Hi Prabhu

Prabhu Ramachandran wrote:
> Been a while since we exchanged emails.
Lots of new projects (and a wee issue of two new babies) has been
keeping me busy. But I'm getting back to doing some CFD/GFD diagnostics
again finally.

> Sounds very cool.  I've been down with a fever past couple of days and
> am still battling my older emails. ;-)
Nothing serious I hope!

> I did notice in the code that you loop over VTK arrays.  If you used
> tvtk you can do this using numpy arrays without the need for an
> expensive for loop in Python.
You are right - I noticed that loops like this are slow. I have replaced
two expensive operation with vtkPointDataToCellData (getting cell-wise
quantities for surface integrals). I can see why tvtk would be better -
another good reason to move to mayavi2.

>     Gerard> For those interested in the hacking aspect, it's
>     Gerard> interesting because you have the change the active scalar
>     Gerard> in a vtk dataset, so there cannot be one dataset, ie. the
>     Gerard> active scalar required for the iso-surface is different
>     Gerard> from the scalar you want to plot on the iso-surface. The
>     Gerard> only way around this appears to be to make a deep copy of
>     Gerard> the output of the contour filter, modify the active scalar
>     Gerard> of this copy and make this object the output of the
>     Gerard> filter.
> 
> I believe you can do this in another way:
> 
> In [1]: import vtk
> 
> In [2]: c = vtk.vtkContourFilter()
> 
> In [3]: c = vtk.vtkContourFilter()
> 
> In [4]: c.SetInputArrayToProcess?
Ah yes - one of the less documented/google-able features of VTK.
However, an additional problem is that I want to add arrays (the newly
computed fluxes to the vtkPolyData output of vtkContourFilter. vtk is
not so keen on filter.GetOutput().GetPointData().AddArray(flux), so it
still looks like I have to make a copy.

>     Gerard> Prabhu, after people have had a change to give feedback,
>     Gerard> are you happy with this being committed? We have started
> 
> Sure, no problem.  I don't think I'll make a new release though.
> MayaVi2 is the future.
That's fine. I have promised myself that I will not write anything new
for mayavi1.x and start working on mayavi2 instead. I'm partly motivated
to have a copy of the filter in the public domain because I refer to it
in a paper I've submitted with a colleague, which contains general
python/vtk/mayavi propaganda.

Thanks for the feedback.
Gerard.

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
MayaVi-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mayavi-users

Reply via email to