[7islands]

Hi,
I just noticed on 3.3-CVS newer than 12th May it is now possible to run the Surface Flow filter after running the Merge Blocks filter over the reader output (Merge Blocks works as a converter from the reader output data type to the data type Surface Flow accepts). However the calculated value of Surface Flow should be taken only as a rough indication. Using calcMassFlow is recommended for accuracy.

Takuya

[mgz1985]

hi takuya..

in response to one of your post

You can use ParaView 3.2.1. In that case you have to add

<Reader name="OpenFOAMReader"
extensions="foam"
file_description="OpenFOAM Files">
</reader>
to somewhere in ParaView3.2.1/Qt/Components/Resources/XML/ParaViewReaders.xml.

I would like to ask an alternative to the above solutionas. i cant locate this directory or file in my version of paraview which is 3.2.1. i had downloaded a diskimage and have it installed on m Mac OSX. everything works with openfoam but when it comes to paraFoam command i get an error for the paraview command not found.

I have tried a lot to debug it by whatever i have on the internet, it does not seem to be working. could you please help me

[podallaire]

Hey Takuya,

I wanted to know if the new PV3FoamReader libs (in OF-1.5) are derived from your patch... ?

I'm still using your patch with the latest cvs version and I find it very good, really stable (good work Takuya) ! However, I tried the new PV3FoamReader but it crashes really easily, not sure what is the problem.

Best regards,

PO

[7islands]

Hi Pierre-Olivier,

I'm so glad you noticed the stability of the native reader. Indeed the fact that PVFoamReader (for ParaView2 at that time) was basically made to crash even with a minor error was one of the good reasons I decided to go to the "native" way. And please don't overlook the contributions of the community, most notably a great amount of help I got from Philippose in testing the reader in its early development stage.

As to your question, the native reader and PV3FoamReader are quite different in philosophy and implementation so I guess (just a guess, have no clue) there was almost nothing they could derive from the native reader.

Takuya

[olwi]

Hi,

Question for Takuya: Is your native reader part of the official paraview cvs, or will the manual patching still be necessary?

Congratulations on your good work.

Best regards,

Ola

[deepblue17]

Hello,

I've patched the ParaView-3.3_pre20080514 package from gentoo Linux with the vtkOpenFOAMReader-20080324.tar.bz2. Compilation was fine. But when I tried to open the cavity case, which was computed with OpenFOAM-1.5, I got the following error message:

ERROR: In /var/tmp/portage/sci-cae/paraview-3.3_pre20080514/work/ParaView3/VTK/IO/vtkOpenF OAMReader.cxx, line 3851
vtkOpenFOAMReader (0x136e6e0): Numbers of faces in owner and neighbor don't match

When I compared the cavity/constant/polyMesh/neighbour file between OF-1.4.1 and OF-1.5, then there are some major differences. It seems to me, that the blockMesh utility has had some changes between the versions.

When I'm trying the native Openfoam Reader in ParaView, then the Paraview GUI gets frozen, without any error messages.

Can somebody repoduce this?

Best regards,
Oliver Borm

[podallaire]

Yes, I confirm, I have the same here.

PO

[7islands]

That's a known issue. I'm working on this (plus some other new extensions introduced in OF 1.5) and will post a new version soon.

@Ola: you still have to do patching, but making the reader to be included into the official paraview cvs is getting higher priority in my to-do list.

Takuya

[podallaire]

Thanks Takuya - that's a good news !

Furthermore, I realized today that, for some reasons, volume rendering is not available in the Paraview version provided in ThirdParty. I really prefer to download fresh cvs versions and to apply your patch + commenting some lines starting at # 3651 in vtkOpenFOAMreader.cxx.

Best regards,

PO

[gschaider]

Hi PO!

I beg to differ: Volume rendering works both on my Mac and my Linux-machines AND it is the first version of paraview where Volume-rendering does not crash paraview on a regular basis (for me)

Bernhard

[podallaire]

That's strange - I compiled yesterday the Paraview3.3-cvs version with PV3FoamReader on both Mac and Linux and could not have the volume rendering option. After seeing that, I download the latest CVS version, applied the native reader patch and recompiled again on both architecture; volume rendering was there.

PO

[gschaider]

It's a bit strange: if I select ONLY the "Internal Mesh" during the initial "Apply" I can do volume rendering. If I select patches as well I can't. Selecting patches afterwards leaves the rendering capability intact. In some very weird way this makes sense

Bernhard

[7islands]

Hi all,
Sorry to have kept you all long - here's a new version of the reader.

This time the main focus of the update was directed to supporting the new extended OF 1.5 file format, resolving cell-to-point filter issues in several corner cases and small usability improvements. Although they may sound rather unimpressive and almost unnoticeable from userland, with increasing number of real productive users I gave them the highest priority.

The reader can coexist with PV3FoamReader side by side. But as usual, please have a careful look through ReleaseNotes and proceed if you think it's worth trying.

This time the packages comes in two flavors, but the reader codes themselves are identical. One is for manual installation procedure as has been used from previous versions vtkOpenFOAMReader-20080728.tar.bz2 and another comes with an automated pre-installation script contributed by Mark Olesen. vtkOpenFOAMReader-scriptedPreInstall-20080728.tar.bz2 as far as I tried, the script works at least for the ThirdParty 3.3-CVS version of ParaView installed under default settings, but not so extensively tested. If it works for you as well, the installation procedure is as simple as
<pre> ./preInstall.sh; buildParaView3.3-cvs 1</pre>and would take only a minute or two if you have built your own binary of the ThirdParty ParaView.

My thanks go to Mark Olesen and Pierre-Olivier Dallaire for helping me a lot with testings and so on.

Hope this is useful to you,
Takuya

[podallaire]

You'll see - this reader is really fast and robust.
Thanks Takuya, it's a very important contribution.

Best regards,

PO

[johannes]

Hi,

thank you very much for the new version of the reader. I was using the former version till now and was very pleased with it.

As I'm currently postprocessing multi region meshes, I've a few questions regarding a quick 'hack' to load a second mesh region until the reader supports this feature itself in a future release.

The solver I'm working with is placing all the files of the second mesh region in a subdirectory within the main case, e.g. &hellip;/system/secondMesh/controlDict instead of &hellip;/system/controlDict, the same for the time and mesh directories.

What I'm thinking about is a duplicate of the original OpenFOAMReader which reads this subdirectory. Ideally, the name of the subdirectory could be changed without the need to recompile the reader.

Unfortunately I even did not succeed in getting a second unmodified OpenFOAMReader to work as ParaView 3.3.0 crashes when trying to open a case with it.
I created copies of ../Paraview3/VTK/IO/vtkOpenFOAMReader.cxx and *.h, renamed them and modified /ParaView3/Servers/ServerManager/Resources/readers.xml and /ParaView3/Qt/Components/Resources/XML/ParaViewReaders.xml accordingly. Recompilation showed no errors. Which parts of the ParaView source do I else have do modify?

Thanks in advance,

Johannes

[7islands]

Hi Johannes,
I guess the only missing piece is to add your renamed reader to ParaView3/VTK/IO/CMakeLists.txt. And you may have to start over from running CMake.

Takuya

[johannes]

Hi Takuya,

thanks for your hint, this entry in CMakeLists was missing indeed.
Unfortunately I'm now getting compilation errors caused by the renamed reader. I'll post them here if can't get them sortetd out soon.

Best regards,
Johannes

[richard]

Hi,

I've been having some problems with ParaView - after getting Qt4.3.4 and recompiling as in the readme I can now open paraView and it wont crash.
However it's not displaying anything when I try to load case data. It uses cpu time as if it's loading data and the time directories are shown in the top right scroll bar but the window remains black and the display tab empty..
Help? I'm a pretty well stuck atm..

Thanks,
Richard.

[7islands]

Hi Richard,
Assuming you are using PV3FoamReader, did you select at least one mesh region in the reader panel?

Takuya

[richard]

Hi Takuya,

Yes, PV3FoamReader and I tried loading in all mesh regions.. I'm using Ubuntu 8.04 on that machine, and I've since tried on suse 10.2 and after the same procedures it worked fine (I can see it's worth the effort, it's fantastic and really fast!).

I still want it to work on Ubuntu, could I be missing a graphics package or something..?