All extensions are shown as incompatible
gabriellima opened this issue · 13 comments
I'm using latest version of Anaconda 3 (python 3.7.1), and followed setup steps for jupyter nbextensions and nbextensions_configurator.
Tried both ways on different machines (using conda install and installing directly by pip)
No matter what I try, the configurator tab show all extensions as "incompatible". Although I can disable the checkbox, and select the extensions that I want (and they're working correctly), I'd like to understand if there's anything going on that I should fix.
I've checked .yaml files of some extensions, and most (if not all) are compatible with 4.x and 5.x versions of Jupyter.
Other packages versions that might be of help.
jupyter (1.0.0)
INSTALLED: 1.0.0 (latest)
jupyter-core (4.5.0)
INSTALLED: 4.5.0 (latest)
jupyter-client (5.3.3)
INSTALLED: 5.3.1
LATEST: 5.3.3
jupyter-contrib-nbextensions (0.5.1)
INSTALLED: 0.5.1 (latest)
uncheck "disable configuration for nbextensions without explicit compatibility (they may break your notebook environment, but can be useful to show for nbextension development)"
uncheck "disable configuration for nbextensions without explicit compatibility (they may break your notebook environment, but can be useful to show for nbextension development)"
Thank you, and for sure that will enable me to check extensions I want.
The point of this issue is to identify why nbextensions_configurator is marking every extension as "not compatible" with my environment. It used to correctly identify extensions that were not compatible with jupyter, but now it is just choosing everything as not compatible at all.
maybe you can kick the button which before of sentence 'disable configuration...' , then you will find it can use as normal. Although i did not konw what bugs will happen. but it can use now~~~
@kaitang2918 I did as you suggested and it appears that the few extensions I selected still work. Nothing seems broken. This is a good work around. But as @gabriellima says, the real issue is to find out why the extensions are marked incorrectly.
I stumbled upon this old issue because I had the same problem. I was disappointed to find no solution here, but Stackoverflow had one.
The cause (at least for me) was that the package notebook was in version 6.0 (as shown by pip freeze or jupyter --version). Since most extensions are marked compatible with 4.x, 5.x, the configurator was right to mark them as potentially incompatible.
The solution was to downgrade the package notebook: pip install "notebook<6".
@pchampin I've been struggling with this for a few days and I see you just posted today. HOW IS THIS STILL A THING? lmao
I'm about to skip all this nonsense and start using Apache Zepplin. Who knows how that will fare...
Where the information about compatibility should be specified? (my custom widget )
This appears to still be an issue - it leads to the need for quite a bit of troubleshooting of "what went wrong" when in fact it's a bug in the configurator
As a user I have to say: Every ever lasting problem with Jupyter, like this one, really makes me want to to use Apache Zeppelin instead.
There is probably so little what needs to be done, just a simple dependency check or specification in some file by author of the widget. But there is no information about where or how.
I had this problem until downgrading to notebook-5.7.10. However, some of the package were still incompatible.
3 years later and this is still an issue -- wow. Even when I disable that checkbox, those extensions do not appear in the notebook itself.
This is just fantastically complex. I was able to get the same extension (autopep8) I needed going in VSCode in 5 minutes. I never could make this work in Jupyter.
@jeffheaton do you think you are able to get ipyvolume package working in VSCode? It requires js widget to display and I cannot make it work in either VSCode or Jupyter.