ismism/src/doc/mdh.ts

type Ln = {
	l: string[],
	n: number,
}

export type Block = {
	tag: "h1" | "h2" | "h3" | "h4" | "h5" | "h6",
	htm: string,
} | {
	tag: "hr",
} | {
	tag: "blockquote",
	htm: string,
} | {
	tag: "ol" | "ul",
	li: string[],
} | {
	tag: "table",
	th: string[],
	td: string[][],
} | {
	tag: "p" | "pre" | "section",
	htm: string,
}

function pln(
	md: string
): Ln {
	return { l: md.split("\n"), n: 0 }
}

function pl(
	ln: Ln
): string {
	return ln.l[ln.n]
}

function pn(
	ln: Ln,
	n: "check" | "back" | "line" | "block" = "line",
): boolean {
	if (n == "back" && ln.n > 0) --ln.n
	else if (n == "line") ++ln.n
	else if (n == "block") while (ln.n < ln.l.length && ln.l[ln.n].trim() == "") ++ln.n
	return ln.n < ln.l.length
}

function pb(
	ln: Ln
): Block[] {
	if (!pn(ln, "check")) return []
	const lb: Block[] = []
	while (pn(ln, "block")) {
		let b: null | Block = null
		for (const p of [ph, phr, pbq, pli, pt, pp])
			if (b = p(ln)) { lb.push(b); break }
	}
	return lb
}

function pbi(
	ln: Ln
): string {
	const l = []
	do {
		if (pl(ln) == "") {
			if (l.length > 0) l.push("")
			continue
		}
		const m = pl(ln).match(/^( {4}|\t)(.*)/)
		if (m) l.push(m[2])
		else break
	} while (pn(ln))
	return l.length > 0 ? htm(pb({ l, n: 0 })) : ""
}

export function pmd(
	md: string
): Block[] {
	return pb(pln(md))
}

export function htm(
	lb: Block[]
): string {
	return lb.map(b => {
		switch (b.tag) {
			case "h1": case "h2": case "h3": case "h4": case "h5": case "h6":
			case "blockquote": case "p": case "pre": case "section":
				return `<${b.tag}>${b.htm}</${b.tag}>`
			case "hr": return "<hr>"
			case "ol": case "ul":
				return `<${b.tag}>${b.li.map(li => `<li>${li}</li>`).join("\n")}</${b.tag}>`
			case "table": {
				const tr = (t: "th" | "td", c: string[]) =>
					`<tr>${c.map(c => `<${t}>${c}</${t}>`).join("")}</tr>`
				const [th, td] = [
					(h: string[]) => tr("th", h),
					(d: string[]) => tr("td", d),
				]
				return `<${b.tag}>${[th(b.th), ...b.td.map(td)].join("\n")}</${b.tag}>`
			}
		}
	}).join("\n")
}

export function mdh(
	md: string
): string {
	return htm(pmd(md))
}

function ph(
	ln: Ln
): null | Block & { tag: "h1" | "h2" | "h3" | "h4" | "h5" | "h6" } {
	const m = pl(ln).match(/(^#{1,6}) (.*$)/) //@ts-ignore
	if (m) { pn(ln); return { tag: `h${m[1].length}`, htm: phtm(m[2]) } }
	return null
}

function phr(
	ln: Ln
): null | Block & { tag: "hr" } {
	if (["***", "---", "___"].some(h => pl(ln).startsWith(h))) { pn(ln); return { tag: "hr" } }
	return null
}

function pbq(
	ln: Ln
): null | Block & { tag: "blockquote" } {
	const l = []
	do {
		const m = pl(ln).match(/^> ?(.*$)/)
		if (m) l.push(pure(m[1]))
		else break
	} while (pn(ln))
	return l.length > 0 ? { tag: "blockquote", htm: htm(pb({ l, n: 0 })) } : null
}

function pli(
	ln: Ln
): null | Block & { tag: "ol" | "ul" } {
	let tag: null | "ol" | "ul" = null
	const li: string[] = []
	do {
		const m = pl(ln).match(/^([*+-]|^\d+\.) (.*)/)
		if (m == null && li.length > 0) {
			const bi = pbi(ln)
			if (bi == "") break
			li[li.length - 1] += bi
			pn(ln, "back"); continue
		}
		if (m == null) break
		const t = m[1].endsWith(".") ? "ol" : "ul"
		if (tag == null) tag = t
		else if (tag != t) break
		li.push(phtm(m[2]))
	} while (pn(ln))
	return tag ? { tag, li } : null
}

function pt(
	ln: Ln
): null | Block & { tag: "table" } {
	const [th, td]: [string[], string[][]] = [[], []]
	let tag: null | "table" = null
	do {
		if (tag == null) {
			if (pl(ln) == "```csv") { tag = "table"; continue }
			else return null
		} else if (pl(ln) == "```") { pn(ln); break }
		if (th.length == 0) th.push(...pl(ln).split(","))
		else td.push(pl(ln).split(","))
	} while (pn(ln))
	return { tag, th, td }
}

function pp(
	ln: Ln
): Block & { tag: "p" | "pre" } {
	const l = []
	let tag: "p" | "pre" = "p"
	do {
		if (tag == "p") {
			if (pl(ln) == "") break
			if (pl(ln).startsWith("```")) { tag = "pre"; continue }
		} else if (pl(ln) == "```") { pn(ln); break }
		l.push(tag == "p" ? phtm(pl(ln)) : pl(ln))
	} while (pn(ln))
	return { tag, htm: l.join("\n") }
}

function pure(
	t: string
): string {
	return t.replaceAll("<", "﹤")
		.replaceAll(">", "﹥")
		.replaceAll("'", "\"")
}

function phtm(
	t: string
): string {
	return pure(t)
		.replace(/\*{3,}([^\*]+)\*{3,}/g, "<b><em>$1</em></b>")
		.replace(/\*{2}([^\*]+)\*{2}/g, "<b>$1</b>")
		.replace(/\*+([^\*]+)\*+/g, "<i>$1</i>")
		.replace(/`+([^`]+)`+/g, "<code>$1</code>")
		.replace(/={2,}([^=]+)={2,}/g, "<mark>$1</mark>")
		.replace(/!\[(.*?)\]\((.*?)\)/g, "<img alt='$1' src='$2'/>")
		.replace(/\[(.*?)\]\((.*?)\)/g, "<a href='$2'>$1</a>")
		.replace(/ {2,}$/g, "<br>")
}