Getting custom environments to appear in Jupyter kernel list on windows

[rsignell-usgs]

@jkellndorfer , okay we spent hours last night trying to get this to work with Miniconda and failed, right?

We should have just read the IOOS environment instructions by Filipe @ocefpaf at
http://ioos.github.io/notebooks_demos/other_resources/
because we missed a step.

After installing miniconda, we should have done:

conda config --add channels conda-forge --force
conda update --all  -y
conda install  nb_conda_kernels -y
conda env create -f cond_edb.yml -q --force

The conda update --all step is important because it updates all the packages to conda-forge.

The above works. Proof:

[jkellndorfer]

Great. I’ll change the installation instructions .Thanks!

…

On Jan 19, 2018, at 10:21 AM, Rich Signell ***@***.***> wrote: @jkellndorfer <https://github.com/jkellndorfer> , okay we spent hours last night trying to get this to work with Miniconda and failed, right? We should have just read the IOOS environment instructions by Filipe @ocefpaf <https://github.com/ocefpaf> at http://ioos.github.io/notebooks_demos/other_resources/ <http://ioos.github.io/notebooks_demos/other_resources/> because we missed a step. After installing miniconda, we should have done: conda config --add channels conda-forge --force conda update --yes --all conda install nb_conda_kernels conda env create --quiet --file cond_edb.yml The update step is important because it updates all the packages to conda-forge. The above works. Proof: <https://user-images.githubusercontent.com/1872600/35157575-64ac1e7e-fd02-11e7-9b00-92447c7ba4ee.png> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AJ95EZ4IlQZtVJ_jAGiQp8G8h8PK-PKcks5tMLLzgaJpZM4Rkn9y>.

[jkellndorfer]

Did Filipe have instructions on how to bundle and install from a memory stick?

…

On Jan 19, 2018, at 10:21 AM, Rich Signell ***@***.***> wrote: @jkellndorfer <https://github.com/jkellndorfer> , okay we spent hours last night trying to get this to work with Miniconda and failed, right? We should have just read the IOOS environment instructions by Filipe @ocefpaf <https://github.com/ocefpaf> at http://ioos.github.io/notebooks_demos/other_resources/ <http://ioos.github.io/notebooks_demos/other_resources/> because we missed a step. After installing miniconda, we should have done: conda config --add channels conda-forge --force conda update --yes --all conda install nb_conda_kernels conda env create --quiet --file cond_edb.yml The update step is important because it updates all the packages to conda-forge. The above works. Proof: <https://user-images.githubusercontent.com/1872600/35157575-64ac1e7e-fd02-11e7-9b00-92447c7ba4ee.png> — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AJ95EZ4IlQZtVJ_jAGiQp8G8h8PK-PKcks5tMLLzgaJpZM4Rkn9y>.

[ocefpaf]

I see that @rsignell-usgs already answered that one via e-mail but here it is for completeness in case someone looks for it in this issue:

https://conda.io/docs/user-guide/tasks/create-custom-channels.html

One alternative is to create your own installable "miniconda-like" distribution with conda-constructor:

https://github.com/conda/constructor

My experiences with constructor in the past where not good but that was one year ago and maybe the project evolved.

[josef-ebd]

Hi Filipe, I tried condo-constructor, but didn’t succeed either. So, now for the first method: I have all the packages (..bz2 files) from the install and creation of the environment to build a channel with, but the pkgs directory also contains still the non-conda-forge bz2 files that I assume came with Miniconda. I assume I would have to delete all those that didn’t come from conda-forge before building the channel. Is that’s how you do it? Thanks, Josef

…

On Jan 19, 2018, at 2:38 PM, Filipe ***@***.***> wrote: I see that @rsignell-usgs <https://github.com/rsignell-usgs> already answered that one via e-mail but here it is for completeness in case someone looks for it in this issue: https://conda.io/docs/user-guide/tasks/create-custom-channels.html <https://conda.io/docs/user-guide/tasks/create-custom-channels.html> One alternative is to create your own installable "miniconda-like" distribution with conda-constructor: https://github.com/conda/constructor <https://github.com/conda/constructor> My experiences with constructor in the past where not good but that was one year ago and maybe the project evolved. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#2 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/APdHga0YOgBQxi_wC0S8RILafdc_JKrZks5tMO8wgaJpZM4Rkn9y>.

[ocefpaf]

I tried condo-constructor, but didn’t succeed either.

I guess that constructor is not mature enough yet.

I assume I would have to delete all those that didn’t come from conda-forge before building the channel. Is that’s how you do it?

No, I leave them all there. Let me know if that approach does not work for you.

[jkellndorfer]

Ok. That will be the approach once it actually runs. Still no luck in seeing any kernels. I installed the extensions, but that did not make a difference.

…

On Jan 24, 2018, at 12:50 PM, Filipe ***@***.***> wrote: I tried condo-constructor, but didn’t succeed either. I guess that constructor is not mature enough yet. I assume I would have to delete all those that didn’t come from conda-forge before building the channel. Is that’s how you do it? No, I leave them all there. Let me know if that approach does not work for you. — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#2 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AJ95EcUcXrO_efXsMD2OSgEy7T8LRFSTks5tN22DgaJpZM4Rkn9y>.

[jkellndorfer]

Ok, I was able to make it work in Windows with:

activate ebd
conda install nb_conda_kernels. # if it was not installed yet with the updated conda_ebd.yml
jupyter notebook

[rsignell-usgs]

Well, at least something works.

But I still want my kernels available for selection in the notebook, they way it worked just 5 days ago:

[josef-ebd]

I agree. It needs to work again …

…

____________________ On Jan 24, 2018, at 2:38 PM, Rich Signell <notifications@github.com<mailto:notifications@github.com>> wrote: Well, at least something works. But I still want my kernels available for selection in the notebook, they way it worked just 5 days ago: [2018-01-19_15-32-35]<https://user-images.githubusercontent.com/1872600/35353106-25b88fe8-0114-11e8-89d5-43f8432095b5.png> — You are receiving this because you commented. Reply to this email directly, view it on GitHub<#2 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/APdHgUzlxIZdlAGUadSs9W8bIDaIBNdtks5tN4bCgaJpZM4Rkn9y>.

[ocefpaf]

jupyter_client 5.2.2 for Windows just got uploaded to the channel. Can you try it again and let me know? I suspect that it should work without the nb_conda_kernels patch, if not I will work on it tomorrow.

[jkellndorfer]

great. Is this a Windows only issue or is Linux affected as well? I have not reinstalled in Linux.

…

On Jan 24, 2018, at 2:47 PM, Filipe ***@***.***> wrote: jupyter_client 5.2.2 for Windows just got uploaded to the channel. Can you try it again and let me know? I suspect that it should work without the nb_conda_kernels patch, if not I will work on it tomorrow. — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#2 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AJ95EdS06lq3V7POxClDbB60KoV9b3b7ks5tN4jEgaJpZM4Rkn9y>.

[rsignell-usgs]

On Windows, just did:

activate root
conda install jupyter_client
jupyter notebook

and..... 🍾 🍾 🍾

[ocefpaf]

Thanks to @minrk for fixing jupyter_client in jupyter/jupyter_client#339!

[rsignell-usgs]

I just did

activate root
conda install terminado
jupyter notebook

to get the terminal support, just to complete the Jupyter experience. No more "Terminals Unavailable":