Description | The basic problem here is as follows. To create a coprocessing visualization script with Catalyst, you are using the coprocessing plugin. The data item in question is a vector variable, such as displacement. You create the coprocessing script by loading an exodus file which is a stand-in for the data which will be generated during the simulation.
So: you bring up paraview, load the coprocessing plugin, and load the stand-in exodus data. You make a simple view which involves the vector variable (e.g. color by the displacement accepting the default use of the vector magnitude for the coloring value). Then dump out a script using the coprocessing plugin.
The script will not work properly during coprocessing visualization. The reason is that a '_' has been appended the the variable name during the process (e.g. the variable is 'displ' but comes out in the script as 'displ_').
A workaround is available, which is to go through the python script and change all the offending strings by removing the '_' (e.g. 'displ_' to 'displ'). However, given the already high bar of using paraview to generate the script, having to go and edit the python before use is a bridge too far for almost any user.
This appears to be somewhat related to to exodus reader. I think (but I'm not exactly sure how it works) that the data in the exodus file actually appears as three separate variables, one for each vector component (e.g. displ_x, displ_y, displ_z). The exodus reader combines these into a single vector variable with a given name (displ_).
I'm not precisely sure of a solution here. Perhaps the thing to do is to allow some automatic editing of the variable names either at exodus read time or coprocessing script write-out time (e.g. optionally remove a trailing '_' from vector names in the coprocessing script output).
|