I recently upgraded Anaconda to include JupyterLab and Notebook 7 and have been frustrated by a number of little annoyances. Code seems to run fine. But:
However, if I run the version “Lite” on the Jupyter website, I don’t see the problem.
Also, I discovered that if I use the old version of the notebook, by running “jupyter nbclassic” (after installing it with conda), the numbering is fine.
In both Lab and Notebook versions, the help menu no longer includes links to the various Python and related docs. This is also fixed by using nbclassic.
The behavior I see happens on both Windows 11 and Rocky Linux 9. I did a fresh install with Anaconda followed by “conda update --all”. I see the same behavior on both desktop and laptop versions.
For now, I will stick with nbclassic. I prefer the way it looks anyway.
This may be the same issue as reported here?
I see you are using Anaconda, too and so maybe it is the same.
The OP said reinstalling the Anaconda Distribution entirely helped.
‘fixed’ is a bit of misnomer here. A better way to fix it would be to point it out in the Issues so that the developers can address it in the latest Jupyter tech. It looks like from here at least for Jupyter Notebook 7+ that is in the works, or there is place for you as an interested party to make a plug for the Help menu getting some attention, too.
Thanks for your quick response! I realize that I am abusing “fixed” in this context; what I really mean is that it’s “fixed for what I want.”
What I am reporting It is not quite the same behavior as reported earlier, since the number skipping happens without any Markdown cells in the notebook, but it is similar. I did not try to reinstall Anaconda, but I see the same behavior after fresh installs on three different systems.
Thanks for the link to the GitHub Issues collection. I will sniff around to see if my cell problem has been reported.
It seems that you may have some kind of extension installed since your screenshot includes and additional magic sparkles icon in the cell toolbar. Could it be that
this extension is executing extra code on the kernel which gets counted as cell executions?
Just from experience users often attribute bugs in extensions to the application, but it is the extensions which are the issue not the application.
Thank you for pointing this out. The “magic sparkles” is the icon for something called “Anaconda Assistant” which is installed by default from the Anaconda distribution. It appears to be some sort of AI coding help, which requires an account sign-in. I have not done that, and I am not interested in it.
But it appears you are right! I read the JupyterLab Reference about extensions, and then went about disabling some to see if one was causing the effect. And, indeed, the culprit was Anaconda Assistant.
From the command line, I executed
jupyter labextension disable @anaconda/assistant
after seeing which extensions are active from
jupyter labextension list
and the cell-numbering problem went away. Replacing disable with enable made it come back.
@RRosio@ericsnekbytes this is a second post this week complaining about Notebook/Lab being broken with the extension shipped by default by Anaconda being seemingly at fault. Can you pass on the word to maintainers? I believe this could be solved by passing the silent: True option to the execute_request as documented in Messaging in Jupyter — jupyter_client 8.6.2 documentation
Thank you for the ping on this issue @krassowski, I’m sorry I didn’t see this earlier! There has been a fix for this which is available in the most recent version of the extension, 4.0.21!
Hi,
I am actually having the similar issue as dbpengra has.
As he/she guided above, I could disable the labextension @anaconda/assistant and cell-numbering problem went away.
I also tried to update the jupyter labextension @anaconda/assistant by typing
jupyter labextension update @anaconda/assistant
however, I could not update the version to 4.0.21, which @RRosio mentioned that the issue is gone.
is there a specific way to update the lab extension version? or should I just diable it if I would like to see the kernel number in sequence ?
