Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • Oct 16 22:06
    sblack-usu edited #967
  • Oct 16 21:47
    sblack-usu synchronize #967
  • Oct 16 21:14
    sblack-usu synchronize #967
  • Oct 16 20:23
    sblack-usu reopened #967
  • Oct 16 20:23
    sblack-usu closed #967
  • Oct 16 18:00
    choldgraf commented #975
  • Oct 16 12:55
    Travis jupyterhub/mybinder.org-deploy (master) passed (1012)
  • Oct 16 12:40

    betatim on master

    Revert "use urljoin" This reve… remove trailing slashes in host… Merge pull request #1206 from g… (compare)

  • Oct 16 12:40
    betatim closed #1206
  • Oct 16 12:40
    betatim closed #1204
  • Oct 16 12:33
    sgibson91 edited #1154
  • Oct 16 11:23
    pdurbin commented #969
  • Oct 16 11:04
    manics commented #975
  • Oct 16 10:20
    sgibson91 edited #1203
  • Oct 16 10:12
    sgibson91 synchronize #1203
  • Oct 16 10:12

    sgibson91 on add-turing-to-federation

    Update .travis.yml Thanks @m… (compare)

  • Oct 16 09:52
    sgibson91 edited #1203
  • Oct 16 09:50
    sgibson91 commented #1203
  • Oct 16 09:48
    sgibson91 edited #1203
  • Oct 16 09:47
    sgibson91 edited #1203
Tim Head
@betatim
and they both have e7dd45.. as the last commit in the ompy sub directory
Min RK
@minrk
Tim Head
@betatim
jupyterhub/binder#173 should i close that as it seems resolved or is there still a problem?
Min RK
@minrk
I think oslocyclotronlab/ompy@fcfe3e9 fixed the build
Aha, I know what the problem is
It was the build cache. Instead of COPY ., the Dockerfile uses git clone repo_url..., which wouldn't invalidate the build cache when there are pushes to the repo
Min RK
@minrk
Tim Head
@betatim
nice find/catch!
another reason to avoid writing a Dockerfile by hand when ever possible :)
Philip Durbin
@pdurbin
Has anyone watched the video or looked at the slides from this webinar from Tuesday? It was really good. https://www.dataone.org/webinars/developing-packaging-and-sharing-reproducible-research-objects-whole-tale-approach
Emmanuelle Gouillart
@emmanuelle
Hi! Is it possible to use jupyterlab >= 1.0 in binder? I came across jupyter/repo2docker#724 which says it is possible to choose, but even with a requirements.txt with jupyterlab==1.0.0 (or >= 1.0) the container is still using 0.35 (as checked by importing jupyterlab and printing its version from the launched notebook). Please tell me if I need to provide other information.
Fred Mitchell
@fm75
I guess that you could create your own Dockerfile in the root of your repo, or in a binder/.binder folder. But that seems like an elephant gun. this seems to suggest what you tried.
Wayne's Bioinformatics Code Portal
@fomightez
Hi @emmanuelle . Definitely possible using requirements.txt or environment.yml.
Let me dig up examples that at least were working and we can check.
Or it might be better to point us at yours that isn't working? I have mostly converted all my repos and they run with 1.2 I think it is.
Example at: https://github.com/fomightez/muscle-binder . That is one of my latest updates.
Sorry it is Version 1.1.4
gnguy
@gnguy
@fomightez Thank you for the very detailed response! Using python run runs the script through but doesn't publish the web app itself. That's an interesting note about the proxying required for other apps. I'll look more into the comments and it seems like the best next step for me would be to follow up with an issue on streamlit to figure out more about JupyterHub compatibility.
Wayne's Bioinformatics Code Portal
@fomightez
That one uses environment.yml@emmanuelle .
Emmanuelle Gouillart
@emmanuelle
Thanks @fomightez I'll give it a try with environment.yml. My repo is https://github.com/plotly/customdata-app.
Fred Mitchell
@fm75
You can confirm it either by importing as you did or running jupyter lab --version in a Terminal
Wayne's Bioinformatics Code Portal
@fomightez
The one that @fm75 pointed out shows the requirements.txt approach. I used the about JupyterLab tab in the program to get 1.1.4 showing.
Oh, you need >=1.0.0 I think.
Wayne's Bioinformatics Code Portal
@fomightez
Nope. jupyterlab>=1.0 is fine. See https://github.com/fomightez/b-roll-binder .
Ohh @emmanuelle . I see the problem with your repository.
You have two sets of instructions to Binder there and it is only using the one in the Binder directory. Where you didn't specify JupyterLab>1.0.
You need to put requirements.txt in binder directory. Or simply move postBuild to main level and delete binder directory.
Wayne's Bioinformatics Code Portal
@fomightez
@gnguy That might be best. I know it took someone quite a bit of hacking to get close to restoring VPython, which uses websockets, to work again in Binder.
@pdurbin Thanks for posting that link. There is some interesting conversation at the bottom, too.
Wayne's Bioinformatics Code Portal
@fomightez
(Sorry, I missed your reply @emmanuelle .) I was really only suggesting the environment.yml for the syntax of the line. If you can, I'd suggest sticking with requirements.txt. I find they build the fastest. Plus less formating to deal with.
Unfortunately it was the example I could get that I knew was working the fastest. Sorry.
Emmanuelle Gouillart
@emmanuelle
oh ok. Yes it works, thank you so much @fomightez !
hum broken line @fm75 (or I miss the credentials).
Wayne's Bioinformatics Code Portal
@fomightez
It is in the documentation for repo2docker as well if anyone doesn't want to read the code --> https://repo2docker.readthedocs.io/en/latest/usage.html#where-to-put-configuration-files .
Emmanuelle Gouillart
@emmanuelle
(or problem on my side for the link)
great. Thanks a lot for the explanations. It's amazing to get such top-quality user-support!
Emmanuelle Gouillart
@emmanuelle
ok :-)
Philip Durbin
@pdurbin
@fomightez sure, and they'll be actively answering questions via Discus for a while so please go ahead and post questions or comments if you have any
Tim Head
@betatim
:wave: @emmanuelle
does someone know what bagit-ro is? saw it mentioned in the slides for the webinar @pdurbin linked
Wayne's Bioinformatics Code Portal
@fomightez
Philip Durbin
@pdurbin
https://github.com/whole-tale/tale_serialization_formats links to https://github.com/ResearchObject/bagit-ro so I assume that's it. A "nascent standard" according to the webinar (around the 40:00 mark). @betatim
Fabio Zeiser
@fzeiser
XX @minrk One again, thanks a lot ;P!
Marcus Christie
@machristie
I'm trying to run the basic pytudes repo with repo2docker and for some reason it appears that repo2docker is generating a 256GB container and then hanging. Anyone else run into this?
Tim Head
@betatim
@machristie which version of repo2docker are you using? we recently fixed a problem where users who have very large user IDs on their local system produce very large images
you'll have to install master for that
Philip Durbin
@pdurbin
Chris Holdgraf
@choldgraf
very cool!
Arnim Bleier
@arnim
Last workshop this year https://twitter.com/arnimb/status/1184054363456704513 but first to be cloud-only ... crossing fingers