Parsing a Markdown-like syntax into a document tree

[SwiftDevJournal]

I'm exploring using the swift-parsing framework to write a parser for a language with Markdown-like syntax. I want to take a string and return a document tree (abstract syntax tree) for the text. To keep things simple at the start, i'm focusing on headings and paragraphs. I have the following code for the document tree:

public protocol DocumentTreeNode {
    var children: [DocumentTreeNode] { get }
}

public struct HeadingNode: DocumentTreeNode {
    public let children: [DocumentTreeNode]
    public let level: Int
    
    public init(level: Int, children: [DocumentTreeNode]) {
        self.level = level
        self.children = children
    }
}

public struct ParagraphNode: DocumentTreeNode {
    public let children: [DocumentTreeNode]
    
    public init(children: [DocumentTreeNode]) {
        self.children = children
    }
}

public struct TextNode: DocumentTreeNode {
    // A text node has no children.
    public let children: [DocumentTreeNode] = []
    public let text: String
    
    public init(_ text: String) {
        self.text = text
    }
}

I know what I have to do to parse a heading.

1. Read 1-6 # characters.
2. Read and discard whitespace.
3. Read everything until the end of a paragraph.

This should return a heading node with one child: a text node with the heading title.

Parsing a paragraph entails Step 3 in the header parsing flow, returning a paragraph node with one child: a text node with the paragraph text.

To parse the text, go through each paragraph. Check if it's a heading. If so, parse the heading. If not, parse the paragraph. Accumulate the document tree nodes into an array.

I have an initial version of the code using swift-parsing.

let heading = Parse {
    Many(1...6) {
        "#"
    }
            
    Whitespace()
    PrefixUpTo("\n\n")
 }
        
let paragraph = Parse {
    PrefixUpTo("\n\n")
}
        
let block = OneOf {
    heading
    paragraph
}
        
let parser = Many {
    block
}

To move ahead on the parser, I have two questions:

1. How do I get the number of # characters that were read so I can set the heading level?
2. How do I deal with the parent-child relationship, where the text I'm reading until the end of the paragraph is a text node that's a child of the heading or paragraph?