fix: full link path replacement applied section by section rather than full file

[HyperEpsilon]

This fixes an error that occurs if a wikilink appears both in the frontmatter and in the main body of a note.

Previous implementation had the regex search the body of the text, but apply the replacement to the entire note. If there existed a [[wikilink]] in the frontmatter that was formatted exactly the same as a [[wikilink]] in the main body, then the link in the frontmatter would be the one replaced when regex.replace() was called.

My solution is as follows:

1. Using regex.exec find the borders of all sections which have to be skipped (frontmatter, code fences, code blocks, excalidraw)
2. Break the text into an array of skippable sections and all other text
3. Apply the wikilink replacement individually to each section not skipped

---

This could potentially be refactored to do the conversion while splitting, but I did it this way to take advantage of asynchronous map

[HyperEpsilon]

Now that I think about it, maybe I should condense this down for readability.

The promises could be added to the array, then call await promise.all after the regex has finished splitting

[HyperEpsilon]

Okay, should be good at this point