Bug#1064777: paraview: CGNS rind cells are not taken into account, not even in interpolation
Francesco Poli (wintermute)
invernomuto at paranoici.org
Sun Feb 25 19:52:51 GMT 2024
Package: paraview
Version: 5.11.2+dfsg-6+b1
Severity: normal
X-Debbugs-Cc: invernomuto at paranoici.org
Hi!
Thanks for maintaining paraview in Debian.
I noticed that I can load a CGNS file containing a (structured) grid
and a cell-based solution with rind cell values.
Rind cells seem to be correctly recognized: their values are not
wrongly assigned to real cells.
And that's good.
However, it seems that rind cell values are entirely skipped and
forgotten.
I haven't found a way to visualize the values in rind cells (is there
one?).
Moreover, rind cell values do not seem to be taken into account, not
even in the interpolation performed by the 'Cell Data to Point Data'
filter.
Please let me try to explain with a test case.
The attached archive 'test_cgns_cellrind.tar.xz' includes a CGNS file
('test1.cgns') and a small Fortran program to generate it.
As can be seen in the attached 'cgnsview_screenshot.png', the analysis of
the file (with cgnsview) reveals that the grid has 4x5x1 cells (2x3x1 of
which are real cells, while the other ones are rind cells) and
that the solution (which consists of the Density scalar field) is equal
to 1.0 in each real cell and 0.0 in each rind cell.
Then, load the file with paraview:
0) start paraview
$ paraview test1.cgns
1) specify "CGNS Series Reader" in the "Open Data With..." dialog window
2) check "Cell Arrays" and click on the "Apply" button
3) from the Filters menu, select Alphabetical / Cell Data to Point Data
and click on the "Apply" button
4) change the representation from "Outline" to "Surface With Edges"
5) change the coloring from "Solid Color" to "Density"
6) click on the "Split Horizontal Axis" button (above the render view)
and click on "SpreadSheet View"
7) check (both visually and in the spreadsheet) that all the density
values in the nodes are 1.0
See the attached 'paraview_screenshot.png'.
I expected that at least the interpolation (performed to obtain point data
from cell data) would take rind cell values into account, thus
producing some nodes with density in between 0.0 and 1.0, but it seems
that all the node values are obtained only by interpolating/extrapolating
real cell values, without taking rind cells into account.
I think the 'Cell Data to Point Data' filter should take rind cell values
into account, where available, in order to produce node values.
It would also be nice, if paraview allowed the user to visualize
rind cells (ideally with an on/off toggle and with the possibility
to set different visualization properties, such as opacity, colors, ...,
in order to visually distinguish rind cells from real cells).
Please forward my bug report upstream, as appropriate.
Thanks a lot for reading so far!
Bye.
-- System Information:
Debian Release: trixie/sid
APT prefers testing
APT policy: (800, 'testing'), (500, 'unstable')
Architecture: amd64 (x86_64)
Kernel: Linux 6.6.15-amd64 (SMP w/4 CPU threads; PREEMPT)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8), LANGUAGE=en_US:en
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
Versions of packages paraview depends on:
ii libadios2-mpi-c++11-2 2.9.2+dfsg1-9
ii libavcodec-extra60 [libavcodec60] 7:6.1.1-1
ii libavformat-extra60 [libavformat60] 7:6.1.1-1
ii libavutil58 7:6.1.1-1
ii libc6 2.37-15
ii libdouble-conversion3 3.3.0-1+b1
ii libexpat1 2.5.0-2+b2
ii libfreetype6 2.13.2+dfsg-1+b1
ii libgcc-s1 14-20240201-3
ii libgdal34 3.8.4+dfsg-1
ii libgl2ps1.4 1.4.2+dfsg1-2
ii libglew2.2 2.2.0-4+b1
ii libglx0 1.7.0-1
ii libgmsh4.12 4.12.1+ds1-1
ii libhdf5-openmpi-103-1 1.10.10+repack-3+b1
ii libjpeg62-turbo 1:2.1.5-2+b2
ii liblz4-1 1.9.4-1+b2
ii liblzma5 5.4.5-0.3
ii libnetcdf19 1:4.9.2-3+b1
ii libogg0 1.3.5-3
ii libopengl0 1.7.0-1
ii libopenmpi3 4.1.6-5
ii libpng16-16 1.6.42-1
ii libpython3.11 3.11.8-1
ii libpython3.12 3.12.2-1
ii libqt5core5a 5.15.10+dfsg-7
ii libqt5gui5 5.15.10+dfsg-7
ii libqt5help5 5.15.10-6
ii libqt5network5 5.15.10+dfsg-7
ii libqt5opengl5 5.15.10+dfsg-7
ii libqt5widgets5 5.15.10+dfsg-7
ii libstdc++6 14-20240201-3
ii libswscale7 7:6.1.1-1
ii libtheora0 1.1.1+dfsg.1-16.1+b1
ii libtiff6 4.5.1+git230720-4
ii libx11-6 2:1.8.7-1
ii libxcursor1 1:1.2.1-1
ii libxml2 2.9.14+dfsg-1.3+b2
ii python3-matplotlib 3.6.3-1+b2
ii python3-mpi4py 3.1.5-5
ii tcl [tclsh] 8.6.13+b1
ii zlib1g 1:1.3.dfsg-3+b1
Versions of packages paraview recommends:
ii mpi-default-bin 1.15
ii paraview-doc 5.11.2+dfsg-6
ii python3-paraview 5.11.2+dfsg-6+b1
Versions of packages paraview suggests:
pn h5utils <none>
pn hdf5-tools <none>
-- no debconf information
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test_cgns_cellrind.tar.xz
Type: application/x-xz
Size: 3368 bytes
Desc: not available
URL: <http://alioth-lists.debian.net/pipermail/debian-science-maintainers/attachments/20240225/ecd3a921/attachment-0001.xz>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cgnsview_screenshot.png
Type: image/png
Size: 40527 bytes
Desc: not available
URL: <http://alioth-lists.debian.net/pipermail/debian-science-maintainers/attachments/20240225/ecd3a921/attachment-0002.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: paraview_screenshot.png
Type: image/png
Size: 118301 bytes
Desc: not available
URL: <http://alioth-lists.debian.net/pipermail/debian-science-maintainers/attachments/20240225/ecd3a921/attachment-0003.png>
More information about the debian-science-maintainers
mailing list