MantisBT - ParaView
View Issue Details
IDProjectCategoryView StatusDate SubmittedLast Update
0011918ParaViewBugpublic2011-03-02 15:462012-04-12 20:57
ReporterGreg Weirs 
Assigned ToRobert Maynard 
PriorityhighSeverityminorReproducibilityalways
StatusclosedResolutionfixed 
PlatformOSOS Version
Product VersionDevelopment 
Target VersionFixed in Version3.12 
ProjectSandia
Topic Namemore-prism-support
Type
Summary0011918: Prism Filter does not accept point data as input
DescriptionThe prism filter expects cell data as pipeline input, but this should not be a requirement. A significant capability that relies on the input being cell data is linked selection - selection in the prism view is associated by GlobalElementID to the data in the other view(s).
Steps To Reproduce
Additional InformationGreg,

There are a two things that would need to be modified in Prism to handle both point and cell data seamlessly. Prism would need to be modified to handle point data at the lowest level. When I say point data I mean that the entities of interest are the points regardless of the cell type. A VTK point can be in multiple VTK cells. Handling point data is more straight forward than what is currently done for cells in Prism. However there are some questions about that would need to be worked out to help Prism know when to process the input as point data versus cell data.

The second thing that would need to be worked out is the linked selection. Prism uses the GlobalElementID to link selection. I believe there is a GlobalPointID that it could use with point data. So it may just be a matter of modifying the existing code to look for GlobalPointID as well as GlobalElementIDs.


I wish that we could have tackled use case sooner. It wouldn't have taken very long to make the necessary changes.


Corey.






On Mon, Feb 28, 2011 at 4:49 PM, Weirs, V Gregory <vgweirs@sandia.gov> wrote:
Hi Corey,
I think I understand the distinction between ParaView/VTK terminology and analysis terminology. The main point is that the analysis data may be point data or cell data; it just turns out that for exodus and spcth it is cell data. For many similar codes it is point data. Tracer data (hscth) is really point data. Prism should not care which it is, as long as the cells or points have the thermodynamic variables matching the SESAME data.

So the question is, what features of Prism rely on the simulation data being cell data? Is it more than linked selection?

Thanks,
Greg
TagsNo tags attached.
Relationships
Attached Files
Issue History
Date ModifiedUsernameFieldChange
2011-03-02 15:46Greg WeirsNew Issue
2011-03-25 15:28Robert MaynardAssigned To => Robert Maynard
2011-03-25 15:28Robert MaynardStatusbacklog => tabled
2011-06-23 10:36Robert MaynardStatusbacklog => todo
2011-06-23 10:36Robert MaynardStatustodo => active development
2011-06-23 15:26Utkarsh AyachitProject => Sandia
2011-07-19 17:21Robert MaynardTopic Name => more-prism-support
2011-07-19 17:21Robert MaynardNote Added: 0027028
2011-07-19 17:21Robert MaynardStatusactive development => gatekeeper review
2011-07-19 17:21Robert MaynardResolutionopen => fixed
2011-07-21 19:18Utkarsh AyachitNote Added: 0027051
2011-07-22 14:13Utkarsh AyachitStatusgatekeeper review => customer review
2011-07-22 14:13Utkarsh AyachitNote Added: 0027066
2011-07-29 16:18Utkarsh AyachitFixed in Version => 3.12
2012-04-12 20:57Alan ScottNote Added: 0028472
2012-04-12 20:57Alan ScottStatuscustomer review => closed

Notes
(0027028)
Robert Maynard   
2011-07-19 17:21   
Only support for reading point based properties have been completed. The selection issue with global point ids is in bug 0012380.
(0027051)
Utkarsh Ayachit   
2011-07-21 19:18   
Scheduled for gatekeeper review on Jul 22, 2011 at 2 pm EDT.
(0027066)
Utkarsh Ayachit   
2011-07-22 14:13   
Merged to git-master.
(0028472)
Alan Scott   
2012-04-12 20:57   
Assuming fixed.