0

I was using nbextension for jupyter notebooks (not jupyter lab) extensively. Recently, I installed the newest version of jupyter (v.6) and most extensions are disabled because they are meant for v.4 or v.5.

As can be seen in this post, this problem is already 3.5 years old. I did try to use some extensions that worked in the past like "Table of Contents (2)" but at least with v.6 of jupyter it doesn't work. What are your experiences? Do I need to downgrade to v.5? Should I switch to jupyterlab?

Kab
  • 43
  • 5
  • Certainly it is not dead. It's just a lot of developers didn't bother updating the YAML associated with it but that isn't really used, see [here](https://github.com/ipython-contrib/jupyter_contrib_nbextensions/issues/1434#issuecomment-513556062). They shouldn't be disabled. I wonder if you are encountering [this issue](https://github.com/ipython-contrib/jupyter_contrib_nbextensions/issues/1635) but because you provide little detail, in your post, I cannot tell. If you are only updating now, you probably should move on to nbclassic anyway. ... – Wayne Jun 06 '23 at 15:06
  • See [here](https://blog.jupyter.org/jupyter-notebook-6-5-0-release-candidate-4c229c6dab55) about that because you probably want to better understand how things will be going forward in regards to the document-centric notebook experience. And using nbclassic is only if you want to keep using the older system for some reason. JupyterLab is approach being on version 4 and a lot of the things that come with extensions in the classic notebook interface come along with JupyterLab. Additionally, Jupyter Notebook version 7 is built on top of the components of JupyterLab ... – Wayne Jun 06 '23 at 15:10
  • but offers the document-centric experience that many still prefer. JupyterLab being drastically different and more like an IDE is sometimes overwhelming for those new to Jupyter or very much used to the classic interface. Another way to go about things since you asked about experiences, would be to switch to the Anaconda distribution if you aren't already. By default it will supply opinionated choices based on what is presently stable in the ecosystem without you needing to make these choices. You'll have both JupyterLab and the classic interface (or similar) there. – Wayne Jun 06 '23 at 15:15
  • @Wayne So you are saying that JupyterLab is implementing most of nbextensions and that the new jupyter is on top of the JupyterLab framework. But this means that nobody will work on nbextensions anymore. That is my definition of being dead - like no further development, no bug fixes, etc.. Thanks for the pointer to nbclassic. I haven't tried it yet. I have JupyterLab but like the minimalistic design of Jupyter. If I want to use an IDE I use VSCode. Just to explain why I ask this question. – Kab Jun 07 '23 at 07:58
  • It's a bit off topic, but with nbopen I can double-click on ipynb files in Win10 and they open in Jupyter. I haven't figured out a way to do that with JupyterLab, yet. – Kab Jun 07 '23 at 07:59
  • Having the `.ipynb` open automatically in JupyterLab vs the older notebook is just a matter of setting up your system properly. Although you are in a way getting away from proper project management by doing that anyway. In the long run that convenience will cost you in not having portable environments and robust versioning. Best bet is to always be setting up an environment and then activating it and then opening the appropriate notebooks. I suspect the nbextensions will still be worked on as nbclassic is meant to be around awhile now. I would definitely suspect bug fixes to keep them ... – Wayne Jun 07 '23 at 14:35
  • working in nbclassic & with the new situation is the horizon for some. However, most of those were always community driven endeavors & if there is no community support for them, they'll languish. Notebook ver 7 & newer versions in that vein, along with JupyterLab are the way of the main development. Also, thankfully no one has voted to close this down; however, I'd posit this discussion is more on topic on [the Jupyter Community Discourse forum](https://discourse.jupyter.org/) than StackOverflow. You'd get more informed commenters sooner there. (Link to there from here if you do.) – Wayne Jun 07 '23 at 14:40

0 Answers0