Create transformer from Markdown doc to LSP data

[mark-wiemer]

The Markdown doc must match the schema defined in #297, see also #296 for all pre-reqs for this work.

Outputting the LSP data itself is its own body of work and will be helpful for others looking to create their own language servers based on official documentation.

[appgurueu]

I would propose an issue which would be: Create parser for Markdown schema. The parser should include validation and helpful error messages, thus including #310. I think that parser and validator should be, more or less, one piece.

The result should be some kind of AST data structure in a decent programming language (say, Python) for which we can then easily write "stringifiers" that produce typedef files, doc pages, etc., which is what this issue would then be about.

[mark-wiemer]

I like the idea, feel free to split issues how you like. I'll clarify that the AST data is stored in-memory with tools by the unified collective, so the best language for serialization would be JSON, not Python. It is data, after all.

And I'll reiterate that I'm stepping away from this project until US news stabilizes. Which might be a long, long while :/