Closed francesco-ballarin closed 2 months ago
It looks like you do not have the latest jupyter-collaboration
packages installed. You would need v3 which is in beta currently.
There is also https://github.com/datalayer/jupyter-server-nbmodel as an alternative. You do not need to install all of jupyter-collaboration
, just "jupyter-docprovider>=1.0.0b1"
and "jupyter-server-ydoc>=1.0.0b1"
.
It looks like you do not have the latest
jupyter-collaboration
packages installed. You would need v3 which is in beta currently.
Well, how silly of me ;)
However, pip install "jupyter-server-ydoc>=1.0.0b1"
results in
Installing collected packages: pycrdt, sqlite-anyio, jupyter-ydoc, pycrdt-websocket, jupyter-server-fileid, jupyter-server-ydoc
Attempting uninstall: pycrdt
Found existing installation: pycrdt 0.8.31
Uninstalling pycrdt-0.8.31:
Successfully uninstalled pycrdt-0.8.31
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
fps-yjs 0.6.0 requires pycrdt<0.9.0,>=0.8.16, but you have pycrdt 0.9.8 which is incompatible.
Successfully installed jupyter-server-fileid-0.9.2 jupyter-server-ydoc-1.0.0b2 jupyter-ydoc-2.1.1 pycrdt-0.9.8 pycrdt-websocket-0.14.1 sqlite-anyio-0.2.3
Should I be installing some other dependencies with --pre
? Adding --pre
to the installation of jupyterlab didn't help, and upgrading jupyter-ydoc
with --pre
didn't help either.
Thanks for the super fast reply!
fps-yjs
sounds like something from jupyverse
, maybe the jupyverse
version you are installing does not yet support newer pycrdt
?
Indeed, I'll have to update jupyverse to support latest pycrdt.
Indeed, I'll have to update jupyverse to support latest pycrdt.
Fixed in the most recent jupyverse release. Closing.
Description
Hi, I saw https://github.com/jupyterlab/jupyter-collaboration/pull/279, and was eager to use it on my system with motivations similar to the one in https://github.com/jupyterlab/jupyterlab/issues/2833. However, cell output does not get preserved even when following the instructions in the PR.
I show below how to replicate with a docker container, but I actually have the same issue on my host machine (without docker).
Reproduce
Open a new docker image with
docker run -it --rm -p 80:80 debian:testing
and run the followingGot to the jupyter URL, create a new notebook with content
and run the cell. Save, close and re-open the tab. The output is not updated after opening the new tab.
It may not be relevant, but I also noted that the html output is not exactly the same as in the video in https://github.com/jupyterlab/jupyter-collaboration/pull/279:
[ ]
next to the second cell while it runs, while I have[*]
as in jupyterlabExpected behavior
Something similar to the video in https://github.com/jupyterlab/jupyter-collaboration/pull/279
Context