Hi team
Yt 4.0.0 was released on July the 6th of last year, and there’s been a bunch of contributions since then.
I would like to propose that we try to cut a 4.1 release before the summer break, roughly a year after 4.0
I think it’s both a reasonable goal and a good time for an important release. I anticipate no one will have time for coordinated work during summer, and the fall is traditionally not ideal either: many of us are on teaching duty, and it’s also when the Python ecosystem is least stable.
There are a couple blockers that need to be adressed to get the dev branch back to a release-ready state, namely
1) in 4.0 we promised (in the form of deprecation warnings) that in 4.1, errors would be raised from ambiguous name-only field keys. Actually implementing this poses a couple difficulties (see https://github.com/yt-project/yt/issues/3381 and https://github.com/yt-project/yt/issues/3839) but nothing insurmountable.
2) We need to reach a consensus on how the new axis flipping/swapping machinery should behave. There’s an open discussion for this here https://github.com/yt-project/yt/issues/3890
To get a broader (possibly more confusing) view of the TODO list, see the open milestone: https://github.com/yt-project/yt/milestone/17
I’ve highlighted what I think are the most crucial points with the “release critical” label.
You can help by discussing and triaging open issues and PRs to and from the milestone.
It’s also a good time to get feature PRs to the finish line.
We haven’t made any big “promises” for yt 4.2 (or nothing as significant as the ambiguous field stuff), so I’m hopeful that getting 4.1 out the door will allow us to make more frequent feature releases in the foreseeable future.
Any feedback is most welcome
Best
Clément
Hi everyone,
tl;dr: tell us how to cite your yt-using papers at
https://forms.gle/Vkvab2RXLf8oTp3z6
For the last few years, we've been working on a paper about yt-4.0,
and the updates since the last paper was published over a decade ago.
I've dragged my feet on it a little bit, but I'm hoping that over the
next few months the last bits can be written up, polished, and the
paper submitted. (See the end of this email for authorship
information on the paper!) The current "build" of the paper is here:
https://yt-project.github.io/yt-4.0-paper/
with a repository at github.com/yt-project/yt-4.0-paper/ .
The paper isn't *yet* complete... which is where this email comes in!
I've been trying to collect a somewhat representative set of
references to the papers that describe or utilize yt for different
simulation platforms. There are two types of citations that we're
trying to track down -- citations that describe the simulation
platform (i.e., the "method paper" that talks about the code) and
those that use yt to analyze data output by the simulation platform
(i.e., maybe you wrote a paper and used yt). To make this easy, I've
made up a google form:
https://forms.gle/Vkvab2RXLf8oTp3z6
If you are able to take a minute and submit a couple references to
your work or work you know about that used yt, it would be very, very
helpful!
And as for the paper itself, we're preparing an email inviting
authorship to anyone who has had changes committed into the yt
repository, but the authorship policy for the paper is by design very
inclusive and described here:
https://github.com/yt-project/yt-4.0-paper#authorship-policy . If you
have contributed, you are likely going to receive this email, but if
you are certain you would like to be included, please follow the
instructions in that link to see how to be added to the manuscript.
Thanks very much,
Matt
Hi team
- With the recent addition of Python 3.11 to the set of versions we actively support, we now have five (3.7 to 3.11).
- [NEP 29](https://numpy.org/neps/nep-0029-deprecation_policy.html) , to which yt complies, only requires supporting “at least two” different versions at any given point, and recommends dropping 3.7 after December 21 (2021 !).
- Many important dependencies (including numpy and matplotlib) are already requiring Python 3.8 or newer.
- [recent history](https://pypistats.org/packages/yt) indicates that Python 3.8 is by far the version yt is most downloaded with (via PyPI)
With all this considered, I believe it is now reasonable to propose dropping Python 3.7.
It’s probably okay that we keep providing binaries for it throughout the upcoming yt 4.1.x release series, but I would like to start requiring Python 3.8 starting with yt 4.2, assuming everyone’s okay with that.
Let me know if you have any questions or arguments against it.
If not, I’ll start working on a patch targeted to yt 4.2
Cheers,
Clément
