Dieter Weber
Dieter Weber
I'm putting this in the backlog for now since I am not aware of anyone having a real-world issue with this. :-)
Related to #1094?
/azp run libertem.libertem-data
/azp run libertem.libertem-data
@sk1p test failure in notebooks was due to [OSError: [Errno 28] No space left on device](https://dev.azure.com/LiberTEM/LiberTEM/_build/results?buildId=1932&view=logs&j=4d6f0f97-9d62-53bc-79d2-96c85fc226be&t=b5ec37f2-4755-55f4-9304-bcf3d588aca8&l=370). On Moellenstedt it looks normal, how are your agents?
I'm re-running the notebook tests, let's see!
/azp run libertem.libertem-data
> Possible bug? A "wart" since variance is always non-negative real, but better to avoid the unsafe cast and make it explicit!
Still reproducible on Windows with current `node` and Firefox 91.11.0esr (64-bit)
@matbryan52 interesting idea to use sparse ROIs! This should make quite a difference for frame picking. The scipy.sparse arrays are currently limited to two dimensions, which means they won't work...