 6.0, 6.0, 6.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0, 10.0]
topologies: 
 mesh: 
   type: "structured"
   coordset: "coords"
   elements: 
     dims: 
       i: 3
       j: 2
       offsets: [2, 2]
       strides: [1, 7]
fields: 
 vert_vals: 
   association: "vertex"
   topology: "mesh"
   offsets: [2, 2]
   strides: [1, 7]
   values: [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 2.0, 3.0, 4.0, 0.0, 0.0, 0.0, 5.0, 6.0, 7.0, 8.0, 0.0, 0.0, 0.0, 9.0, 10.0, 11.0, 12.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
 ele_vals: 
   association: "element"
   topology: "mesh"
   offsets: [2, 2]
   strides: [1, 7]
   values: [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 2.0, 3.0, 0.0, 0.0, 0.0, 0.0, 4.0, 5.0, 6.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]

[uturuncoglu]

@cyrush No worries. Thanks for the information and help. Let me try to see how it works for the regular quad grid and update about you if I have any more clarification. I might also need to apply same for the unstructured grid (spatial mesh with triangular and data in layers/depth) but that is a work for later.