[7islands]

Hi Richard,
The empty display tab sounds strange. Can you check your mesh is really loaded through the Information tab of Object Inspector by checking the parts marked in red in the figure below? Does the same case load on suse 10.2?


Takuya

[richard]

Hi,

Yes, the information is there on both Ubuntu and suse - e.g. the cavity tutorial has "unzoned" and "patches" data heirarchy, 1280 cells etc for both machines.
But the only thing in the ubuntu display tab is the grayed out "visible" box..

[lord_kossity]

Hi Takuya,

to continue the conversation I started per Email I will quote mine and yours...

Hello Takuya,

when I use paraview 3.3 in order to display Lagrangian data by VTK-files, I'm witnessing a difference of one (or two timesteps - not so sure about that) between the Lagrangian and Eulerian field. It certainly depends on the creation of a *_0.vtk and *_1.vtk file for the Eulerian field, where the first (and second) Lagrangian time step get matched to, resulting in the situation I described above.
Since that looks quite strange, I read the Release Notes and could find a point starting with "To separate Eulerians and Lagrangians [...]". But I can't manage to display the Lagrangian part. And yes, I created a *.foam file...

In the moment I'm helping myself out by just deleting the two .vtk files *_0.vtk and *_1.vtk leading to expected results.

Could you please explain to me a bit more in detail how to display the Lagrangians and if deleting the two vtk's is a possible way either!?

Kind regards and many thanks,
Andreas

Well, Takuya you answered the question partially, but I have to admit that the second question is quite hidden in the last sentence of the E-mail...

Could you please tell me in detail how I can display Lagrangian data by using my *.foam file. The description in the Release notes is a bit too short, at least for me. I loaded up the case an used the "Extract Block" Filter and the Glyph filter. The resulting picture was the following . But that is not the one I'm getting if I am using the way via VTK.

So am I doing here something wrong?

[7islands]

Hi all,

I posted a development snapshot of the reader on the Wiki as the package size exceeded the postable limit of the forum. This time I'm focusing on multi-region cases and parallel reading of decomposed cases. The snapshot is not meant for everyone since installation and usage are complicated and untested, but if anyone is interested, have fun with it.

In my initial testings I got typically 2.3x - 2.5x speedups in mesh loading time by 4 process parallel execution. However everything is not so rosy especially in that I often got slower rendering performance by parallel executions. As I myself am quite new to parallel usage of ParaView, I appriciate sharing your findings on parallel visualization.

http://openfoamwiki.net/index.php/Contrib_Parallelized_Native_OpenFOAM_Reader_fo r_ParaView

Takuya

[7islands]

Hi Andreas,
Which block did you extract by ExtractBlock1?

T

[lord_kossity]

Hi Takuya,

I would say I tried every combination of extraction without the desired result.
The picture above was made after extracting Block0 and its two subdirectories as well as DataSet1.

[7islands]

So the pipeline browser display in the picture above does mean nothing, right? Sorry, really sorry, but I still don't get none of the point of your question, what you want to do, nor what you want me to do...??

At least, as the very baseline of our talk, I'd like to confirm if you surely know that ParaView places the first file of each (Eulerian or Lagrangian) *.vtk file sequence at timestep 0 no matter what the beginning number of each file sequence is? i.e. I'd like to confirm if you know, if your Eulerian *.vtk starts with say Eulerian_0.vtk and Lagrangian with Lagrangina_2.vtk, ParaView draws Eulerian_0.vtk and Lagrangian_0.vtk at timestep 0. Of course foamToVTK assigns an identical sequence number to both of Eulerians and Lagrangians at an identical timestep, so you'll see each data in different timesteps plotted in a single view.

If you are trying to suggest me to write it in the release notes, I would say writing such information is completely out of scope, since the reader is not a VTK file series reader.

T

[7islands]

ParaView draws Eulerian_0.vtk and Lagrangian_0.vtk at timestep 0

Correction: ParaView draws Eulerian_0.vtk and Lagrangian_2.vtk at timestep 0

T

[lord_kossity]

Ok, let me start again.

Until now my way to display Lagragians was to use foamToVTK and read in the Eulerian and Lagrangian part separately. Like that I have been able to get pictures like this:

a row of drops injected in a pipe evaporating (Eulerian field is temperature, Lagrangian diameter).

In order to speed up post processing I want to quit using foamToVTK an read the data set directly into paraview. But using this way I am not able to create images like with foamToVTK

[7islands]

Err... OK your motivation is OK, but, I have to say I'm so sorry that I'm not an espar... Since I'm a normal human, with the lacks of your answer to my question and knowledge to your case I don't think I can say anything about your question. At least can you post the case here (send me via mail if not a postable size) and provide all the information (incl. the answer)?

T
p.s. The next contact will be at next morning in JST.

[mabinty]

dear all!!

i m working with OF 1.5.x and compiled paraview and the tPV3FoamReader module as described in the readme file. the problem i face is in connection with the multiRegionHeater tutorial (chtMultiRegionFoam solver) where i m not able to display the results. the case has 5 regions: left/rightSolid, heater, top/bottomAir, where for each a <regionname>_1.vtk file is
produced and stored in a VTK folder. i ve no problem to import the region meshes in paraview (file->open) but cannot display any quantity.
the inspector/properties menu of the different regions is empty (blank) so that non of the field variables can be chosen in the inspector/display/color by menu. the same is true for the multiRegionHeater.OpenFOAM file (even when all patches and the internal mesh is ticked in the region status).

i ve no idea whats going wrong here and i m really stuck since a while now. could anybody give me a hint? i
would greatly appreciate any comments!

thx in advance!
aram

[deepsterblue]

Takuya,

Your Reader compiles fine, but when I use pvFoam with my case, I get the following error, and paraview quits:

ERROR: In /home/smenon/OpenFOAM/linux64/ParaView-3.4.0/Servers/Common/vtkProcessModule.cxx , line 1079
vtkProcessModule (0x767330): Object type: vtkOpenFOAMReader, could not find requested method: "SetTimeStep"
or the method was called with incorrect arguments.

while processing
Message 0 = Invoke
Argument 0 = id_value {78}
Argument 1 = string_value {SetTimeStep}
Argument 2 = int32_value {0}



ERROR: In /home/smenon/OpenFOAM/linux64/ParaView-3.4.0/Servers/Common/vtkProcessModule.cxx , line 1080
vtkProcessModule (0x767330): Aborting execution for debugging purposes.


/home/smenon/bin/pvFoam: line 72: 23141 Aborted ${paraviewPath} --data="${PWD}/system/controlDict.foam"

Any idea what could be wrong? I'm using OF-1.4.1-dev, and I compiled your reader from the OpenFOAM wiki page (vtkPOpenFOAMReader-r64-20081005.tar.gz). I also compiled paraview-3.4.0 from sources, and not from the CVS.

Thanks.

[philippose]

Hello Sandeep,

It almost looks like you forgot to modify the "readers.xml" file which is found in the following path:

ParaView3\Servers\ServerManager\Resources\readers. xml

This XML file defines the actual graphical interface that Paraview provides when you try to load an OpenFOAM case.

The reader is actually a backend which interacts with Paraview (and hence the user), via the interface defined by this XML code.

The original version of the OpenFOAM settings in the XML file are not valid for the reader written by Takuya.

As part of the installation package from Takuya, you should either find a modified "readers.xml" file, or a "diff" file with only the differences, which you need to patch onto your "readers.xml" file.

Philippose

[deepsterblue]

Thanks Philippose!

The archive on the OpenFOAM wiki doesn't contain the diff file. Takuya, I hope this can be updated when you get the time.

[7islands]

Hi Sandeep, Philippose,

The newest version on the wiki (20081005) requires a plugin to be built as well as ParaView itself and loaded, instead of patching readers.xml. The instruction for building the plugin is written in the "Installation: parallelized reader" section of the wiki page. Can you make sure about it?

Takuya

[deepsterblue]

Takuya,
I'm not sure how this works exactly, but I'm using the serial reader, and that required me to patch readers.xml from an older archive of yours; for it to work. From a first glance at POpenFOAMReaderSM.xml, it seems to be a parallel-specific portion. And ReleaseNotes requires installation of the serial reader first. Any ideas?

Also, after a bit of reading, I see that mpeg support is no longer available due to a licensing issue. But the vtk-FFMPEG encoder uses a deprecated img_convert function that's been replaced by an equivalent swscale call in the new ffmpeg implementation, so that doesn't compile. Is this rectified on the CVS distribution?

[7islands]

As you guessed, in the newest version all UI interactions are handled by the parallel reader and it forwards requests to the serial reader as necessary. Thus you have to build the parallel reader plugin even if you only need the serial reader. I guess that's the most confusing part.

Patching readers.xml with a older version may (or may not) work if you indeed only need the serial reader, but at least you'll miss several useful features like case reloading.

Takuya

[podallaire]

Hi Takuya,

I did not realize that the reader had evolved so much in the last month, thanks very much for your effort ! I'm building a fresh version of ParaView with your latest version and I'm following the new instructions.

One quick question I have for you is where do you store the plugins in the paraview application , mac version ? Is there a way to include the plugin in the standalone bundle ? Is the best practice to launch paraview from the shell instead of double-clicking on the app ?

Best regards,

PO

[7islands]

Hi Pierre-Olivier,

I haven't tried by myself as I usually launch from the shell, but you can see the list of search paths from the Tools->Manage Plugins/Extensions... dialog. If there's a path inside the application bundle directory in the list, I guess that's the place where you would want to put the plugin.

Takuya

[podallaire]

Hi Takuya,

I built paraview and the new reader dylib but the application crashes when lauched (PV_PLUGIN_PATH is defined). Same result if I try to load manually the plugins.

Any idea ?

Thanks

PO