In Jupyter notebook, all cell ids change on save.

[dtabell]

This problem means that git diff on a .ipynb file can report many changes even in the absence of any “real” change in the notebook. Aside from the diff noise, it means the file hash changes and git tells me I need to commit the “updated” file.
This issue, opened on 20 January in the JupyterLab repo, is (as of 15 May) now closed, which implies this problem is now fixed. Is it possible to update JupyterLab, or whatever piece lower down that is responsible for this behavior? Or was the problem not truly fixed?

[robnagler]

It looks like this fix is part of JupyterLab 4.0, which has yet to be released.

The problem is the newest ipython version added cell ids, but jupyterlab wasn't updated to support them so they get regenerated each time. JupyterLab should 4.4, not 4.5.

I don't think there's an easy fix for us.

[dtabell]

I'm confused: they’re working on 4.5, but have not released 4.0? Anyways, I came across the following possibility: jupyterlab/jupyterlab#9729 (comment) Have not yet tried it.

…

On 17 Sep 2021, at 14:09, Rob Nagler ***@***.***> wrote: It looks like this fix is part of 4.0, which has yet to be released <https://github.com/jupyterlab/jupyterlab/projects/12>. The problem is the newest ipython version added cell ids, but jupyterlab wasn't updated to support them so they get regenerated each time. JupyterLab should 4.4, not 4.5. I don't think there's an easy fix for us. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#236 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACXTUHEOBBC4A7HSSJOUWE3UCON7RANCNFSM5EIHCBRA>. Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[robnagler]

Notebook version 4.5. JupyterLab version 4.0.

This is part of our version hell. We will need to downgrade ipython to match this. It's not as easy as:

python -m pip install --upgrade 'nbformat<5.1.2'

[dtabell]

I guess that means for now we

sed -i -e '/"id": /d' *.ipynb

(or similar) before committing our notebooks in JupyterLab.