Eric Prestat

Closing as this should be dealt with in https://github.com/hyperspy/rosettasciio

This has been like this for a while and I think this is a desired behaviour. You may find inconvenient in this specific case, but for example when you are...

The behaviour of this method shouldn't behave differently for different matplotlib backend. Even when using the widget backend in the jupyter lab, I am convinced that it should close and...

@magnunor, any more thoughts on this? If you want to replot in a different output, you can simply `s.plot` in the input of the corresponding cell. IMHO, we should avoid...

Yes, I agree that we should find a better approach. Particularly, we should implement a way to _add_ widgets interactively without having to initialise it first, as it is currently...

See #1927 for current progress on this.

As it is possible to setup "Github Code Scanning" right now, would it be sensible to do that? - https://github.blog/2020-09-30-code-scanning-is-now-available - https://github.blog/2020-09-30-code-scanning-is-now-available Would it be worth doing this on `RELEASE_next_major`...

For the review, it would be good to use the analysis tool to understand how it should be fixed. Did you use https://lgtm.com/projects/g/hyperspy/hyperspy to find what needs to be changed?

I think before considering adding such feature, we should first implement adding a roi interactively, as mentioned in other PRs (for example https://github.com/hyperspy/hyperspy/pull/1785). There are a few PRs which requires...

This `mask` argument is for the navigation dimension and in your case, it seems that the navigation dimension dimension is 0 and therefore the shape is `()`. I label this...