Closed martinjanssens closed 2 years ago
Looks good to me :) Does it makes sense that this wasn't caught by a test?
@martinjanssens just double-checking this, should we add a test for this? Seems like this should've been caught by a test earlier if there was something amiss...
Sorry, I had forgotten about this. It makes sense that this wasn't caught by a test, as the fourier metrics are only tested for correctness on periodic domains. I've added a test that tries to just run the code for all possible inputs to make sure it doesn't crash, that should catch errors like these in the future :)
A colleague of mine found that my refactored spectrum calculation only applied windowing if
periodic_domain==True
, while we want the opposite. Have fixed this here!