Hi all,
We’re proud to announce the first release of the yt-widgets package,
widgyts, which is a growing collection of Jupyter widgets for use in yt.
The package is on PyPI so you can install it with:
pip install widgyts
This package was designed so that you can interactively explore your data
with the goal of making data exploration more accessible to yt users. You
can pan, zoom, and update view parameters (like the colormap or the
colormap bounds) quickly and learn where interesting features of your data
are. We’ve uploaded a couple of demonstration notebooks to give you a feel
of the widget, one with IsolatedGalaxy and the others with slices and
projections (made using yt) of Britton’s Pop II Prime dataset (thanks for
making this dataset available Britton!). You can experiment and play around
with the notebooks on the yt hub
<https://girder.hub.yt/#raft/5c1ab9f5323d12000134e095>. Alternatively, you
can watch <https://www.youtube.com/watch?v=5dl_m_6T2bU> or look at the
slides <https://munkm.github.io/2018-07-13-scipy/> from Madicken’s talk at
SciPy 2018 about the initial stages of this project.
At present, widgyts consists of a few components. The foundational widget
is that of a variable-mesh image viewer, using Rust compiled to
WebAssembly. The WebAssembly code is used to do the most
performant-limiting calculations in the widget, resulting in much higher
framerates than pure javascript rendering. Because WebAssembly was designed
to interface with javascript, widget features can be built out using
standard Python/javascript aspects of Jupyter widgets. You can read a
little bit more about WebAssembly here
<https://developer.mozilla.org/en-US/docs/WebAssembly> and here
<https://medium.com/javascript-scene/what-is-webassembly-the-dawn-of-a-new...>
if
you’d like to learn more about it. Widgyts also currently has a widget that
quickly converts data arrays to rgba colormapped arrays, and a tool to set
up a standard set of widget controls.
This should work in all major browsers in the Jupyter notebook; at present
Jupyterlab does not have support for easy distribution of wasm binaries.
When able, we will build out support for this widget with Jupyterlab.
We’d like to thank the hard work done on other projects to make the
development of widgyts possible. In particular, we’d like to thank the
Jupyter team and the developers of ipywidgets. We’d also like to thank the
rust core team and the wasm-pack developers for their awesome platforms.
Thanks for reading! If you’d like to give us any feedback on the widget,
please don’t hesitate to reach out via e-mail, slack, or in either the rust
<https://github.com/data-exp-lab/rust-yt-tools> or the widgyts
<https://github.com/data-exp-lab/widgyts> repository.
Madicken Munk, Matthew Turk, and Nathanael Claussen
Hi all,
With the advent and maturation of the YTData frontend, is there any
use case we would still recommend folks use the functions
Index.save_object
Index.load_object
Index.save_data
Index.get_data
or are they candidates for deprecation and removal in yt-4.0?
-Matt
Hi all,
Bili put in a PR to reverse the orientation of yt's default angular
momentum vector to be closer to what most people expect given conventions
about right handed coordinate systems.
See https://github.com/yt-project/yt/pull/2043
I'm a little nervous about merging this since it might be disruptive. Does
anyone see any issues with merging? People on master branch should know
what they're in for?
If no one responds here or comments in the PR I'll probably merge the PR
next week.
-Nathan
Dear all,
I am just starting to get my head around yt and especially it's part on the construction of light-cones. I have written my own code so far to construct light-cones which extend beyond the box of my cosmological simulations. However I would love to see a general software which can transform any cosmological simulation box into any light-cone.
Therefore I was wondering whether anyone of you could tell me if there are any open issues which I could start working on immediately?
(Depending on how one would use the light-cone in the end I can already see some points of improvement, e.g.: interpolation scheme for positions and velocities of sub-&halos between snapshots, ...).
Cheers,
Christovis
P.S.: I also have a nice pipeline to do gravitational strong lensing I could add to yt (currently using ctypes). as far as I have seen there is not much on gravitational lensing at the moment. Into which directory should I add the code once I have cleaned it up?
