Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

SciPy 2023 tracking issue #170

Closed
19 of 34 tasks
dcherian opened this issue May 18, 2023 · 9 comments
Closed
19 of 34 tasks

SciPy 2023 tracking issue #170

dcherian opened this issue May 18, 2023 · 9 comments

Comments

@dcherian
Copy link
Contributor

dcherian commented May 18, 2023

cc @scottyhq @lsetiawan @JessicaS11 @TomNicholas @andersy005 @negin513

Any one is welcome to contribute here, please reach out!

Logistics

As reminder, last year we created a outline for attendees (max 90) to follow. The outline links to a canonical copy of the material that gets updated every year, and serves as a syllabus for a future learner who did not necessarily attend the event.

I propose we do the same this year, and use the opportunity to develop some new canonical notebooks.

Here's our proposal outline converted to a todo list. I've added links to material that exists already. For bulllet items without a llink we need to develop new material (in most cases adapt the narrative documentation to tutorial format). Clearly we have a bunch of work to do :)

Outline

Notes

At a high level, we have to work on the indexing material, and the "duck array" material. The xarray docs on indexing are fairly comprehensive so it's mostly presenting that in a friendly manner. The duck array material is fairly minor, so that would be a decent lift.

@dcherian
Copy link
Contributor Author

dcherian commented Jun 3, 2023

Meeting Notes

@dcherian
Copy link
Contributor Author

dcherian commented Jun 20, 2023

Pasting the email about Nebari for the SciPy tutorial.

I think we should do this!


We are emailing SciPy presenters to share an offer from Quansight to provide a hosted platform (JupyterHub, via Nebari) to run your tutorial.

Your workflow, and all participants' workflow, will involve: - Registering on the platform [1] (with any email and password), - Selecting a suitable machine [2] (i.e., adequate compute and storage for your tutorial), - Cloning the tutorial (git) repository in the JupyterLab interface that opens by default, - Teach your tutorial :)

As the presenter, you will have one extra step before the tutorial: create a suitable environment for your tutorial. If you have an environment.yml or requirements.txt file, this is a quick step which will take ~5 mins at most, and the Quansight team will assist you with this.

The platform (Nebari) has Dask integrated, and the capability to share files and dashboards within the platform. You can consider using this as an alternative to Binder, which currently has reduced capacity and funding. The Quansight team will be at SciPy in-person and will set up a dedicated Slack channel for support during the tutorial sessions.

If you're interested, please fill out this Google Form: https://forms.gle/recAwR1DaSau2Gsu6 by 25th June, and the Quansight team will reach out to share more information and schedule a live demo session to onboard you to the platform. If you have any questions, please contact Dharhas Pothina ([email protected]) or Pavithra Eswaramoorthy ([email protected]).

[1] Quansight won't collect or store your email on registering, you can also register with a fake email. [2] They will offer two types of machines for everyone: Small (2 CPUs, 8 GB RAM), Medium (4 CPUs, 16 GB RAM). If you need larger instances or a GPU instance, let them know in the Google Form and they will include it.

@scottyhq
Copy link
Contributor

I think we should do this!

Sounds good to me! Small instances should be fine. @lsetiawan would you be willing to fill out the form and be the point of contact for this?

@lsetiawan
Copy link
Member

lsetiawan commented Jun 20, 2023

I think we should do this!

I think this is great that they're offering Nebari! Yes. I totally agree that we should do this 😄 I've used Nebari in the past, and I think it's cool that they're willing to host for tutorial.

@dcherian
Copy link
Contributor Author

dcherian commented Jul 6, 2023

@andersy005 @negin513 @TomNicholas can we get a quick update here please?

I think at this point, we could just plan to use the existing material as-is to reduce stress. There's definitely plenty of content!

@dcherian
Copy link
Contributor Author

dcherian commented Jul 6, 2023

@lsetiawan do you have any more info on the nebari deployment? I don't see any info on the scipy tutorial website so far

@lsetiawan
Copy link
Member

The rep from quansight will try to deploy it from the repo. I'm currently on vacation. I'll ping her again tomorrow.

Hi Don,

I hope you have a nice vacation!

The onboarding meeting diverged off-topic, so I'll share a fresh set of instructions in a day or two. I'll also test the xarray tutorial on Nebari and let you know (or maybe open issues/PRs against the repo) if we need anything from you. :)

@negin513
Copy link
Contributor

negin513 commented Jul 7, 2023

Hello @dcherian ,

I created a PR for the indexing notebooks. #192

Regarding the last task under Indexing in the list above :

I can write about broadcasting, aligning, etc in these notebooks. But was not sure if it fit nicely under indexing section. Do you envision having these under advanced indexing? Does it fit the advanced indexing or the computation section better?

Here is the notebook for reference:
https://tutorial.xarray.dev/workshops/online-tutorial-series/03_computation.html

@dcherian
Copy link
Contributor Author

dcherian commented Jul 7, 2023

I can write about broadcasting, aligning, etc in these notebooks.

Sorry for being unclear. We can skip that for now. We already have material on alignment and broadcasting here

@dcherian dcherian unpinned this issue Jul 12, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants