refactor prosemirror conversion for understandability

[vezwork]

Changes

• Apply formatter to changed files (mostly normalize whitespace, semicolons)
• Big refactor of mapTokens (renamed to mapTokensRecursive)
  • Move nested functions out
  • Clarify isToken logic by removing redundant logic
  • This function has somewhat brain-bending recursive functionality (I was confused and I also saw it being misused in resolvePandocBlockCapsuleText, resulting in redundant logic). It has two interrelated recursive parts. A "recursive array data mapping" part and a "recursive function output content mapping" part. Because of the recursive nature of this function, I feel justified in going full functional programmer on its refactor: I refactored into two separate functions that return functions: mapRecursiveArray factors out the "recursive array data mapping" part, and mapTokenAndContentTokensRecursively factors out the "recursive function output content mapping" part. It is still a difficult function to understand, but I've removed some redundancy and made how its parts depend on eachother more direct and explicit.
  • Add docs
• After refactoring mapTokensRecursive I understood it better and was able to understand and significantly simplify resolvePandocBlockCapsuleText
• Change writeToken to prefer Array.find over iterating through the array until the element is found.
• Refactor createHandlers
  • Pulls out the non-looping part into createHandler which is nice because it can now early return instead of modifying a variable.

TODO

• add tests to increase confidence that this refactor does not change roundtripping behaviour. DONE in Snapshot test function improvements and more snapshot tests! #790

Could we add tests to see if the prosemirror doc was instantiated properly as well?

[vezwork]

The new tests in CI from #790 caught an error caused by this refactor! The visual editor is unable to open valid-nesting.qmd in this PR. It's exciting that the tests helped me!

On the other hand it's disappointing that my refactor is not flawless and perfect.

I know one thing that could have different behaviour than the original code (The fact that mapTokenAndContentTokensRecursively doesn't check if a token has children like the code previously did). I could take a look at that... Edit:I tried changing what I thought could be causing the error and it didn't fix it. I'd have to incrementally try the refactor's changes to see what's going wrong...

But I think I'm going to set aside this refactor for now. Writing this PR helped me understand the code, even if my refactor is not perfect. I will leave it up as a draft so I can reference it when I come back to this area.