Extract anchors from AST

[frankharkins]

Closes #11 and fixes #12

• 549c592: A prefactor moving the AST-parsing and walking to its own module. The mdx::walk_ast function calls the links::extract_from_node for each AST node.
• 7291e7f: Add tests for new behaviour
• 38f0b72: The main change in this PR, extracting anchors from the AST. For headings, we concatenate the values of all text and inlinecode nodes. For ids, we look for the id prop.
• 38c2c67: A quick follow-up to handle inline math. Tests expectations come from the website anchors.

The rest of the commits are minor changes (renaming / formatting / etc.)

I ran this new code against Qiskit/documentation and it (correctly) detects one broken link.

[Eric-Arellano]

Excellent!

Btw, I recommed simplifying the file organization to not use a folder per file, given how simple our project is. You should be able to move, for example, src/anchors/mod.rs to be src/anchors.rs.

[Eric-Arellano]

Incomplete comment

[Eric-Arellano]

Why is this part removed?

[frankharkins]

It's moved to

mdx-link-extract/src/mdx/mod.rs

Lines 25 to 34 in 38c2c67

	/// Walk the markdown AST and call a function on each node
	pub fn walk_ast<'a>(node: &'a Node, f: &mut impl FnMut(&'a Node) -> ()) -> () {
	f(node);
	if let Some(children) = node.children() {
	for child in children {
	walk_ast(child, f);
	}
	}
	}

Before, extract_from_node was the only function that used the AST, so it walked it too. Now we walk it in walk_ast.