New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Release v1.6 #2251
Comments
I think the support for non linear axis #1927 is a clear candidate for v1.6. I'll be happy to contribute to the implementation. |
The following
It would also solve the issue that 'function_nd' is not defined for some of these functions, see #2225 |
Good point, indeed the best way to fix #2225 is to implement as many components as possible using
|
I hope #2035 fits the brief of "streamline the syntax" and makes it in |
Could we get #1497 in here too please? I don't really understand how to solve the gui related bits. Picking up our other conversation about electron diffraction packages - one way to solve issues emerging in the hyperspy community is to upstream this kind of core functionality that's not really got anything to do with diffraction into hyperspy. |
Thanks, I've just added it to the list. For those who don't know which conversation @dnjohnstone is talking about pyxem/pyxem#409 |
Hi Everyone, I happen to have already implemented the Voigt lineshape as an Expression for a personal project. It is identical to the original function in hyperspy, but without the parameter estimation and PES specific parameters. You can find the code using this link. I'm a little out of the loop lately, but I'll be happy to contribute this to hyperspy if its interesting! Best regards, |
@AEljarrat Check your link? :p |
Fixed the link. Thanks @thomasaarholt |
@francisco-dlp I also have a large-ish PR to come focusing on #1159 and the rest of the decomposition stuff. Hoping to have it up by the weekend/early next week. Would be nice for that to make it in. |
I think that trying to keep up with the plan of making v1.6 the last release of the 1 series is counterproductive because we are not anywhere near the target of merging most of the open pull requests, while many PRs has been merged since the latest minor release. So, what about releasing in the next few weeks v1.6? Which PRs could we realistically merge in that time span so that they make it into the release? |
I agree. I would add a couple of low hanging fruit:
Shall we include the non linear axis branch in the next release too? I think it would be more rather than making a separate beta release. |
+1 for including the non linear axis branch (there were a few more fixes necessary though) |
I suggest we fix a date for the release, otherwise we'll be tempted to keep on delaying it until this or that feature is ready. What about the 15th of May with a feature freeze on the 13th of May? I agree that it will be good to add the non-linear axes feature if ready. The already implemented features not too far off. There are missing ones that we could leave for later, since the partial support that is already implemented is a big step forward. In my mind, the most important things to get right for this release are: i) saving to file ii) the syntax. |
This sounds a very good plan to me! For the syntax side of non linear axis, it would be good to address #2345 at the same time. |
#2262 should be basically ready to be included, has a first review, but might need another one. |
#2383 is quite big but I think would be a worthy addition - it improves a lot plus fixes a fair few things that never quite worked. |
I suggest delaying the release 1 week. I've been busier than expected with other tasks and I bet that the same is true for several of us. |
In a current discussion in #2299 some of us are considering further delaying this release until some remaining key issue with the non-uniform axes features are addressed. Are there any argument against further delaying this release? |
@francisco-dlp how many other significant things have been added? 25 days ago you wrote:
That would be a good thing to compare against waiting for the non-uniform axes. |
That's definitely true. Which are the advantages/disadvantages of delaying the release say by ~1 month? Advantages from #2299, mainly pointed out by @jlaehne :
Can somebody think of disadvantages, that can outweight the benefits above? In particular, are there people out there who will be badly impacted by such a delay? |
Asked a different way, what "content" will be provided with 1.6? Perhaps we should write a draft of the release notes (would have to be done anyway with either release date). That would make it clearer what the major pros/cons of delaying are. |
For that we could have a look at https://github.com/hyperspy/hyperspy/milestone/37?closed=1 . Of course updating CHANGES.rst would be better. I've assigned the Rnp PRs to v1.6 because I thought that there wasn't going to be a v1.5.3. However, since we are still discussing when to release v1.6, if we finally go for a delay, we could consider releasing v1.5.3 if identify PRs merged into Rnp that justify releasing asap. |
Looking at the checklist on the top of this page, other mentions in this thread and the open issues in the milestones (which partly overlap of course), there are a couple of other things that might make it into v1.6 if we delay by a month. |
A cursory glance, some key things already in are absorption correction, serpentine scan pattern, decomposition overhaul (+lots of additions to docs), performance improvements to FFTs across the board, threading control for parallel map, and many other doc improvements. |
I think that there is no question that there are way enough important PRs merged since v1.5 to justify a new release. The main issue is, should we wait 1 month for the non-uniform axes and other PRs to be included in this release? I think that it is worth it unless some communities out there are itching for some of the features already merged and will be badly impacted by the delay. |
@ericpre, is there any reason to keep the freetype < 2.10 constraint for the tests? |
|
Why waiting for using |
Yes, it is one option, |
I have just tagged some of the items in the v1.6 milestone with the |
I think that it is time to start planning the next minor release that is going to be the last one of the 1 series i.e. the last one before the split.
The split will make it more difficult to contribute code that you have started developing based on the current HyperSpy structure because chances are that at least part of your code will then belong to a different package. Therefore, I suggest we focus our efforts on finishing and merging as many pending contributions as possible, rather than in making brand new contributions.
Since we will be breaking the API, this is also a good time to streamline the syntax by e.g. pruning methods are no longer relevant, changing the default value of arguments when that leads to a better outcome, renaming functions etc. All those API changes should come with a deprecation warning in v1.6 so that users have time to adapt their scripts. I propose to list and discuss in this issue those API changes in a checklist.
PRs
Support for non-uniform axes #1927New: Linear fitting 2.0 #1462Multivariate curve resolution #2172The text was updated successfully, but these errors were encountered: