[philippose]

Hello BastiL,

Good Evening to you!!

That almost looks like a problem with your "readers.xml" file... are you sure you have modified that file as per the instructions given by Takuya? If I am not mistaken, the original version of Paraview still contains the old OpenFOAM reader by Terry Jordan.... and it must be conflicting with the version from Takuya....

Have a nice day...!

Philippose

[bastil]

Originally Posted by philippose

That almost looks like a problem with your "readers.xml" file... are you sure you have modified that file as per the instructions given by Takuya?

Well, of course it is conflicting. I followed the instructions given here:

http://openfoamwiki.net/index.php/Ti...r#Installation

I can not find anything about a "readers.xml"-File in there - except the hint that it does not need to be patched anymore...?

Regards, BastiL

[7islands]

Hi BastiL,

The contents of the readers.xml should be included in the parallel plugin (libPOpenFOAMReaderSMPlugin.so) thus perhaps the plugin was not loaded to ParaView correctly. If you don't mind building ParaView again try another "build everything as builtin" version

PHP Code:

svn co https://of-interfaces.svn.sourceforge.net/svnroot/of-interfaces/branches/vtkPOpenFOAMReader-r23-VTKcontrib 


(the installation instruction are in the included README and INSTALLs). This also has some bugs found in the 20081005 version fixed.

Takuya

[bastil]

Originally Posted by 7islands

If you don't mind building ParaView again try another "build everything as builtin" version

PHP Code:

svn co https://of-interfaces.svn.sourceforge.net/svnroot/of-interfaces/branches/vtkPOpenFOAMReader-r23-VTKcontrib 


(the installation instruction are in the included README and INSTALLs). This also has some bugs found in the 20081005 version fixed.
Takuya

Ok fine I will try. Do I have to rebuild all of paraview? (make clean, make)? Or can I rebuild just modified as with OpenFOAM? If yes, how? Thanks.

BastiL

[7islands]

A safe bet would be to start over from the vanilla 3.6.1 source and re-run CMake from a fresh build directory (It may not be necessary to go through all the steps in reality but that would help pinpointing the causes if you had more problems).

Takuya

[bastil]

Takuya,

thanks this works fine now.

[pcaron]

Hello everybody. Regards from Argentina!

I've used OpenFOAM since January 09. I've ran OF in Ubuntu 8.04, Kubuntu 8.04 without problems. Always desktop machines.

By June, I bought a laptop: HP 550, Intel Core 2 Duo 1.8GHz, RAM:4Gb,Video: Intel GME965. I installed an Ubuntu 9.04 64bit, with several known problems with the Intel video driver. So I've waited until Ubuntu 9.10 64bit.

At the beginning everything looks fine. But sometimes the system hanged up. I thought the problem was related again with the video driver (almost the last version available) so I installed Opensuse 11.2 64bit last weekend (with the last video driver available). All partitions using ext4.

Today my system hanged up again. The difference with previous times is: I know what I have to do to hang up the system.

I'll explain the details:
Objetive: Read an OpenFOAM mesh. Just the mesh, no results. Apply opacity to the surface.

Option 1: Read the mesh using paraFoam. After read the mesh, change to Surface+Solid Color. Apply opacity=0.5. After a few seconds the system freezes. Then the only option is to force shutdown

Option 2: Read the mesh using pvFoam. Same results as Option 1.

Option 3: export the mesh with foamToVTK and read using paraview. Everything works fine, just a little slow.

I'm really confused. Could it be a problem with the reader?

I'll appreciate any help.

Best regards

Pablo

[7islands]

Hi Pablo,
Sorry but have no idea. If you can reproduce with one of the datasets that is included in ParaViewData3.6.zip, you might ask help at the ParaView list.

Takuya

[pcaron]

Hi Takuya!

Thank you for your quickly response. I'll try to reproduce the problem with the paraview data.

Have a nice day!

Pablo

[7islands]

Hi,

A version of the reader has been in ParaView 3.7-cvs and VTK 5.5-cvs since 9 Jan 2010. As far as I understand, the commit is still considered tentative as the Kitware developers are setting up a test soon and the reader still has to pass it before everything settles.

The version is based on a previous release of r64-20081005 since I had to do some specific works of addressing compilation issues on some ParaView/VTK dashboard platforms. The most notable differences from the current release of r173-20090920 are

• The version supports dictionary syntaxes up to OpenFOAM 1.5 (no support for regular expressions).
• The reader does not have the custom Qt UI (the Reload and Watch buttons) since the developer used a GPLed version of Qt 4.3 at that time.

However anyhow, the reader has all the basic features including parallelization and is expected to work at least better than "chose OpenFOAM file format from the list of the readers but then ParaView just crashed" which we were long used to. Please feel free to give it a try.

Have a nice weekend,
Takuya

[pcaron]

Hi Takuya!

I'll give it a try!

BTW, is it posible to add the "view patch names" option to the reader? When using the standard reader I can see this option. It's very useful to check the mesh.

Regards

Pablo

[7islands]

Hi Pablo,

I admit its usefulness, but sorry, currently I have no plan to implement the option because

• it requires either quite a tricky implementation which makes the reader incompatible with the client/server mode, or an exceptionally tight integration between the client side and the server side codes, and
• supporting decomposed cases requires too complicated works.

Takuya

[pcaron]

Takuya,

Thanks for your quickly response. I understand the problem.

I'll tell you if the problem persist in my laptop.

Regards.

Pablo

[shirazbj]

Hi,


I am trying to build the r173-20090920 Reader on Ubuntu 8.1 & 9.04 for paraview3.6.2. But has problem:

Follow Step 4.2.2 Building and installing the plugins

When ccmake <the top directory of the reader source tree>

setting ParaView_DIR = <the directory where ParaView was built>


I got an error saying ParaView_DIR can't be found. It asks me to locate
a file named ParaViewConfig.cmake. But I can only found one named
ParaViewConfig.cmake.in under the folder ParaView3 extracted from the source.

Where the this file?

Does this Reader need a working paraview to build?

Thanks

Cean

[7islands]

Hi Cean,

ParaViewConfig.cmake should be in the top directory of the build tree. You need to do a full build of ParaView 3.6.2 first.

Takuya

[deepsterblue]

Takuya,
Do you think you could add support for surfaceFields on your reader? I'd like to be able to visualize them like say, a pointField, or something similar. I'm hoping that this isn't too much of a problem, because boundaries for volFields are already displayed correctly, and they lie on faces.

Thanks

[olesen]

Originally Posted by deepsterblue

Takuya,
Do you think you could add support for surfaceFields on your reader? I'd like to be able to visualize them like say, a pointField, or something similar. I'm hoping that this isn't too much of a problem, because boundaries for volFields are already displayed correctly, and they lie on faces.

From what I know, supporting surface fields would require a significant rewrite of the reader.

The only way I can see representing a surface field as a point field would be to create new points corresponding to the face centroids - but this would be a reader feature entirely.
Simply extending the current volFields handling to include surface fields isn't really an option either. The boundaries faces are indeed treated as independent geometric entities in VTK and have their own values, but the internal faces do not exist independently.
If internal faces were to be added as independent entities, I would suspect that the best would be a switch in the GUI to handle that. In one position, the reader would handle surface fields, in the other position it would handle volume fields - you would then need two reader instances to display everything.

[7islands]

Hi Sandeep, Mark,

Both you two are quite right -- from pure technical standpoint, adding support for surfaceField should not be that difficult. The problems are what entity we would create in order to represent the surfaceField mesh and fields and what amount of work is required in order to implement the feature, because VTK does not have its own support for face fields. I wonder if creating a point field of face centroids is enough? Or should we create a polygonal mesh that contains all faces? (The latter is what I did in gmshFoam by the way)

Takuya

[olesen]

Originally Posted by 7islands

Or should we create a polygonal mesh that contains all faces? (The latter is what I did in gmshFoam by the way)

That is probably the best solution. If you want to avoid having two reader instances, it'd be reasonable easy to use a GUI switch to enable surfaceFields and then place the resulting vtkPolyData on a separate output port.

[7islands]

Here is what it looks like in the case of gmshFoam, and would look like (sort of) in ParaView when a clip of a polygonal surfaceField mesh is created.

(Surface with edges, emulated by Gmsh with additional element outlines)
surfaceWithEdges.png

(Wireframe, emulated by Gmsh with Element boudary mode Dimension - 1)
wireframe.jpg

Note that you will typically see different colors depending on the direction of the faces since the face fluxes quite depends on the direction of the flow. Are these representations look useful?

Takuya