Quarto Markdown and Python Markdown2 can treat some extensions
differently. To be able to use all Quarto Markdown features, it can be
beneficial to read the quarto generated HTML directly instead of
convertion Markdown to HTML using Python. This change allows reading
HTML and uses the HTML file as default.

This addresses #8

Co-authored-by: Tobias Koelling <tobias.koelling@mpimet.mpg.de>

This change uses markdown2 to actually parse the markdown files
in order to correctly identify headers (and e.g. not accidentally read
code comments as headers).
It also formats the generated summary a little different (subheadings
end up as list-items, not as another heading level).
This way, the generated summary page is a little less noisy.

Co-authored-by: Florian Ziemen <ziemen@dkrz.de>