Backport PR #86 on branch main (Import shared model)

[fcollonval]

Backport PR #86

This breaks the API of 0.2.x because the cells metadataChange is no more providing two map values but an array of map changes.

This is unfortunate but definitely easier to work with for consumer that most of the time will listen to some specific keys.

[fcollonval]

@davidbrochart @hbcarlos should we switch the notebook metadata to be a YMap too?

[davidbrochart]

You mean to be able to observe it individually? Right now we have a "meta" YMap that consists of "metadata", "nbformat" and "nbformat_minor".

[davidbrochart]

Just to be sure, the main difference is that we now have a YMap for the cell metata, that was not in PR 86, right?

[fcollonval]

Thanks for the review

Just to be sure, the main difference is that we now have a YMap for the cell metata, that was not in PR 86, right?

Yes

You mean to be able to observe it individually?

Yes to align with how we handle the metadata at the notebook and the cell level.

[davidbrochart]

Yes to align with how we handle the metadata at the notebook and the cell level.

Maybe yes.

[fcollonval]

Yes to align with how we handle the metadata at the notebook and the cell level.

Maybe yes.

Let's do that in another PR.