Plot over line not working properly with vtkNonOverlappingAMR data

ParaView Support
1

Hello,

I noticed that “Plot Over Line” filter is not working properly when applied to vtkNonOverlappingAMR : the plot appears to be trucated to the very first AMR block encountered.

Is this a known limitations ? (I was not able to find a reference stating this filter is not working well with this kind of data). In such a case, would it be hard to fix ?

thank you.

2

I’m unable to reproduce using latest developement version, which version are you using ? Can you share your data ?

3

Hello Mathieu,

Testing data are generated using the following small piece of code:
https://github.com/pkestene/vtk_amr_examples

I was testing under ubuntu 22.04, using paraview 5.10.0-RC1 (the one from ubuntu).

I’ve just tested a nightly build of paraview 5.12 (ParaView-5.12.20240626-MPI-Linux-Python3.10-x86_64), things have improved but there is still a problem when I select “Sample Uniformly” the plot is still truncated while the 2 others choices are working fine for me, they display the full plot.

With 5.10.0-RC1 all choices where providing a truncated plot (i can provide a screenshot if it necessary).

I was using

4

Works fine here with ParaView 5.12.0 binary release on Linux:

image
image396×657 4.3 KB

5

No, it doesn’t.
If you selected the diagonal, you cross 2 blocks with each 4 differents values; you should have 8 steps, and x axis should go from 0 to 8*sqrt(2) = 11.3. Here your x-axis is truncated to a little bit less than 6 (that is about 11.3/2).

6

Indeed, results seems incorrect, do not hesitate to open an issue: https://gitlab.kitware.com/paraview/paraview/-/issues

1 Like
7

ok, thanks.

8

Hi @pkestene , I’m trying to reproduce this issue but your github repo is no longer available. Could you please share it again ?
Thanks!

9

hello Léon,
here are the data:
tmp.tgz (2.5 KB).
thanks.

10

Hi @pkestene,
The problem was due to a wrong handling of 2D images by VTK in some cases which vtkNonOverlappingAMR relied on. It should now be fixed in both VTK and ParaView masters (https://gitlab.kitware.com/vtk/vtk/-/merge_requests/11333)

Thank you for providing your data, it definitely helped identify the origin of the bug !
Let us know if you still encounter your problem

11

Thanks, I tested with a paraview nightly build, it works perfectly as expected.