1D to Strip

[agcapps]

Adds a verify() function and strip-conversion routines

[agcapps]

This is at the discussion stage. In particular, the results need to be tested against other libraries that expect strip meshes, and we need to examine assumptions.

[agcapps]

Currently, the oneD::verify function expects a structured mesh, as do the conversion routines. verify, at least, should support uniform meshes (specifying a dx and an origin). The conversion routines should support uniform meshes if needed by users. In addition, we should probably also support conversion between 1D structured and 1D uniform and back.

[cyrush]

Instead of a new verify case, can we identity existing verified meshes with a subset of properties for this case?

[cyrush]

Suggested methods for strip creation:

Suggested top level method:

blueprint::mesh::generate_strip(conduit::Node& mesh,
      const std::string& src_topo_name,
      const std::string& dst_topo_name);

Suggested topology specific method:

blueprint::mesh::topology::generate_strip(const conduit::Node &topo,
       conduit::Node &topo_dest,
       conduit::Node &coords_dest)

[cyrush]

Q: For the result strip mesh, does that have to be uniform/rectilinear/structured for the carter use case, or could it be unstructured for any case we could care about?

[agcapps]

I don't know for sure if Carter wants uniform/rectilinear/structured. I will try to track down a test case and model off that.

[agcapps]

@cyrush , any more thoughts on this PR?

[cyrush]

Yes, I have been pondering.

We can chat at our meeting today.