ww_bounds bug fix in setter and getter - #2275

Thanks for this fix @eepeterson
I was thinking is it worth adding a unit test to protect against this bug returning in the future.
If you are keen on a unit test then I've made a PR against your branch eepeterson#1

The code looks fine to me, just to check my understanding, is this reshape needed when using WW with multiple energy groups

@shimwell this is essentially just fixing the from_xml method of the WeightWindows class. If you read in weight windows from a settings.xml file into a python script the indexing will be wrong (transposed to what it should be). If you did happen to write them out again they will not be ordered in the way the openmc executable assumes.

Thanks Ethan this might explain why the minimal magic script @pshriwise and myself have been looking into has been struggling with improving the ww on each iteration as we read in the settings.xml after writing it each loop

Thanks Ethan this might explain why the minimal magic script @pshriwise and myself have been looking into has been struggling with improving the ww on each iteration as we read in the settings.xml after writing it each loop

I think this is fixing a read into Python from XML as opposed to fixing a read into the C++ API. I've had success generating WWs with multiple energy groups on several models with that script as-is. I think our ww gen problem may have more to do with an instantiation of a Numpy array as integers rather than floats -- a change that I have in my branch on this but not that gist I sent you. My bad!