[vtk-developers] Patch for vtkHierarchicalBoxDataSet

[Berk Geveci]

Hi Matthew,

Correct me if I am wrong but this patch is not enough by itself. Did
you by any chance modify the wrapping hints file?

Also, it should be relatively easy to make vtkAMRBox subclass of VTK
data object as long as we are not storing it in the meta-data. If that
is going to make your life easier, I would be happy to give some
pointers.

Also, are you considering expanding the capabilities of the AMR
support in VTK? If the answer is yes, I would love to collaborate.

-berk


2008/9/19 Matthew Turk <matthewturk at gmail.com>:
> Hi there,
>
> I'm a developer of an analysis toolkit for Adaptive Mesh Refinement
> data, and we're very interested in using VTK as a means of exploring
> that data.  After browsing through the various available classes and
> means of accessing this data, as well as reading over the plans for
> inclusion in the future, it seemed to us that our best bet for
> enabling this exploration now was to use the vtkHierarchicalBoxDataSet
> interface.  In particular, we wanted a solution that would be
> wrappable by Python, as our entry point will be via a Python-based
> toolkit, probably through TVTK.
>
> At first, we prepared a fairly extensive patch attempting to expose
> vtkAMRBox as a vtkObject, implementing the various pure virtual
> methods available there.  This would enable Python applications to
> instantiate vtkAMRBoxes, set their properties, and feed them to the
> existing vtkHierarchicalBoxDataSet::SetData(...) method.  However, the
> reference counting was problematic, and it became clear that this
> could be avoided by wrapping the creation of a vtkAMRBox.
>
> The attached, rather modest, patch implements a new method
> (vtkHierarchicalBoxDataSet::SetData(level, id, LoCorner, HiCorner,
> dataSet) ) that accepts the metadata affiliated with a vtkAMRBox and
> generates it inline to pass to the original, non-wrapped,
> vtkHierarchicalBoxDataSet::SetData(...) .
>
> I was hoping to have this considered for inclusion, as it would enable
> us to have collaborators use an off-the-shelf VTK to visualize our
> data.  If I can do anything to encourage its inclusion, please let me
> know.
>
> Thanks!
>
> Matt Turk
>
> _______________________________________________
> vtk-developers mailing list
> vtk-developers at vtk.org
> http://www.vtk.org/mailman/listinfo/vtk-developers
>
>