Closed philippjfr closed 8 years ago
The code looks correct and I've tested it on JupyterHub.
I've noticed that this fix really helps a lot with memory consumption: previously, the kernel would regularly run out of memory but I after applying this fix I haven't seen it happen once!
I'll go ahead and merge.
Includes two fixes to avoid loading and caching cube data. The first avoids using the
slices
method in favor of theextract
method since the former loads the data into memory. The second change uses copy, whereever the code tries to access the cube data. This should address #16.