Closed wangenau closed 4 days ago
Thanks @wangenau! Sorry for my late feedback since github doesn't send notification to my email anymore (weird).
No worries @hainm Thanks a lot for merging it! Do you plan on creating a new patch release for this?
Do you plan on creating a new patch release for this?
@wangenau yeah, I am working on it https://github.com/nglviewer/nglview/pull/1127
Do you plan on creating a new patch release for this?
@wangenau yeah, I am working on it #1127
@wangenau 3.1.4 is released with your change. Cheers.
@hainm Thanks a lot for your efforts!
The package now works nicely under Python 3.13.
write_html
now also works again :+1:
I also tried to get it to work in VSCodium, but only got the old error message:
Failed to load model class 'NGLModel' from module 'nglview-js-widgets'
Error: No version of module nglview-js-widgets is registered
I saw that it should work according to https://github.com/microsoft/vscode-jupyter/issues/15757 Is there something special you have to do to get it to work in VSCode?
Um I will investigate. Thanks for trying.
@wangenau weird, I've tried with python 3.1.3 in VSCode
@wangenau By the way, it seems python 3.13 is quite new so jupyter doesn't work well yet.
After pip3 install nglview==3.1.4
, I have to mamba install jupyter
again.
After doing above, please doublecheck if /python/path/to/share/jupyter/nbextensions
have the nglview-js-widgets
folder.
(Mine is in $HOME/mambaforge/envs/py3.13/share/jupyter/nbextensions
)
@hainm Now it works under Codium somehow. Thanks a lot! Maybe I was selecting the wrong virtual environment yesterday, not sure what I did wrong :sweat_smile:
Hey, currently nglview is not installable under Python 3.13. The reason is a deprecation concerning the jupyter_packaging build packaging dependency: https://github.com/jupyter/jupyter-packaging/pull/153 Simply unbounding the dependency allows the installation under 3.13 (installation tested in a Docker container, but maybe one should check if everything still works correctly). For 3.13 we need a version of at least 0.12.2, but I opted to not set
jupyter_packaging>=0.12.2
since 0.12.2 does only support Python 3.7+, while nglview supports Python 3.6+. It would be cool if we could get a new version on PyPI to make the package installable again :) Best, Wanja