Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 08:00
    adah1972 opened #4989
  • 01:22
    yousuftufail opened #4988
  • 01:01
    M-M-Rezaee commented #1334
  • 01:01
    M-M-Rezaee commented #1334
  • Oct 18 23:09
    gnestor commented #1238
  • Oct 18 20:02
    AdityaJakkam commented #4985
  • Oct 18 19:42
    AdityaJakkam commented #4985
  • Oct 18 19:28
    AdityaJakkam commented #4985
  • Oct 18 17:46
    yousuftufail commented #3631
  • Oct 18 16:02
    takluyver commented #4987
  • Oct 18 15:47
    takluyver opened #4987
  • Oct 18 13:32
    takluyver commented #2653
  • Oct 18 13:29
    takluyver closed #4956
  • Oct 18 12:59
    prof-lupin commented #2653
  • Oct 18 11:33
    swebalaji commented #4973
  • Oct 18 10:21
    datainvestor commented #2558
  • Oct 18 07:12
    eric-wieser commented #4066
  • Oct 18 05:59
    BjornFJohansson commented #3035
  • Oct 18 03:15
    reaCodes commented #4986
  • Oct 18 03:01
    kevin-bates commented #4986
Aaron Hall, MBA
@aaronchall

I'm trying to use export but I keep getting this kind of error message on the server (and when I access the api/nbconvert it gives me {}

Content security violation: {csp-report":{"document-uri":"http://localhost:8889/api/nbconvert"...

Aaron Hall, MBA
@aaronchall
It seems to me that tornado.web.authenticated is checking for current_user - how do I ensure I have that bool(current_user) -> True?
Andreas Klostermann
@akloster
I have a custom kernelspec and would like to add paths to the PYTHONPATH environment variable without overwriting the old entries... how would I do that?
Kevin Bates
@kevin-bates

Hi @akloster, this topic has popped up from time to time, and is something I'd like to have as well. So, this being the triggering inspiration, I went ahead and created a PR to allow for this, assuming jupyter/jupyter_client#472 provides the functionality you need. If this fits your needs, it would be great if you could take it for a spin.

Since this, if approved, won't be available for a bit, you could workaround this by wrapping the launch of your kernel in a script that alters the env to your liking prior to the kernel's actual launch. This is a common trick when launching kernels as spark drivers. I can point you to such examples if you like.

Andreas Klostermann
@akloster
@kevin-bates thank you for your work, I am sure this will be useful. I created a new python package, https://github.com/akloster/freecad_jupyter , to solve the particular problem I had.
@kevin-bates the package includes a script which sets up the path correctly.
Kris Wilson
@kwlzn
:wave:
I'm part of a small team at Twitter working on Jupyter Notebook-based remote development environments for our DS/ML use cases.
We'd love to make more upstream contributions to the project. I was wondering if there was a recommended approach to discussing issues and seeking our reviewers for PRs? We have one out here that solves a data privacy/security compliance issue: jupyter/notebook#4835
Andreas Klostermann
@akloster
@kevin-bates I solved my problem largely by making a custom kernel launcher
I am trying to make integration between Jupyter and FreeCAD better. I have now a custom kernel class derived of IPythonKernel, but I am having trouble customizing anything with that
particularly I would like to do things like register line magics and display hooks from the kernel (not from extensions in a Profile , because that would require more user interaction when installing)
Andreas Klostermann
@akloster
I'm using IPKernelApp.launch_instance(kernel_class=FreeCADKernel) to launch the kernel. However, I can't find any way to hook up a function to be executed once the shell and ioloop are installed. Before that I can't register the display hooks
Jason Bratt
@jasonhero
Does anyone have any examples of creating a custom websocket message from a server extension and passing it along to aJS extension?
Ben Thayer
@bthayer2365
Hey everyone, I wrote some code for a pull request to fix a problem with backspacing while using multicursor. Is anyone here familiar with CodeBullet and willing to review this pull request? jupyter/notebook#4804
Jason Grout
@jasongrout
Thank you very much! I haven't had time to come back to this PR in the notebook yet, but it is in my queue to review (and anyone else willing to review it, please don't wait for me!)
My time is mostly concentrated on JupyterLab at this point. Do you know if this is also an issue in JupyterLab? Just curious.
Ben Thayer
@bthayer2365
Not sure, I can check and fix if needed. I had some issues with my local git repo, so this PR got closed. I'll submit another PR soon
Jason Grout
@jasongrout
whew, thanks. I was afraid you closed it because you were getting frustrated. Sorry, I wish I could do more to review and merge more quickly.
Ben Thayer
@bthayer2365
No worries! I had all my code on master, so it was messing with everything from my other branches. It was a mess... I just put it on it's own branch. Here's the new PR: jupyter/notebook#4880
Jason Grout
@jasongrout
yes, definitely good practice to do a PR from a branch, not master
another reason is it is easier to collaborate with others on a PR on a branch, but often github will prevent people from pushing to your master
Jan Nylund
@jannylund
Howdy. I got a question that I asked in jupyterhub, but probably belongs here. If I use the basenotebook and include influx-python, it breaks urrlib3's certificate validation. This does not behave similar with standard python container, but I haven't been able to figure out what happens. Reproducable code here: jannylund/jupyterhub-influx-cert-issue@2e534f2
Kevin Bates
@kevin-bates
Hi Jan. I'm not a web dev but when I run the Makefile, I get the same output from each invocation sans the expected timestamp diffs and the value of some NID entry (which I guess is a cookie). Could you please amend your great repo with a README.md file that illustrates how things break with perhaps expected outputs? Some interested web dev may then be able to help easier.
Jan Nylund
@jannylund
@kevin-bates Sure, I'll do that. Right now the master is working but the feature branch throws ssl errors.
Jan Nylund
@jannylund
I'll refactor it to be one branch for simplicity.
Kevin Bates
@kevin-bates
:+1:
Jan Nylund
@jannylund
@kevin-bates sorry for slow reply (had to take wife to restaurant). I identified the issue, it's related to how urllib3 changes handling of cacerts when going from version 1.24.x to 1.25.x and then 1.25.3. Since the jupyter-base-notebook uses 1.24.2, urllib needs to be told specifically to import cacerts. (and 1.25.x as I tried to use causes other issues). So the solution is to stick with 1.24.x and https://stackoverflow.com/questions/35015762/apple-lookup-api-certificate-error-when-called-with-python-3-4-and-urllib3-reque/35023741#35023741
Now, next thing is how to get python-influx to load these. I'll dig into that now. But seems no issues exist in jupyter-notebook. Just my misunderstanding of urrlib with the change sin versions.
Jan Nylund
@jannylund
Ok, that was simple. For some reason they disable ssl verification by default when using SSL, enabling it in the client will propagate through and all things work properly. :tada:
Sylvain Corlay
@SylvainCorlay
john
@jcoady
Is there a way to determine from javascript code or python code in a notebook if the notebook is running on JupyterHub. I need to construct a websocket from javascript to the python code and in order to create a correct url for JupyterHub I need to somehow determine that the notebook is running on JupyterHub.
Min RK
@minrk
@jcoady JUPYTERHUB_API_TOKEN (and a couple other JUPYTERHUB_) environment variables will be defined on the kernel side. Maybe harder on the javascript side.
Aaron Hall, MBA
@aaronchall
Is there a good demo notebook that walks through the features of the notebook (so I don't have to reinvent the wheel)?
Aaron Hall, MBA
@aaronchall
I'm seeing some in the notebook github repo
Min RK
@minrk
ipython/ipython-in-depth is the place where such things have lived for a long time, but I fear it is extremely outdated
(as perhaps evidenced by the fact that it's still an IPython repo)