[egp]

Hi everyone,

I'm using the latest snapshot of Paraview 3.3.0 on OSX 10.5.2 with the pyFoam script from Takuya.

Paraview launches from the command line (pyFoam . case) and successfully grabs system/controlDict, reads time-level 0 data, and allows visualization of time-level 0 data.

However, I can't get paraview to read any of the timesteps beyond 0. Is this an issue with the built-in OpenFOAM reader, with 3.3.0, or with me (most likely!)?

Eric

[7islands]

Hi Eric,
The reader tries to list time directories according to the description in controlDict whenever possible (an inherited feature from Terry's original code which I found sometimes useful). However it may be causing the problem in your case. Could you comment out (add // to) lines 3651-3663 of vtkOpenFOAMReader.cxx except line 3662 as follows<pre> 3651 // if(((adjustTimeStep == "off" || adjustTimeStep == "no"
3652 // || adjustTimeStep == "n" || adjustTimeStep == "false"
3653 // || adjustTimeStep == "") && writeControl == "timeStep")
3654 // || ((adjustTimeStep == "on" || adjustTimeStep == "yes"
3655 // || adjustTimeStep == "y" || adjustTimeStep == "true")
3656 // && writeControl == "adjustableRunTime"))
3657 // {
3658 // return this->ListTimeDirectoriesByControlDict(&dict);
3659 // }
3660 // else
3661 // {
3662 return this->ListTimeDirectoriesByInstances(&dict);
3663 // }</pre>, recompile (go to [your build directory]/VTK/IO and run make) and see if it works?

Takuya

[philippose]

Hello Takuya,

A Good day to you... and long time no see :-)!

I was thinking of the same thing... I am quite sure the issue lies within that snippet of code...

Maybe we should consider shifting to time step extraction via instances, but keep the controlDict extraction method as an option which can be enabled or disabled from the reader GUI within ParaView as discussed hmmmm.... couple of months ago :-)!

What say ?

Philippose

[7islands]

Hi Philippose,
You are probably right, I'm inclined to make the feature optional in the next release (if any).

Takuya

p.s. Do you still see the cell-to-point filter problem? I'm always interested in fixing it (if it's confirmed to be a problem).

[olesen]

I would favour dropping the controlDict parsing entirely since it can incorrectly ignore existing solutions. For example, when a steady-state solver uses the Time::writeAndEnd() method after reaching convergence.

[7islands]

Yep that behavior will be the default (if there will be another release).

T.

[egp]

Hi Takuya,

I have not tried to compile version 3.3.0. I am using the binary downloaded from paraview.org.

I'll try to download the 3.3.0 source and compile the code with your recommended changes to lines 3651-3663.

Eric

[7islands]

Hi Eric,
Now I got it: the reader enabled in version 3.3 by default is Terry's code, which doesn't support VCR Play/Stop/etc. buttons. You could also try (warning: advertising own stuff) the reader posted to another thread on 26/Mar/2008 [1].

[1] http://www.cfd-online.com/cgi-bin/Op...1231#POST21231

Takuya