[Paraview] Problem with temporal filter and multi-block data
Moreland, Kenneth
kmorel at sandia.gov
Tue Jun 2 11:48:15 EDT 2009
OK. What you describe is not part of the original posting of 6662, but I'll take your word for it that there is a larger issue encompassing both problems. Perhaps this is documented in the voluminous notes attached to 6662 that I'm not going to bother to read.
-Ken
On 6/2/09 9:18 AM, "John Biddiscombe" <biddisco at cscs.ch> wrote:
Ken
I disagree with your disagreement. TemporalDataset is a TypeOf CompoundDataset. In the executive, there is a check for IsTypeOf(insert compound stuff here), and the test returns true when it gets temporal, but should be returning false because the filter wants multi-block - which is a sibling of temporal. 6662 is causing a great deal of errors related to temporal data and fixing it will also fix 9089 (at least I think so). 6662 really addresses 2 serious problems in my mind. One is the wrong type checks above, the second is the incorrect creation of Temporal types when not really needed. Although they are two bugs, I propose a pipeline /executive fix which should address both.
On a side note. Does the original poster know that the temporalPathline filter already does what he wants? (build trajectories of particles etc?) - though not multi-block.
JB
Re: [Paraview] Problem with temporal filter and multi-block data I actually disagree with that statement. The issue with 6662 is that the pipeline is sometimes stuffing data in a vtkTemporalDataSet when it should not. If you look carefully at 9089, you will see that in fact the data is supposed to be in vtkTemporalDataSet; the downstream temporal interpolation filter will not work otherwise. The error occurs even when the data really should be in temporal form.
In my mind 9089 is a fairly serious issue; the executive is failing to properly use a filter that does not directly support temporal data. Issue 6662 is just a minor inconvenience; as long as temporal data is being handled correctly, the only effect to the user should be an odd reporting of temporal data set in the information panel.
-Ken
On 6/2/09 12:57 AM, "John Biddiscombe" <biddisco at cscs.ch> wrote:
This is yet another manifestatrion of
http://public.kitware.com/mantis/view.php?id=6662
which I will try to look at this week or next.
I'm linking the bugs.
JB
Re: [Paraview] Problem with temporal filter and multi-block data That clarified it. I am able to reproduce the problem now, and I see that it is in fact a bug. I submitted a report for this.
http://www.paraview.org/Bug/view.php?id=9089
Thanks for bringing this to our attention.
-Ken
On 5/31/09 1:07 AM, "Christian Wellmann" <wellmann at ikm.uni-hannover.de> wrote:
Hi Ken,
the order of operations is:
-load the data
-attach the ExtractBlock to the reader
-attach my temporal filter to the ExtractBlock filter
If my data doesn't consists of multiple parts, so I don't have to use
the ExtractBlock, my temporal filter works fine.
christian
2009/5/29 Moreland, Kenneth <kmorel at sandia.gov>:
> I am confused about the order of operations you are doing to get the error
> (since it apparently only happens if you do things in a certain way). Are
> you loading the data, attaching your temporal filter to the reader, and then
> attaching the extract block to your temporal filter? Or are you loading the
> data, attaching extract block to the reader, and then attaching your
> temporal filter to extract block? Or are you loading the data, attaching
> the temporal filter to the reader, and then attaching extract block to the
> reader too?
>
> -Ken
>
>
> On 5/29/09 3:32 AM, "Christian Wellmann" <wellmann at ikm.uni-hannover.de>
> wrote:
>
> Hi all,
>
> I use paraview to view results of transient combined finite discrete
> element analyses.
> My *.pvd input files look like this:
>
> <?xml version="1.0"?>
> <VTKFile type="Collection" version="0.1" byte_order="LittleEndian">
> <Collection>
> <DataSet timestep="0,0000000000e+00" group="0" part="0"
> file="se_relax_p_000000000.vtp"/>
> <DataSet timestep="0,0000000000e+00" group="0" part="1"
> file="se_relax_c_000000000.vtp"/>
> <DataSet timestep="0,0000000000e+00" group="0" part="2"
> file="se_relax_r_000000000.vtp"/>
> <DataSet timestep="0,0000000000e+00" group="0" part="3"
> file="se_relax_f_000000000.vtu"/>
> <DataSet timestep="1,0000000000e-03" group="0" part="0"
> file="se_relax_p_000002000.vtp"/>
> <DataSet timestep="1,0000000000e-03" group="0" part="1"
> file="se_relax_c_000002000.vtp"/>
> .
> .
> .
>
> The results are stored for each part (DE, contact, rigid-boundaries,
> FE) and time step in a single *.vtp / *.vtu file.
> I can load the *.pvd file into paraview apply non-temporal filters to
> the single parts and animate without problems.
> Now I wrote a temporal filter to show the trajectories of data points.
> It requires a time series of datasets
> with constant number of points and generates the trajectories of the
> points as a single PolyData.
> The filter works fine if my input does not consist of multiple parts:
>
> <?xml version="1.0"?>
> <VTKFile type="Collection" version="0.1" byte_order="LittleEndian">
> <Collection>
> <DataSet timestep="0,0000000000e+00" group="0" part="0"
> file="se_relax_p_000000000.vtp"/>
> <DataSet timestep="1,0000000000e-03" group="0" part="0"
> file="se_relax_p_000002000.vtp"/>
> <DataSet timestep="2,0000000000e-03" group="0" part="0"
> file="se_relax_p_000004000.vtp"/>
> .
> .
> .
>
> But if I load my complete data with multiple parts and apply the
> ExtractBlock filter first to show the
> trajectories for one part only, ParaView complains:
>
> ERROR: In
> /home/wellmann/program/ParaView-3.4.0/VTK/Filtering/vtkDemandDrivenPipeline.cxx,
> line 822
> vtkCompositeDataPipeline (0x1e6bbc0): Input for connection index 0 on
> input port index 0 for algorithm vtkExtractBlock(0x11f4ddc0) is of
> type vtkTemporalDataSet, but a vtkMultiBlockDataSet is required.
>
> In writing the filter I tried to follow the "Time Dependent Processing
> in a Parallel Pipeline Architecture" paper
> and was geared to the vtkTemporalInterpolation filter. However I am
> not confident about how this temporal
> stuff works out: In my understanding my filter tells the upstream part
> of the pipeline to loop over the requested
> time steps and join the data into a TemporalDataSet that my
> vtkTrajectories filter can work with. Indeed if I connect
> any non temporal filter to the upstream pipeline part (instead of the
> vtkTrajectories filter) I can animate over the time-steps
> so the upstream pipeline is able to provide data sets for each time
> step. Furthermore, as said above, the filter works
> fine if there is no multiblock data as source of the pipeline.
>
> Thanks for any hints, Christian
> (the filter files are attached)
>
> --
> Christian Wellmann
> Institute of Continuum Mechanics
> Leibniz Universitaet Hannover
> Appelstr. 11
> 30167 Hannover, Germany
> phone: +49 511 762 2285
> fax: +49 511 762 5496
> email: wellmann at ikm.uni-hannover.de
>
>
>
> **** Kenneth Moreland
> *** Sandia National Laboratories
> ***********
> *** *** *** email: kmorel at sandia.gov
> ** *** ** phone: (505) 844-8919
> *** web: http://www.cs.unm.edu/~kmorel <http://www.cs.unm.edu/%7Ekmorel> <http://www.cs.unm.edu/%7Ekmorel>
>
>
--
Christian Wellmann
Institute of Continuum Mechanics
Leibniz Universitaet Hannover
Appelstr. 11
30167 Hannover, Germany
phone: +49 511 762 2285
fax: +49 511 762 5496
email: wellmann at ikm.uni-hannover.de
**** Kenneth Moreland
*** Sandia National Laboratories
***********
*** *** *** email: kmorel at sandia.gov
** *** ** phone: (505) 844-8919
*** web: http://www.cs.unm.edu/~kmorel <http://www.cs.unm.edu/%7Ekmorel> <http://www.cs.unm.edu/%7Ekmorel>
________________________________
_______________________________________________
Powered by www.kitware.com <http://www.kitware.com> <http://www.kitware.com>
Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView
Follow this link to subscribe/unsubscribe:
http://www.paraview.org/mailman/listinfo/paraview
**** Kenneth Moreland
*** Sandia National Laboratories
***********
*** *** *** email: kmorel at sandia.gov
** *** ** phone: (505) 844-8919
*** web: http://www.cs.unm.edu/~kmorel <http://www.cs.unm.edu/%7Ekmorel>
**** Kenneth Moreland
*** Sandia National Laboratories
***********
*** *** *** email: kmorel at sandia.gov
** *** ** phone: (505) 844-8919
*** web: http://www.cs.unm.edu/~kmorel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20090602/9a03fe1b/attachment-0001.htm>
More information about the ParaView
mailing list