(0036569)
|
Alan Scott
|
2016-07-25 19:24
|
|
From Ben:
Thanks for taking a look at this ticket, to provide some context to the request, I've attached a (fairly complex) example of the kind of plots I had in mind. These are located in the second column, you can ignore the plots in the first column and the inset.
In these plots, the vertical axis is position (along a line) and the horizontal axis is time. The color is determined by the values of a variable along this line.
At present, this figure is generated by loading exodus output into Paraview, selecting points that (more or less) fall along the line, using "Plot Selection Over Time," using "Save Data" to output these as csv files, and then doing my own post-processing in matplotlib. It would be helpful to get the same results without all of the hoop jumping.
Some desirable features:
* Choice of an arbitrary vector (ala "Plot Over Line") which automatically interpolates from nearby values.
* Some control over the temporal/spatial discretization
* Export to a interchangeable format for additional post-processing
* Custom ranges for the various axes ala "Plot Over Line"
Additionally, the axes need not always be space versus time, though I'd argue it is a good default.
A distinct but related capability would be the ability to average/integrate over a particular dimension. Say I ran a 2D simulation in x and y, but only cared about behavior in y, it would be useful to integrate/average over all (or part) of x (or theta, or r, or phi, or ...).
Sorry if this is too much information. Let me know if you have any questions, |
|