Mailman 3 - yt-dev - python.org

Dropping support for Python 3.8 ?
by Clément Robert 07 Sep '23

07 Sep '23

Hi all, It is that time of the year again; I would like to ask you guys how you feel about ending support for Python 3.8 soon. Here’s a bit of context: - In 2021 and 2022, we’ve dropped 3.6 (resp 3.7) about the same time we added support for 3.10 (reps 3.11) - Python 3.12 is coming soon (currently scheduled for early October), and we’re already working on adding support for it. - this year is special, because numpy is preparing a new major release, which is basically unprecedented. We already know that supporting the new version (2.0) will take extra effort and that it won’t be compatible with Python 3.8, so if we don’t drop it by then, it’ll probably make the task more difficult than it could be. - yt 4.3.0 is currently aimed at late September, so that would be the earliest we could make a release without 3.8 support. There used to be a major blocker in the way, in that our tests on Jenkins (which is were real example datasets are stored) were running on Python 3.8 and upgrading was not trivial. Kacper recently overcame this problem and managed to complete the upgrade (kudos to him !) so that’s why I’m asking you now. Please engage on the following GitHub issue if you have strong opinions or questions. https://github.com/yt-project/yt/issues/4302 [4302.png] [TST: how to drop CPython 3.8 ? · Issue #4302 · yt-project/yt](https://github.com/yt-project/yt/issues/4302) [github.com](https://github.com/yt-project/yt/issues/4302) Cheers ! Clément

2 2

What's up in yt for June 2023
by Matthew Turk 08 Jun '23

08 Jun '23

Hi everyone! We're trying something new -- there's a lot going on in the yt community, and we wanted to tell you about some of it. ## yt 4.2 As you may have seen recently, yt 4.2 was released at the end of May: https://github.com/yt-project/yt/releases/tag/yt-4.2.0 Many, *many* thanks to Clément Robert for pushing forward so much of this release. It includes a number of enhancements, but perhaps one of the most important is a major improvement to how yt stores and handles index files for particle datasets. This work was a collaboration including John ZuHone and Kacper Kowalik. We also want to take this opportunity to remind folks that yt development can be community-driven -- which means sharing with us feedback, thoughts on what can be improved, and even chipping in to help out with code, documentation, and even just talking about what you do with yt on Slack or over email -- see the bottom of this message for more! We have a list of fun, introductory issues for new contributors: https://github.com/yt-project/yt/labels/new%20contributor%20friendly and some we're looking for help with: https://github.com/yt-project/yt/labels/help%20wanted ## Workshop Presentations In April, we had an informal online workshop. The videos from that are now online and we invite you to check them out -- they cover an introduction to using yt and some of the basics of how to get started analyzing data. Day One: https://www.youtube.com/watch?v=h6NE32WagHs Day Two: https://www.youtube.com/watch?v=fn45FK7hX7Q Matt also streams development once in a while at twitch.tv/powersoffour , but be sure to manage your expectations on that front! ## Widgyts for Jupyter and yt One of the yt family of projects, Widgyts (https://widgyts.readthedocs.org/ and github.com/yt-project/widgyts ) just had another release in early April. Widgyts, as the *extremely* clever name suggests, is widgets for yt within Jupyter and Jupyterlab. The new version includes a mesh viewer for finite element meshes, as well as preliminary support for displaying particles in the viewer. If you're using AMR datasets or FEM, check it out! ## Libyt Shin-Rong Tsai and Hsi-Yu Schive are in the final stages of preparing the library 'libyt' for release, with initial connections to the GAMER code and Enzo. libyt is an in-situ analysis library that provides the ability to execute analysis and visualization during the course of large-scale computations. Look for more information soon, but the code is currently available at github.com/calab-ntu/libyt , and Shin-Rong will be giving a talk at Scipy 2023 about libyt. ## Scipy 2023 At Scipy 2023 this year, there will be two presentations about yt or yt-related projects. The first, from Chris Havlin, will be on July 13 at 3PM in Grand Salon C, in the "Earth, Ocean, Geo, and Atmospheric" track, on `yt_xarray`: https://cfp.scipy.org/2023/talk/QYHD3G/ The second, from Shin-Rong Tsai, will be on July 12 at 1:55PM in Grand Salon C, in the "Astronomy and Physics" track, on `libyt`: https://cfp.scipy.org/2023/talk/JTXC9W/ If you're at SciPy, come see them and say hi! Usually all the talks are recorded, and we'll be sure to include them in a future newsletter. ## Slack As a reminder, we have a Slack community where you can ask questions, suggest things for development, or just generally chat. You can join by visiting https://yt-project.org/slack.html . ## Tell Us About Your Stuff! We'd love to share news of your papers in our next newsletter -- so please reach out to matthewturk(a)gmail.com to share when you publish a paper that uses yt you'd like to have shared out. Even if you don't want us to include it in a newsletter, we still like to hear about them! And we would also *love* to have new blog posts for https://blog.yt-project.org/ , which Madicken Munk set up with a really cool autodeployment system -- you just need to make an issue in the repository (https://github.com/yt-project/blog ) to contribute an article! (for more info: https://yt-project.github.io/blog/posts/contributing/ ) That's all for now, folks -- happy computing!

1 0

Splitting ewah_bool_utils
by Matthew Turk 08 May '23

08 May '23

Hi folks, Just a heads up that we're going to merge https://github.com/yt-project/yt/pull/2711 soon. This takes some code that was yt specific and moves it to an external package (where perhaps it can find use outside of yt). The only impact this will have is if you are accustomed to installing yt from source, where you will now need to also have this installed. If you are used to using yt from conda, pip, etc, then the dependency resolution will pick up the change and install it. *But*, if you are installing yt from source, you'll also need to install 'ewah_bool_utils' which is located at: https://github.com/yt-project/ewah_bool_utils , as well as on conda-forge, pypi, etc. -Matt

1 0

ANN: Virtual short-form course on using yt the week of April 10, 2023
by Matthew Turk 11 Apr '23

11 Apr '23

Hi all, I'm writing to announce that the week of April 10, we'll have a short-form course streamed online on how to use yt. It will be live broadcast through Twitch and later uploaded to YouTube for on-demand viewing. The presentations will be staggered across several different times-of-day to accommodate viewers in different time zones and will initially cover materials drawn from the "yt Quickstart" notebooks, building to an in-depth exploration of a dataset. Each day will include time for Q&A and interactive discussions. The times for the workshop will be: - Monday, April 10: 12:00-14:00 CDT / 17:00-19:00 UTC - Tuesday the 11th: 8:30-10:30 CDT / 13:30-15:30 UTC - Wednesday the 12th, 15:00-17:00 CDT / 20:00-22:00 UTC - Thursday the 13th, 11:00-13:00 CDT / 16:00-18:00 UTC In the days leading up to the workshop I will post here with more information about joining the live stream, as well as a rough breakdown of what topics will be covered on which days. Best, Matt

1 2

Intel Cython survey
by Nathan 03 Feb '23

03 Feb '23

Hi all, Intel just put out a request for feedback about Cython and possible directions to invest development effort. Please fill it out if you have strong opinions about Cython! Forwarding this because there are a lot of cython users in this community. https://docs.google.com/forms/d/e/1FAIpQLSfL8QGbZ53Q6iGUwyLzmtPHx_7m3yEGBl3… Nathan

1 0

[ANN] yt 4.1.4 and yt_astro_analysis 1.1.2
by Clément Robert 24 Jan '23

24 Jan '23

We are happy to announce 2 releases from the yt dev team this week: - yt 4.1.4 (https://github.com/yt-project/yt/releases/tag/yt-4.1.4) - yt_astro_analysis 1.1.2 (https://github.com/yt-project/yt_astro_analysis/releases/tag/yt_astro_analy…) Both are bugfix releases with no new features, upgrading is recommended for all users. Pleaserefer to the release notes on GitHub for more detail. Clément

1 0

Heads up: splitting out EWAH code
by Matthew Turk 30 Dec '22

30 Dec '22

Hi folks, I wanted to give y'all a heads up that in #2711 ( https://github.com/yt-project/yt/pull/2711 ) we are working to split the EWAH code out into its own library, ewah_bool_utils. ewah_bool_utils is now packaged by conda and available on PyPI, so this *shouldn't* add any major issues, but it's possible that in the near term you may need to install a new package if you're running on main rather than a release. I'll follow up here when the PR goes in as well, just to make sure we're all on the same page. -Matt

1 0

ANN: yt 4.1.3 is out
by Clément Robert 20 Dec '22

20 Dec '22

yt 4.1.3 is out with an important patch for everyone who cares about particles ! see the release notes on github for details https://github.com/yt-project/yt/releases/tag/yt-4.1.3 Happy holidays Clément, on behalf of the yt team

1 0

[ANN] yt 4.1.2
by Clément Robert 19 Nov '22

19 Nov '22

Announcing yt 4.1.2 ! This is a bugfix release with no new features, all users are encouraged to upgrade. See the release notes for details https://github.com/yt-project/yt/releases/tag/yt-4.1.2 Clément, for the yt dev team

1 0

What's new in Python 3.8 (digest)
by Clément Robert 21 Oct '22

21 Oct '22

Hi all, yt 4.2 (current main branch) has officially dropped support for Python 3.7 While this is mainly done to reduce technical debt, it also means we can now start using Python 3.8+ syntax, so I wanted to highlight a couple things from the “What’s new in Python 3.8” page, to get everyone on the same page. FTR this isn’t me campaigning to modernise all the things, but it’s always nice to have access to a larger subset of what Python can do :) - Assignment expressions, aka “walrus operator” https://docs.python.org/3/whatsnew/3.8.html#assignment-expressions In some places, using `:=` makes for shorter code without sacrificing readability. I find it really shines when binding a variable that’s immediately going to be tested in an `if/else` block, for instance. - Positional-only parameters: https://docs.python.org/3/whatsnew/3.8.html#positional-only-parameters Though rarely useful, making arguments positional-only means we don’t have to commit to non-meaningful names forever (looking at you, `yt.load(fn=…)`) - “debug fstrings” https://docs.python.org/3/whatsnew/3.8.html#f-strings-support-for-self-docu… Python 3.8 makes fstrings even nicer and more useful ! - new features for the `typing` module https://docs.python.org/3/whatsnew/3.8.html#typing We now have easy access to `typing.Final` (declare that a variable should be considered constant and not be assigned to), and `typing.TypedDict`, which allows to mark dictionaries as having an exact set of keys with known types. These have no effect at runtime but are very useful in making sure the code is correct at type-check time with mypy (e.g, am I rebinding a varname that was suppposed to be constant ? Did I forget a key in my return dict, or does it have an unexpected type ?) Cheers ! Clément

1 0