Situation of jupyter accessibility, where to start?

Hi all,
I have just discovered thisForum but it seems that discussions are outdated; is it abandonned or has it migrated elsewhere ?
I am working in an university project on inclusion of students. I am trying to find software whose interface can provide solutions for students with special needs. At the moment I am focused on dyslexia but anything can enter the scope. I would be gratefull if anybody can brief me on the ongoing projects and the places where I can find help and if possible participate to the efforts.

1 Like

There are weekly meetings, with the next meeting (usually on Wednesdays) appearing on Jupyter Community Calendar

It’s a slow boat to turn, but there is work ongoing on here:

The goal is usually to decide what ends up as the next PR to the relevant repos (e.g. jupyterlab/jupyterlab, jupyterlab/lumino), as well as community events that lead to additional work around non-purely-code topics, such as documentation and outreach.

Specifically for dyslexia, we’ve done some looking at some of the prospects of integrating customized fonts, such as dyslexie and atkinson hyperlegible on jupyterlab-fonts, but there hasn’t been a lot of motion on that recently.


Many thanks @bollwyvl for all these informations.
I will try to read as many pages as I can in the time dedicated to accessibility on my timetable, I’ll do my best to join your next meeting with the best knowledge of your team’s work as I can. It will be evening time for me (7pm) but I can manage, I have worked in standardization and I am used to around the world meeting schedules.
Specifically for dyslexia I have found that jupyter-themes might be a track to explore another being to upgrade jupyterlab-fonts which brings a lot of errors today. I’ll try to further investigate .
Thanks again

jupyter-themes might be a track to explore

From a broader accesibility perspective: notebook 7 will be replacing jquery/codemirror5 with the jupyterlab 4 stack and codemirror6, which has dramatically improved accessibility for screen reader users… this has been one of those “big rocks” that it’s tough to be serious about.

upgrade jupyterlab-fonts which brings a lot of errors today

Please feel free to leave an issue there with the issues you are seeing with that extension, and I’ll see what i can do about it, though I have been using it with recent jupyterlab 3.x releases without much issue.

1 Like

I have written an issue with a link to a possible other issue. I am wondering if there is a dependency problem on version of nbconvert While exploring I have seen somewhere that nbconvert has version <7 but 7.0.0 is intsalled.

I am sorry I do not understand, my jupyter’s knowledge of the dev side is not sufficient, I’ll need some more reading and probably explanations.

Sorry for the delay it took me some time to find out the cause of the issue (see below).

Things are OK now with jupyterlab-fonts module I can see the default 3 fonts with their various faces that are available.
As the jupyter configuration informations are spread in various places (see below) I have not been able to locate where and how to add fonts. I have opendyslexic fonts in otf format, I suspected that I have to add the font in .local/share/virtualenvs/FlaskVenv-XZ2RGVQf/lib/python3.10/site-packages/jupyterlab_fonts/labextensions/@deathbeds/ . Is that a correct start ? It seems not to be sufficient … If I succeed would you like me to make a PR ?

Origin of the previous installation problems :

I am migrating to the use of virtual environments built with pipenv, I am unfortunately in a weird transitional situation where I have got interferences between the modules installed by the system, the modules installed by the user pip install command and (shame on me) modules installed with sudo pip install. On top of that jupyter seems to look for its configuration from the ~/.jupyter folder which yields peculiar effects when you have modules in your virtual environment that require versions not compatible with modules in the user environment ! On top of that if you add that jupyter adds some more information in ~/.local folders (etc, share, lib) you are done. It might be quite usefull and robust that jupyter writes all the configuration informations it needs inside ~/.local/share/virtualenvs/venv-dirname-XXXXX