Obsidian to Anki has one of the most horrifying regex setup I have seen.
It involves the idea of capturing a whole segment of text that will be used for flashcard, then regexing inside it for questions (first capture group) and answers (second capture group).

Examples:
^(.*[^n:]{1}):{2}([^n:]{1}.*) (RemNote single line style)

^Q: ((?:.+n)*)n*A: (.+(?:n(?:^.{1,3}$|^.{4}(?<!<!--).*))*) (Question answer style)

I want to do this for a bunch of notes that are setup with a lot of callouts, they look like this:

> [!PDF|] [[James W. Kurose, Keith W. Ross - Computer Networking.pdf#page=296&selection=32,0,33,25|James W. Kurose, Keith W. Ross - Computer Networking, p.296]]
>
> > Describe why an application developer might choose to run an application over UDP rather than TCP.
> 
> An application developer may not want its application to use TCP’s congestion control, which can throttle the application’s sending rate at times of congestion. Often, designers of IP blah blah.

I want to capture ALL lines starting with > > as question and ALL lines starting with > as answer. First and second capture groups in regex, specifically. Only it seems the presence of > in first and second capture group causes a mess because regex barely has negative matching?

I tried a lot of things including [^ >] after first capture group and (?! >) but it doesn’t seem to work.