Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 05:31
    betatim commented #1086
  • 05:31
    betatim labeled #1086
  • 05:31
    betatim labeled #1086
  • Apr 02 21:58
    sje30 opened #1086
  • Apr 02 05:40
    betatim commented #865
  • Apr 01 20:47
    betatim commented on 56a5734
  • Apr 01 18:31
    choldgraf commented #1398
  • Apr 01 18:31

    choldgraf on master

    Switch from conda to pip in RTD Merge pull request #1398 from G… (compare)

  • Apr 01 18:31
    choldgraf closed #1398
  • Apr 01 18:31
    choldgraf closed #1397
  • Apr 01 17:53
    GeorgianaElena edited #1398
  • Apr 01 17:48
    GeorgianaElena opened #1398
  • Apr 01 17:35
    agahkarakuzu commented on 56a5734
  • Apr 01 17:29
    agahkarakuzu commented on 56a5734
  • Apr 01 17:27
    agahkarakuzu commented on 56a5734
  • Apr 01 17:26
    agahkarakuzu commented on 56a5734
  • Apr 01 17:24
    agahkarakuzu commented on 56a5734
  • Apr 01 17:22
    agahkarakuzu commented on 56a5734
  • Apr 01 17:19
    agahkarakuzu commented on 56a5734
  • Apr 01 17:18
    agahkarakuzu commented on 56a5734
Wayne's Bioinformatics Code Portal
@fomightez
@bryevdv Note the case of the true should be different now according to that stackoverflow comment.
Correct link about new host is here (I hope).
Bryan Van de Ven
@bryevdv
@fomightez I have seen both, and tried both
Wayne's Bioinformatics Code Portal
@fomightez
Darn.
Bryan Van de Ven
@bryevdv
:/
Chris Holdgraf
@choldgraf
hmmm - I'm not exactly sure...I probably would check in the nbviewer repository, as I'm pretty sure it's a purely "community-run" service, there is no dedicated person or team for running it
I think that @parente was once involved in it...maybe still? maybe @minrk ?
or, you can also ask in the community forum
Bryan Van de Ven
@bryevdv
I did ping Min in the issue a few days ago. Given that there is infrastructure involved I assumed there had to be some at least quasi-official group running things, but perhaps not. I think I'll just delete all the links to nbviewer from our own docs. Thanks!
Tim Head
@betatim
@rabernat i think you need to look into the jupyter notebook server API to tell it to shutdown the kernels. I think you can also tell it to shutdown itself which will lead to the container exiting. Freeing up the resources in binder
@bryevdv I think Min is the (only) person who knows about nbviewer.org right now. It got moved to a cluster at OVH a few weeks back (now runs on the same cluster as ovh.mybinder.org). Maybe in the move something got broken?
I think Min is currently not available so responses will be slow. However this is a chance to get involved in running nbviewer :) I also have access to the cluster so i can potentially restart/prod a container if that is what it needs
Bryan Van de Ven
@bryevdv
@betatim I wish that I could but I am myself deeply underwater
Tim Head
@betatim
:)
i added some notes to the issue about what i found
seems like it needs someone with a bit of experience of how nbviewer is setup ... or spare time to get acquainted with it
let's see what happens
Bryan Van de Ven
@bryevdv
I saw, they are consistent with what I observed re: page shows "rendered a few seconds ago" hopefully it's something easily fixable by whoever can get to it, thank you for the added info
Ryan Abernathey
@rabernat

i think you need to look into the jupyter notebook server API to tell it to shutdown the kernels. I think you can also tell it to shutdown itself which will lead to the container exiting.

Thanks @betatim! I am able to shut down kernels, but afaik, shutting down all kernels does not actually terminate the notebook server itself. Is this what a "session" is (https://github.com/jupyter/jupyter/wiki/Jupyter-Notebook-Server-API#sessions-api)? If I shut down my session, will the server shut down? My goal is to get kubernetes to actually shut down the pod and free up associated resources, but without any explicit kubernetes access.

Benjamin Feagin Jr.
@bfeaginjr
The organization I work at doesn't allow us to disclose code to third parties; is there a script we can run locally, instead of via this website, to achieve the same functionality?
Sarah Gibson
@sgibson91
@bfeaginjr repo2docker is the tool that builds the images of repos and that can be installed and run locally. You could also deploy your own BinderHub and restrict access to only those within your company
https://repo2docker.readthedocs.io/en/latest/
https://binderhub.readthedocs.io/en/latest/
Benjamin Feagin Jr.
@bfeaginjr
thank you!!!
Tim Head
@betatim
@rabernat i don't know what/if there is a API command for getting the server to shutdown. On mybinder.org the config we use for the notebook server is that it shuts itself down (after a few minutes) if all its kernels have stopped running
a session doesn't sound right, but yeah, gotta check the docs
Developer
@foobar-coder
hi, I am having some troubles executing a Jupyter Notebook (NB) in binder that tries to access our FTP site. We serve some large files through FTP and have tabix indexes for them. The idea of the NB is to show users how to use tabix to perform queries against these large files. However, we are unable to, as the requests time out. That same NB works fine inside our network and in Google Colab. When we noticed that the NB was not working, we tried to investigate by creating a minimal NB that uses ftplib to simply open a connection to the FTP site and prints “Connection successful” if it works. However, we got this error: "[Errno 99] Cannot assign requested address". We can run our other NBs that access HTTP locations in binder.org without any issue. The problem only happens for NBs that try to use an FTP connection. Could somebody from the Binder team please investigate?
Wayne's Bioinformatics Code Portal
@fomightez
@foobar-coder There are no FTP connections allowed by the public Binder for reasons of avoiding the potential for using it for abuse.
See the note about it at https://github.com/binder-examples/getting-data#large-public-files .
Towards the bottom of that section.
Wayne's Bioinformatics Code Portal
@fomightez
@foobar-coder Funny enough in the discussion NCBI was mentioned as a potential issue if that was the case, see jupyterhub/mybinder.org-deploy#427 .
Developer
@foobar-coder
@fomightez Sad to hear that, but thanks for the info
Tim Head
@betatim
@foobar-coder for examples I'd recommend making a smaller version of your dataset and storing it in your repository. then illustrate how to use them in your examples. avoids our FTP limit and saves bandwidth
kaibawong
@kaibawong
hi all, not sure if you have this problem before. I did a commit and push for a previously working repo in binder and now it doesnt run anymore and gives me this error, i cant seem to understand what went wrong. I was just testing the commit and push command. Then the whole thing just fail! Can someone please help, thank you in advance!
image.png
image.png
image.png
image.png
Sarah Gibson
@sgibson91
It sounds like there are conflicts in the dependencies of the packages you are trying to install. I would try building the repository locally with repo2docker (builds much more quickly than on the website). You can output the logs and see which packages are conflicting and change the versions.
Docs for repo2docker are here: https://repo2docker.readthedocs.io
jlchulilla
@jlchulilla
hi, I would appreciate any help
I'm not a developer but an analyst
I have connected binder to a small repo of mine. A notebook uses mathplotlib binder and it produces a crash
https://mybinder.org/v2/gh/jlchulilla/COVID-19-scripts/d22f323e7796c90ae74503cfc19ac0e9921c4ae8
the error message is this: Error loading notebook
Unreadable Notebook: /home/jovyan/ECDCcovid19TOJUPYTER.ipynb NotJSONError("Notebook does not appear to be JSON: '%%matplotlib widget\nimport pandas as pd...")
Lukas
@lukasheinrich
hi
is launching repos from ZENODO DOIs broken?
i'm trying out the example shownin the mast
https://mybinder.org/v2/zenodo/10.5281/zenodo.3242074/
but can't see any image build