[vtkusers] Question about in-memory processing

[Consuelo Mendez]

Dear Sirs, 


I am completely new to the use of vtk and tools like PARAVIEW but I was really pleased to seen that with python I can really manage to get a fast prototyping for my purposes. By one of the task I am focusing on my research I have nevertheless the following problem: 


I have a large number of simulations (> 300), each of them is also big ( up to millon grid points), I have a method that uses 1 of the simulations and calculates an appoximation of all other 299 simulations (including node variables). What I would like to do is simply load the 1 simulation in PARAVIEW and then contruct a kind of transformation (filter??) that generates the corresponding coordinates for a desired simulation and update the information in the memory according to my transformation (no need to write the corresponding file). A very fast visualization should be the result of such an update (since the transformation is very cheap), even for very big meshes if one have enough RAM. From the python side I could not realize how to make those changes and update the corresponding visualization part without going over a temporary file. I have heard from sockets, that should allow something like this, also maybe filters are somehow useful. Even for
 something like reading a compressed file and decompressing in memory have not found good information or a simple example. I have no idea where to start with and maybe more important for me is to know the amount of effort I must invest in order to get this implemented. My data are unstructured vtk files containing a triangle mesh and there also information attached at the nodes.   


I appreciate any help you can provide with respect to this topic.  

Best regards

Consuelo Mendez 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20130619/1638026b/attachment.htm>