John Omotani

I did implement creating-in-memory in #132. I thought it was being used in most places (there's a save-to-disk argument to the input-creation functions that defaults to *not* writing). It's possible...

I had an idea for caching the input data, to avoid re-creating it. Working on it now, PR hopefully coming soon...

Looking at how long different tests take, the biggest chunks of time are the parallel interpolation tests (which you'd expect to be computationally heavy because they have four FFT transforms...

hmm... wonder if we actually need to deep-copy at all? I suspect the idea at the time was that it would be safer, but maybe it's not needed...

The `DataArray`s in attrs were in the 'regions'. They were actually just a few scalars, so probably makes sense to switch stick in `.values` to have them as just numbers....

Would be interesting to see if this is improved by the latest xarray release - https://github.com/pydata/xarray/pull/7824 may help things.

> Might be better to just move all the tests to a top-level directory so they're not installed at all? I think having tests installed is a feature - it...

> Ah, I didn't realise people tested installed packages! When developing, I usually do an editable install into a virtual env, which does test the package as it would be...

There isn't. Looking at it again, maybe we should remove the warning and trust BOUT++ that if ZMAX was set by the user then dz=2*pi*(ZMAX-ZMIN)/nz. If I remember right, I...

For a method, in BoutDataset add something like (?) ``` """ docstring... """ def poincare_plot( self, rz, yind=0, num=100, early_exit="warn", direction="forward", new_fig=True, show=False, save_as=None ): ds = self.data mytracer =...