feature: Parse URLs out from plaintext paragraphs

[keithfancher]

Did you check the tree-sitter docs?

• I have read all the tree-sitter docs if it relates to using the parser

Is your feature request related to a problem? Please describe.

Currently, URLs are just treated as "normal" text within a paragraph/node. This means they can't be specifically queried for things like syntax highlighting, turning off spell-checking, etc. Particularly re: spell-checking, this makes for some noisy documents... e.g. note the squiggles below:

In the pre-tree-sitter days, one could simply regex match to disable spell-check as described here, for example.

Describe the solution you'd like

If URLs could be targeted with queries, it would be easy to turn off spell-check for them and also for color schemes to highlight URLs as they see fit. (Which is a common practice in various text editors, markdown viewers, etc.)

Describe alternatives you've considered

Just for the spell-check issue, my current workarounds involve one of the following:

• Making all links explicitly [markdown links](http://blah.org)
• Making all links explicitly <http://auto-links.org>
• Turning off tree-sitter for markdown 😢

I've also experimented with the additional_vim_regex_highlighting option as described in the nvim-treesitter docs, but enabling it (and explicitly setting syntax on) still doesn't seem to allow the :syn match UrlNoSpell '\w\+:\/\/[^[:space:]]\+' contains=@NoSpell call to behave as expected -- tree-sitter must be completely disabled.

Additional context

No response