[Paraview] DatasetTriangleFilter "creating" non-tet cells during volume visualization

[Gage, Kenneth]

Will, Lisa -

Thanks for looking into this; I didn't find this in an application that I myself built but in ParaView stable and development; Does ParaView perform the cell type check you mentioned prior to calling vtkDataSetTriangleFilter before volume visualization?  

The grid I have is quite fine in some areas due to the need to resolve flow accurately; could there be some merging of near-coincident nodes during the import-triangularization process that could cause "non-tet" cells to be created? Unfortunately, I don't have a simple dataset that could be an example.  The dataset I have is about 1.4 - 1.5 million tets but I could ftp it if you would like to test it on your installation. I can also try to put a small program together to see what dumps out of vtkDataSetTriangleFilter and let you know what I find. 

Thanks and best regards -
Ken

Kenneth Gage
MD-PhD Student
McGowan Institute for Regenerative Medicine
University of Pittsburgh
228 MCGOWN
3025 East Carson Street
Pittsburgh, PA 15203-2155



-----Original Message-----
From: Will Schroeder [mailto:will.schroeder at kitware.com]
Sent: Tue 8/16/2005 7:44 PM
To: Lisa Avila; Gage, Kenneth; paraview at paraview.org
Subject: Re: [Paraview] DatasetTriangleFilter "creating" non-tet cells  during volume visualization
 
The way the filter is right now it assumes that people will not use 
vtkDataSetTriangleFilter if they already have tetrahedra. There is an 
easy test for this: if you invoke vtkDataSet:GetCellTypes() it 
returns a list of all the types of cells found in a dataset. Using 
this test you can short circuit (i.e., skip using) the filter.

Having said this, I believe VTK should make life easier, not harder; 
and pipelines that change based on datatypes are hard to program in 
ParaView So IMHO the filter should be modified so that it simply 
passes data through in cases where the input is already simplices. 
This could be done pretty easily, and I'd be willing to do it before 
VTK 5.0 is cut.

Further, I am concerned that non-tetra cells are being created. While 
this is hard to believe, it is after all software and nothing 
surprises me anymore :-)  I'd like to see a simple dataset where this 
happens so we can find the problem and fix it.

Will

At 05:10 PM 8/16/2005, Lisa Avila wrote:
>Hi Ken,
>
>I am calling in some expert help on this since I am not the author 
>of vtkDataSetTriangleFilter.
>
>Will: when given an input that consists solely of tetrahedra, what 
>does vtkDataSetTriangleFilter do? Does it actually copy its output? 
>Can it possibly produce non-tetra cells?
>
>Lisa
>
>
>At 10:34 AM 8/16/2005, Gage, Kenneth wrote:
>>Dear ParaView list -
>>
>>I am exporting Ensight Gold format files from a FLUENT CFD 
>>simulation for reading into ParaView.  The simulation consists of 
>>all tetrahedrons (and triangle surfaces) - I checked the output 
>>file manually to confirm this fact.  ParaView reports the presence 
>>of non-tet cells when attempting volume visualization, which I 
>>assume are created during the use of DatasetTriangleFilter prior to 
>>visualization.  Has anyone else seen this problem, and is there 
>>some sort of workaround?
>>
>>Best regards -
>>Ken
>>
>>Kenneth Gage
>>MD-PhD Student
>>McGowan Institute for Regenerative Medicine
>>University of Pittsburgh
>>228 MCGOWN
>>3025 East Carson Street
>>Pittsburgh, PA 15203-2155
>>
>>_______________________________________________
>>ParaView mailing list
>>ParaView at paraview.org
>>http://www.paraview.org/mailman/listinfo/paraview
>
>