Problem running Miller example

[rsignell-usgs]

I tried to run the Miller Composite notebook (because I thought it might be a good one for @julienchastang poster) from
https://unidata.github.io/python-gallery/examples/miller_composite.html#sphx-glr-examples-miller-composite-py

and it's dying in cell [7] here:

https://gist.github.com/anonymous/3da3480e71a62a36983b68dc7ad398d6

what up?

[dopplershift]

You'd need to install numpy 1.11. Since then, they mucked with gradient to restrict the dx and dy to scalars. See Unidata/MetPy#174 .

[dopplershift]

You must not be running the environment.yml file for this repo, since that dependency is already encoded there...

[rsignell-usgs]

@dopplershift, True.

I went to https://unidata.github.io/python-gallery/examples/miller_composite.html#sphx-glr-examples-miller-composite-py and clicked on download Jupyter notebook, and since I wasn't missing any packages in the custom environment I was using, I didn't suspect that it would be a version issue in one of the packages.

Would it be possible in sphinx gallery to add a link to the environment.yml close to where the download as Jupyter notebook link is?

[dopplershift]

No help from sphinx-gallery here, but a pull request adding install instructions and linking the environment from from the main gallery would be welcome!

[rsignell-usgs]

Not sure the best way to do this. Closing since miller runs fine with specified environment .

[julienchastang]

@rsignell-usgs Open an issue at https://github.com/Unidata/xsede-jetstream/ if you think the problem lies there. (I did wonder, after we discussed this issue that the environment may simply be out of date.)

[julienchastang]

@dopplershift Would it be alright if I submitted a PR to add nb_conda_kernels to this project's environment.yml? Seems like a relevant addition to this notebook project. Moreover, I could then reuse this environment.yml file directly for Unidata/science-gateway#264.

[dopplershift]

@julienchastang in general, that's fine. The reason it's not is because this environment isn't intended for running a server that hosts different kernels. Is that what you need it to do? You only need to install nb_conda_kernels in the server's environment--the individual environments for each of the kernels don't actually need it installed.

[rsignell-usgs]

The individual kernels need ipykernel, right?

…

-Rich

On Nov 27, 2017, at 17:15, Ryan May ***@***.***> wrote: @julienchastang in general, that's fine. The reason it's not is because this environment isn't intended for running a server that hosts different kernels. Is that what you need it to do? You only need to install nb_conda_kernels in the server's environment--the individual environments for each of the kernels don't actually need it installed. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[dopplershift]

Right--this is pulled in by jupyter listed in the environment.yml for this repository. Believe me, I've used this very environment with my own nb_conda_kernels setup. 😁

[julienchastang]

OK thanks for the clarification. No PR then.