Håkon Wiik Ånes

> Hi @onatlandsmyr, this sounds like an excellent start. My only thought is that in general you will always want to flatten 2D images at some stage (for speed reasons)...

So based on discussions (off GitHub) yesterday, we agreed to not set up this "matching" package now, but wait for when other work has progressed in pyxem and the design...

the paper referenced in this comment https://github.com/pyxem/diffsims/pull/137#issuecomment-731808228 is relevant for the potential "matching/indexing" package, i think, since, as morawiec emphasises in the paper, the indexing procedure is "technique agnostic". note...

Yes, please go ahead and copy the deprecation utilities and tests, keeping the links to Matplotlib and scikit-image, where we originally took it from in kikuchipy/orix (orix' utilities were modified...

This is a good idea. We [instruct pytest](https://github.com/pyxem/kikuchipy/blob/develop/kikuchipy/conftest.py#L75) to download any external data to the local cache (on a contributor's machine or any CI machine) before running any tests, so...

Thanks for taking the effort to improve the use of Dask in HyperSpy and starting this discussion! > The lazy and non-lazy map functions should be identical using dask.array.from_array Without...

We use `mpl_toolkits.axes_grid1.anchored_artists.AnchoredSizeBar` (https://matplotlib.org/3.1.0/api/_as_gen/mpl_toolkits.axes_grid1.anchored_artists.AnchoredSizeBar.html) in `orix` (https://github.com/pyxem/orix/blob/master/orix/plot/crystal_map_plot.py#L207).

I can indeed recommend pooch for larger files! Following what scikit-image are doing, I found it easy to implement a [kikuchipy.data](https://kikuchipy.org/en/latest/reference.html#module-kikuchipy.data) module with some datasets packaged with kikuchipy, while some...

This looks interesting, although I haven't used the `DiffractionVectors` class in pyxem to comment on the benefit of the proposed class. My use of decomposition has mainly been on denoising...

Thanks for pinging me, @CSSFrancis. I'm interested in this class, although I don't have any use for it myself at the moment. > The thought process behind this is because...