Jupyter on Raspberry Pi, Android? miniforge arm64 packages

[Wes Turner]

Does anyone have experience with Jupyter on a Raspberry Pi and/or Android devices?

While it's already possible to pip install JupyterLab on a Pi or Android device (and wait for a number of builds to compile because there are very few ARMv7 or ARMv8/ARM64 already-built wheels on PyPI for anything), it's now possible to install conda-forge packages on a Pi with miniforge (because conda-forge builds for ARM64, too, now)!

https://github.com/conda-forge/miniforge/

https://github.com/conda-forge/miniforge/releases

wget https://github.com/conda-forge/miniforge/releases/download/4.8.3-4/Miniforge3-4.8.3-4-Linux-aarch64.sh

wget https://github.com/conda-forge/miniforge/releases/download/4.8.3-4/Miniforge-pypy3-4.8.3-4-Linux-aarch64.sh.sha256

./.sh --help

IDK if there are any plans to release a Docker containers with miniforge installed; just like the miniconda docker images.

The newest Raspberry Pi 4 has 8Gb of RAM (for $75 for just the board)

https://conda-forge.org/ is sponsored by NumFocus. Conda-forge feedstocks are built by conda-smithy on a number of free CI services: "travis-ci.com, appveyor.com, circleci.com, dev.azure.com"

[Wes Turner]

Use cases:

- Jupyter Notebook/Lab on an ARM64 Android drvice

- Jupyter Notebook/Lab on a single-board ARM64 device like a Raspberry Pi

- JupyterHub on a single-board ARM64 device like a Raspberry Pi

- JupyterHub on an ARM64 cluster (k3s)

- Jupyter Notebook/Lab + Conda-forge on a $100 [$200] Pinebook [Pro] ARM64 notebook

- Jupyter Notebook/Lab + Conda-forge on a MacOS ARM64 notebook (eventually?)

How many 8Gb RPi4's would it take to serve a class of 20-30?

(If Jyve + Brython doesn't solve the [offline] deployment takes an hour for everyone problem for intro lessons that don't require any packages)

There are a number of benefits to ARM64.

- Cost in an educational context is one constraint.

- Energy efficiency is another.

There are also ARM HPC clusters these days; due to the memory bandwidth.

[Jonathan Gutow]

I have been developing direct data acquisition from sensors inside Jupyter on Raspberry Pis. It is a little slow on a 3B+, but not bad. Based on my experience a 2GB  Pi 4B would probably be fine for student use. That said, more RAM is better. I use a 4GB 4B running Xubuntu driving an artists drawing display as an electronic white board in class and for screen casting. I often run a few web browser tabs along with the white board (Xjournal ++) and screen capture (vokoscreen) software. I have run sympy and jupyter notebooks on this set up for demonstrations. Some details on my setup can be found at: https://gutow.github.io/Pi-FrakenTablet/Pi_based_artist_drawing_slate.html.

I am just in the process of releasing the first beta of the data acquisition inside Jupyter (see https://github.com/JupyterPhysSciLab/JupyterPiDAQ).

Hope this info is useful to you.

Jonathan

[Wes Turner]

Thanks. This sounds like a good solution for "Options for giving math talks and lectures online" https://terrytao.wordpress.com/2020/03/10/options-for-giving-math-talks-and-lectures-online/ ... 

https://news.ycombinator.com/item?id=22539537

> [...]

> Jupyter notebooks can be saved to HTML slides with reveal.js, but if you want to execute code cells within a slide, you'll need to install RISE: https://rise.readthedocs.io/en/stable/

>

> Here are the docs for CoCalc Course Management; Handouts, Assignments, nbgrader: https://doc.cocalc.com/teaching-course-management.html

>

> Here are the docs for nbgrader: https://nbgrader.readthedocs.io/en/stable/

https://github.com/quobit/awesome-python-in-education/blob/master/README.md#jupyter

If Xjournal + vokoscreen don't overload a Pi 4, that sounds great. Haven't heard of vokoscreen, but have worked with simplescreenrecorder and OBS (on x64 workstations, at least). 'Screenkey' is useful for showing which keys are pressed as captions in the screencast. https://www.reddit.com/r/Fedora/comments/h92iik/any_screen_recorder_you_can_recommend/?utm_medium=android_app&utm_source=share

--
You received this message because you are subscribed to the Google Groups "Teaching with Jupyter Notebooks" group.
To unsubscribe from this group and stop receiving emails from it, send an email to jupyter-educat...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/jupyter-education/8e4c23fe-2073-4bea-949a-c667bf1ea233n%40googlegroups.com.

[Wes Turner]

(At $120 with a detachable backlit keyboard, the PineTab would be significantly less costly than $400 if it could handle the workload)

[Jonathan Gutow]

The PineTab is definitely cheaper. However you give up the much higher resolution and larger screen, pressure sensitive pen and the digitizer is usually more responsive than typical capacitive touch screens. Note, the drawing screen cannot be used with your fingers.

So it strongly depends upon your usage. I do use it for teaching, but bought it with my own money so that I could play with digital drawing and painting.

On 7/15/20 8:33 PM, Wes Turner wrote:

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you recognize the sender and know the content is safe.

-- 
Dr. Jonathan Gutow
Chemistry Department
UW Oshkosh
web: https://uwosh.edu/facstaff/gutow
e-mail: gu...@uwosh.edu
Ph: 920-424-1326