Tutorial Refactors

[CSSFrancis]

Just thinking about the changes from last year and what we should think about changing:

Dask/ Lazy Processing

• Dask Distributed (How to use it and what does everything mean)
• File Formats --> Zarr vs HDF5
• Lazy Marker Plotting?

Data Visualization

• Plotting using (new) Markers
• Documentation: Add Examples for overlaying quad mesh. hyperspy/hyperspy#3333
• New Feature: Add sub-figure plotting hyperspy/hyperspy#3343

EELS

• New edges for fitting!

Pyxem

• Redo Tutorials based on 1.0.0 release

Rosettasciio

• General session on how to use Rosettasciio separately. @smribet maybe in the py4dstem session?

Are there any other additions? Feel free to edit my post if you want.

[smribet]

Hi Carter,

I'd be interested to chat more about Rosettasciio either during the workshop or another time. I don't think it would fit in the py4DSTEM section, since we have a lot of other material. Data IO is a pretty general topic. If it is included, perhaps it would make senes to go towards the beginning or end of the workshop as part of a larger discussion?

~Stephanie

[CSSFrancis]

I'd be interested to chat more about Rosettasciio either during the workshop or another time.

Sure, send me an email, and we can set up a meeting anytime!

I don't think it would fit in the py4DSTEM section, since we have a lot of other material. Data IO is a pretty general topic. If it is included, perhaps it would make senes to go towards the beginning or end of the workshop as part of a larger discussion?

I think that in the hyperspy section, we can quite easily talk about using Rosettasciio. For the py4dstem section, it would be nice to show something like loading a .mib or .hspy dataset into py4DSTEM using rosettasciio quickly but not go into the details. Mapping the metadata from one to the other/ having a working code example so that people have it if they need it seems useful.

If you would like, I can write up an example showing this. I think a common workflow at Diamond is to load the data using hyperspy/pyxem and then transfer it to py4dstem for certain processing (@M0hsend you could correct me on that!) so it would be usefully regaurdless.

[M0hsend]

@CSSFrancis That is true. You can see an example in this notebook here.

[magnunor]

@CSSFrancis, I added the same dask distributed as last year: #6

I have not tried running that one specifically (I tested the lazy big data one I ran last year).

[CSSFrancis]

@magnunor that should be good. I can test it a little later (probably will end up being tomorrow at this point)!