So, it took a few attempts & steps, but I finally succeeded in exporting some of my lecture notes to LaTeX. I took them on my Surface device, in OneNote, as that is the easiest & quickest way to "prototype" the notes. Given my experience and all shortcut possibilities, I am faster typing the equations than actually writing them out by hand. In order to provide my notes to other people for them to edit them, I can use the provided export options in the OneNote Windows app: docx, pdf or onnenote again. The first two do not provide a sensible export, since neither allow the editing of equations as it should be possible.

Now, I wanted to export them to LaTeX, as laTeX does provide awesome equation & math capabilities. Unfortunately, this is not an option available by default. And even though the OneNote format is open, I did not find a way to decompress the actual OneNote file to access the xml.

For another project, I already worked with the OneNote API, so I tried to export it with that one. No luck either, as the equations are not in there. Finally, one way I found to programatically access the OneNote equations is to use OneNote online: There, you can find the equations as MathML in the HTML source. To extract the code from the OneNote webpage, I wrote a small script around puppeteer: check it out!

This HTML code could then be converted to LaTeX using pandoc, which actually does support MathML. Problem is, the exported code looks horrible. So, what we have to do is clean it up: both the HTML first and the LaTeX afterwards.

In the HTML, you will want to do the following RegEx replacements:

<script[\s\n]*type="math\/mml"[\s\n]*(id="[a-zA-Z0-9\-]*")>(([\s\S](?!script))*)<\/script> => $2
<script[\s\n]*type="math\/mml"[\s\n]*>(([\s\S](?!script))*)<\/script> => $1
<span[\s\n]*class="EquationPlaceholderText"[\s\n]*style="display: none;">\[Equation\][\s\n]*</span>[\n\s]*<span[\n\s]*class="RawEquation"[\n\s]*data-length="[0-9]*"[\n\s]*style="display: none;">[^<]*</span>

Then, you may run the conversion.

The code to convert all html files in a directory (the export dir) could look like this:

#!/bin/bash

for f in *.html 
do
  pandoc -f html -t latex -s --mathjax -o "out/$f.tex" "$f"
done

Finally, the following RegExs are a few deletions & replacements (notice the =>) I found I needed to further improve the output:

\\textenglish\[variant=[a-z]*\]\{\{\}\}\{~\}
\\textenglish\[variant=[a-z]*\]\{\}\{~\}
\\textenglish\[variant=[a-z]*\]\{\{([\S ]*[\n]*[\S ]*)\}\}\{~\} => \n$1
\\textenglish\[variant=[a-z]*\]\{\{([\S ]*[\n]*[\S ]*)\}\} => \n$1
\\textenglish\[variant=[a-z]\]\n\n
\{\}\{\{\{\[\}Equation\{\]\}\}\{[\S\\ ]*\}\{\{\} \}\}
\{\{\{\[\}Equation\{\]\}\}\{[\S\\ ]*\}\{\{\} \}\}
\{\{\{\[\}Equation\{\]\}\}\{[\S\\ ]*\n[\S\\ ]*\}\{\{\} \}\}
\{\}\{\{\{\[\}Equation\{\]\}\}\{[\S\\ ]*\}\{\{\}[\s\n]*\}\}
\}\}\{\{\{\[\}Equation\{\]\}\}\{[\S ]*\}\{\{\}
\{\}\{\{\{\[\}Equation\{\]\}\}\{[\S ]*\n*[\S ]\}\{\{\}[ \n]*\}\}
^[\s]*\{[\s\{\}~]+$
\{\}\{~\}
^[\s]*\{\}[\s]*$
^[\{\}\s]*$
^\n\n*$ => \n
\\protect\\hypertarget\{MathJax-Element-[0-9]*-Frame\}\{\}\{\n\\protect\\hypertarget\{[\S]*\n\}
  \\protect\\hypertarget\{MathJax-Element-[0-9]*-Frame\}\{\}\{\n  \\protect\\hypertarget\{[\S ]*\n  \}