[Paraview] particletracer, vectors menu?

[Berk Geveci]

I think a flag or maybe a new domain is best.

-berk

On 2/15/08, John Biddiscombe <biddisco at cscs.ch> wrote:
> Berk
>
>  Good point. I forgot all about that. We should add a check. The trouble
>  is, how does the pv gui know that ParticleTracer needs time? Detecting
>  the output is easy becuase the filter already exists and stesp can be
>  queried, but the input is only declared in the XML as input <dataobject>
>  - do we have to add an XML flag to say Input requirements point vectors,
>  dataobject and Time steps?
>
>  I'll add the bug/feature request
>
>
>  JB
>
>
>  > John,
>  >
>  > Did you try what happens if you connect a source that doesn't support
>  > time? I didn't want to commit a fix without checking that. We should
>  > probably add support to paraview to detect a dataset that has time in
>  > it. This is already done for sources/readers and can be expanded to
>  > any algorithm. If you want, post a request in the bug tracker and I'll
>  > look into it in a few weeks.
>  >
>  > -berk
>  >
>  > On 2/15/08, John Biddiscombe <biddisco at cscs.ch> wrote:
>  >
>  >> Berk wrote:
>  >>  > This is actually simple to fix. In the filters xml
>  >>  > (Servers/ServerManager/Resources/filters.xml), ParticleTracer says it
>  >>  > is expecting a vtkTemporalDataSet input. However, in some cases, the
>  >>  > input may look like a vtkUnstructuredGrid, for example. The input type
>  >>  > for ParticleTracer should be set to vtkDataObject to fix this. I am
>  >>  > attaching a patch. I am not sure this is the best way to fix this in
>  >>  > the main repository so I am not committing it yet.
>  >>  >
>  >>
>  >>
>  >> I don't know why I had trouble yesterday, but as Berk says, it should be
>  >>  vtkDataObject as input not vtkTemporalDataset - and when I put this in,
>  >>  I was able to do particle tracing on a volume (vtkImageData) dataset.
>  >>  (Didn't try the LEA one, but it's safe to assume it'll work).
>  >>
>  >>  For those that care - vtkTemporalDataset should not generally be
>  >>  imported directly by a filter, instead the filter should request
>  >>  multiple time steps and let the executive create the temporal wrapper.
>  >>  We have one use case (referred to in my xdmf post which causes trouble,
>  >>  but we'll fix that eventually).
>  >>
>  >>
>  >>  JB
>  >>
>  >>
>  >>
>
>
>
> --
>
> John Biddiscombe,                            email:biddisco @ cscs.ch
>  http://www.cscs.ch/about/BJohn.php
>  CSCS, Swiss National Supercomputing Centre  | Tel:  +41 (91) 610.82.07
>  Via Cantonale, 6928 Manno, Switzerland      | Fax:  +41 (91) 610.82.82
>
>
>