Upgrade MathJax from v2 to v3

[eeopd]

Currently MathJax 2.7.4 is used, but the latest version is 3.1.2. In v3, a complete rewrite of the project, some nice improvements were made, including much faster rendering and some new packages. However, because MathJax 3.0 is a complete rewrite it includes quite a few incompatible changes, so this upgrade would require quite a bit of work to get working, and in the end result in some differing behaviours/broken TeX blocks, and also cause several extensions to break.

Would you be open to a patch? Or would working on integrating KaTeX seem preferable (#3094, #4066)?

[malouke2020]

Hi,
i am wondering if we can use jupyter whitout Mathjax , my Browser crash all time when Mathjax load page whit safe.js ???
Thanks

[jtpio]

FYI the next version of the Jupyter Notebook (Notebook 7) is based on JupyterLab 4 and will use MathJax 3.

For reference this was implemented in #6777 and jupyterlab/jupyterlab#13877.

You can already try the beta with: pip install --pre notebook.

[jtpio]

Closing as this has now been implemented in Notebook 7.

[jtpio]

i am wondering if we can use jupyter whitout Mathjax , my Browser crash all time when Mathjax load page whit safe.js ???

If you are using JupyterLab or Notebook 7, you could look into switching to KaTeX via an extension: https://github.com/jupyterlab/jupyter-renderers

[malouke2020]

Hi,
I use the folowing version of jupyter notebook
notebook 6.5.4
notebook_shim 0.2.3
-and i try different Browser in Incognito mode and regular mode

• i have a big data Frame that i can't split in diffrente sub part
• when i start processing after clear all notebook and restare from scratch two things can Hapen;
• 1 crash of google chrome or fireforx
• 2 i lose the connection to my jupyter server (all process are in local no ssh no proy .....)
  i suspect Mathjax and Safe.js to triger that error beceasu the majorty of time i can see the paths to mathjax or safe.js loading when crash happen.
  i can't use jupyterlab for the reason i have to change css to dispaly dataframe side by side and this it's too simple with
  CSS = """
  .output {
  flex-direction: row;
  flex-flow: row wrap;
  font-size: 10pt;
  }
  """
  HTML('<style>{}</style>'.format(CSS))
  But with jupyterlab i found that hard to control all output dispaly ...
  can i ask you advice .
  Thanks