Open rtapella opened 11 months ago
I am able to successfully install jupyter-fs, but I cannot find a current jupyter_server_config.json and am having trouble creating a persistent one myself. The paths from jupyter --paths
that exist are in in opt/conda (which doesn't persistent on restarts) and .projects/jupyter which I thought should persist on restart, but I am unable to create a file with touch
in the dockerfile and creating a file manually in the ade in this folder is not persisting on restart. RUN jupyter server --generate-config
also isn't working. Any suggestions?
There are some bugs in previous versions of jupyter-fs (0.3.1 and lower), and I was only able to reliably get jupyter-fs v0.4.1 to work locally with jupyterlab4, so I recommend we hold off on this until we upgrade to jupyterlab4 This does seem like a handy package though
Okay thanks for trying it out. I was not able to successfully install it. I wanted to just try it out to see if it's worth adopting
@wildintellect
I have been unable to get jupyter-fs
successfully working in the ADE. I have tried the following:
jupyter_server_config.py
with jupyter server --generate-config
in the jupyter --config-dir
folder. Adding
c.ServerApp.contents_manager_class = "jupyterfs.metamanager.MetaManager"
c.ServerApp.jpserver_extensions = {"jupyterfs.extension": True}
worked or also
c.NotebookApp.contents_manager_class = "jupyterfs.metamanager.MetaManager"
c.NotebookApp.nbserver_extensions = {"jupyterfs.extension": True}
Or, adding a jupyter_server_config.json
in the jupyter --config-dir
folder with
{
"ServerApp": {
"contents_manager_class": "jupyterfs.metamanager.MetaManager",
"jpserver_extensions": {
"jupyterfs.extension": true
}
}
}
worked too
All these different variations of jupyter_server_config.json
and jupyter_server_config.py
do not work in the ADE. I was using the url osfs:///projects
. We don't have s3 access without credentials so can't be doing s3:// links, need to be doing osfs://
jupyter-fs
with conda, pip, and npm and none workOther notes:
jupyter_server_config
file jupyter labextension list
has a different version of jupyter-fs
(0.4.1) than conda list
(1.0.0) and what I installed in the environment.yml (1.0.0). But also, jupyter-fs
v0.4.1 should still work and even when I directly install 0.4.1 and the versions for conda list
and jupyter labextension list
match at 0.4.1, the extension still doesn't work. So I am not sure if this is really an issue since 0.4.1 works locally too
Does anyone have any other recommendations of things to try? cc @marjo-luc @batpad @yuvipanda - what library did we decide to try on VEDA for S3 file browsing?
Would like to install this in my jupyter workspace: https://github.com/jpmorganchase/jupyter-fs
Need to pip install && add to jupyter_server_config.json