[Paraview] problems with ServerManagerConfiguration and vtkImageReslice

Fri May 20 05:51:00 EDT 2011

On Fri, May 20, 2011 at 9:31 AM, Jérôme <jerome.velut at gmail.com> wrote:
> It is definitively not an out-of-memory problem!!
> Your data makes my paraview crashed too, but not only. I tried with a vtk
> python script (thus nothing to do with paraview) and the same occurs.
>
> The reason comes from the way your Image.vti is built: the output of the
> vtkXMLImageDataReader contains one PointData array, but scalars are not
> set. You have to write somewhere when you build your image something like
> that :
>    myImage->GetPointData()->SetScalars(
> myImage->GetPointData()->GetArray(0))

I generated the image by a custom filter operating on a generic
vtkDataSet, it just happens to be vtkImageData at this time. Your code
resolves the issue indeed, many thanks! But should the same be done
with other data types or only images? Do Scalars/Vectors always need
to be specified? These scalars/vectors/field data etc. remain unclear
for me for years.

Regards,
Dominik

>
> For a ParaView-exclusive solution, just plug a Calculator after the
> Image.vti source
> and set the operation to "FractalIterations" only. It will set properly the
> output scalars.
>
> HTH
> Jerome
>
> 2011/5/20 Dominik Szczerba <dominik at itis.ethz.ch>
>>
>> Hi,
>>
>> My data is here, I would be glad if you tried it out.
>>
>> http://otoro.itis.ethz.ch/~dominik/tmp/Image.vti
>>
>> My ParaView is several days old git.
>>
>> Thanks and regards,
>> Dominik
>>
>> On Fri, May 20, 2011 at 8:51 AM, Jérôme <jerome.velut at gmail.com> wrote:
>> > I don't have any problem running Median or Gradient on image of type
>> > double. If you are allowed to share your data, I may give a try here.
>> >
>> > Maybe are you running out of memory? You can try to extract a subset
>> > of your data and see if the segfault occurs.
>> >
>> > Jerome
>> >
>> > 2011/5/20 Dominik Szczerba <dominik at itis.ethz.ch>
>> >>
>> >> Hi Jerome,
>> >>
>> >> Thanks for the references, but there does not to be any solution in
>> >> there?
>> >> And yes, "strange extents" are also taking place in some cases, i.e.,
>> >> extents are actually correct, but the content of the image data is
>> >> smaller if the new spacing is finer than the original.
>> >> Due to the lack of detailed description of this whole mechanism I can
>> >> not say if this is my lack of knowledge or a bug.
>> >>
>> >> Regarding segfaults, the only non-standard thing I can say about my
>> >> input image data is that it contains a scalar type "double". The said
>> >> filters will work with image generated by Sources/Wavelet, but the
>> >> type of that one is float. I am afraid something is hardcoded
>> >> somewhere.
>> >>
>> >> Regards,
>> >> Dominik
>> >>
>> >> On Fri, May 20, 2011 at 8:25 AM, Jérôme <jerome.velut at gmail.com> wrote:
>> >> > Hi,
>> >> > regarding the vtkImageSlice exposition in ParaView, you should take
>> >> > a look at these threads:
>> >> > - issue report:
>> >> >    http://www.mail-archive.com/paraview@paraview.org/msg02225.html
>> >> >    http://www.mail-archive.com/paraview@paraview.org/msg08260.html
>> >> > - issue deep explanation:
>> >> >    http://www.cmake.org/pipermail/paraview/2010-June/017738.html
>> >> >
>> >> > regarding the segfault, I feel surprise ! I use Median and Gradient
>> >> > very
>> >> > often, and I never noticed a bug. Can you give the step to reproduce
>> >> > ?
>> >> >
>> >> > Best regards,
>> >> > Jerome
>> >> >
>> >> >
>> >> > 2011/5/19 Dominik Szczerba <dominik at itis.ethz.ch>
>> >> >>
>> >> >> I think my theory that it has to do with threading may be correct...
>> >> >>
>> >> >> I just tried instead of my custom plugin to run the built-in
>> >> >> Gradient
>> >> >> or Median on my vtkImageData and Paraview (3.11) silently segfaults.
>> >> >> Both these built-in filters derive from vtkThreadedImageAlgorithm.
>> >> >> Built-in vtkExtractVOI. however, works correctly, just as my other
>> >> >> custom plugins deriving from vtkImageAlgorithm.
>> >> >>
>> >> >> Any pointers in this bug hunting are appreciated.
>> >> >>
>> >> >> Dominik
>> >> >>
>> >> >> On Thu, May 19, 2011 at 9:30 PM, Dominik Szczerba
>> >> >> <dominik at itis.ethz.ch>
>> >> >> wrote:
>> >> >> > So far I was able to add a few simple custom plugins, but am
>> >> >> > slowly
>> >> >> > losing the grip.
>> >> >> > I am trying to resample an image like this:
>> >> >> >
>> >> >> > <SourceProxy name="ImageResample" class="vtkImageResample"
>> >> >> > label="ImageResample">
>> >> >> > <InputProperty name="Input" command="SetInputConnection">
>> >> >> >        <ProxyGroupDomain name="groups">
>> >> >> >                <Group name="sources"/>
>> >> >> >                <Group name="filters"/>
>> >> >> >        </ProxyGroupDomain>
>> >> >> >        <DataTypeDomain name="input_type">
>> >> >> >                <DataType value="vtkImageData"/>
>> >> >> >        </DataTypeDomain>
>> >> >> > </InputProperty>
>> >> >> > <DoubleVectorProperty name="Magnification Factor"
>> >> >> > command="SetAxisMagnificationFactor" number_of_elements="3"
>> >> >> > number_of_elements_per_command="1" repeat_command="1"
>> >> >> > use_index="1"
>> >> >> > default_values="1 1 1"/>
>> >> >> >
>> >> >> > I get these 2 problems:
>> >> >> >
>> >> >> > 1) There appear two (!) outputs (named Output-0 and Output-1). I
>> >> >> > get
>> >> >> > this message:
>> >> >> >
>> >> >> > ERROR: In
>> >> >> > /home/dsz/pack/ParaView/VTK/Filtering/vtkDemandDrivenPipeline.cxx,
>> >> >> > line 827
>> >> >> > vtkPVPostFilterExecutive (0x3aa4f60): Input for connection index 0
>> >> >> > on
>> >> >> > input port index 0 for algorithm vtkPVPostFilter(0x2cfb850) is of
>> >> >> > type
>> >> >> > vtkImageStencilData, but a vtkDataSet is required.
>> >> >> >
>> >> >> > 2) The first output has the expected Spacing, but it does not
>> >> >> > contain
>> >> >> > the same scalar array as in the input. Instead, it contains a new
>> >> >> > array ImageScalars ranging [0,0]. I get this message:
>> >> >> >
>> >> >> > Warning: In
>> >> >> >
>> >> >> >
>> >> >> > /home/dsz/pack/ParaView/ParaViewCore/VTKExtensions/vtkTexturePainter.cxx,
>> >> >> > line 179
>> >> >> > vtkTexturePainter (0x339fa70): Failed to locate selected scalars.
>> >> >> > Will
>> >> >> > use image scalars by default.
>> >> >> >
>> >> >> > I have been so far successful with other image algorithms, however
>> >> >> > they were all deriving from vtkImageAlgorithm. This one derives
>> >> >> > from
>> >> >> > vtkThreadedImageAlgorithm - could this have anything to do here?
>> >> >> > Could anyone point me into the right direction? After inspection
>> >> >> > of
>> >> >> > the xml files in the ParaView source tree I am still clueless.
>> >> >> >
>> >> >> > Thanks and regards,
>> >> >> > Dominik
>> >> >> >
>> >> >> _______________________________________________
>> >> >> Powered by 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
>> >> >
>> >> >
>> >
>> >
>
>