Markdown parsing (link and parenthesis)

main.rs

// see fediverse thread: https://astolfo.social/notes/9kgqfhd4f9gqlh39
//
// Written by @benaryorg with the idea of enforcing balanced parens in links, following guideline:
//
// I think the easiest, best, and most incorrect solution would be to enforce balanced parens in the links.

use ::
{
	nom::
	{
		IResult,
		Finish,
		bytes::complete::
		{
			tag,
		},
		character::complete::
		{
			anychar,
			char,
			one_of,
			none_of,
			alphanumeric1,
		},
		combinator::
		{
			complete,
			map,
			recognize,
			opt,
		},
		multi::
		{
			many0,
			many1,
		},
		branch::
		{
			alt,
		},
		sequence::
		{
			delimited,
		},
	},
	std::
	{
		io::stdin,
	},
};

#[derive(Debug, Hash, Clone, PartialEq, Eq, PartialOrd, Ord)]
enum Markdown
{
	Char(char),
	Link
	{
		text: String,
		url: String,
	}
}

fn markdown_link_ref_balanced_parens(input: &str) -> IResult<&str, ()>
{
	// does not support:
	//   unicode
	let (input, _) = many1(alt(
		( alphanumeric1
		, recognize(one_of("-._~:/?#[]@!$&'*+,;%= "))
		, recognize(delimited(char('('), opt(markdown_link_ref_balanced_parens), char(')')))
		)))(input)?;

	Ok((input, ()))
}

fn markdown_link_ref(input: &str) -> IResult<&str, ()>
{
	// schema
	let (input, _) = alt((tag("https://"), tag("http://")))(input)?;
	// domain
	//   does not support IDN
	//   does not validate for multiple of: '@', ':', etc.
	let (input, _) = many1(alt((alphanumeric1, recognize(one_of("-.:@")))))(input)?;
	// path
	let (input, _) = opt(markdown_link_ref_balanced_parens)(input)?;

	Ok((input, ()))
}

fn markdown_link(input: &str) -> IResult<&str, Markdown>
{
	let (input, _) = char('[')(input)?;
	let (input, text) = recognize(many1(none_of("\\]")))(input)?;
	let (input, _) = char(']')(input)?;
	let (input, _) = char('(')(input)?;
	let (input, url) = recognize(markdown_link_ref)(input)?;
	let (input, _) = char(')')(input)?;

	Ok((input, Markdown::Link
	{
		text: text.into(),
		url: url.into(),
	}))
}

fn markdown_escape(input: &str) -> IResult<&str, Markdown>
{
	let (input, _) = char('\\')(input)?;
	map(anychar, Markdown::Char)(input)
}

fn markdown_element(input: &str) -> IResult<&str, Markdown>
{
	alt(
		( markdown_escape
		, markdown_link
		// always fall backto reading raw chars, this way the parser cannot hard-fail it will just produce plaintext
		, |input| map(anychar, Markdown::Char)(input)
		)
	)(input)
}

fn markdown(input: &str) -> IResult<&str, Vec<Markdown>>
{
	many0(markdown_element)(input)
}

fn main()
{
	for line in stdin().lines().take_while(Result::is_ok).map(Result::unwrap)
	{
		match complete(markdown)(&line).finish()
		{
			Ok((_, out)) => println!("{:?}", out),
			Err(err) => eprintln!("{:?}", err),
		}
	}
}

#[cfg(test)]
mod test
{
	#[test]
	fn parse_markdown()
	{
		use crate::{markdown, Markdown};

		// no parens in link with parens around text
		assert_eq!(markdown("foo (see: [something](https://en.wikipedia.org/wiki/Text)) bar"), Ok(("",
			vec!
			// opening text, parens, and more text
			[ Markdown::Char('f'), Markdown::Char('o'), Markdown::Char('o'), Markdown::Char(' '), Markdown::Char('('), Markdown::Char('s'), Markdown::Char('e'), Markdown::Char('e'), Markdown::Char(':'), Markdown::Char(' ')
			// link (no parens)
			, Markdown::Link { text: "something".into(), url: "https://en.wikipedia.org/wiki/Text".into(), }
			// closing parens, and following text
			, Markdown::Char(')'), Markdown::Char(' '), Markdown::Char('b'), Markdown::Char('a'), Markdown::Char('r')
			])));

		// balanced parens in link with parens around text
		assert_eq!(markdown("foo (see: [something](https://en.wikipedia.org/wiki/Text_(literary_theory))) bar"), Ok(("",
			vec!
			// opening text, parens, and more text
			[ Markdown::Char('f'), Markdown::Char('o'), Markdown::Char('o'), Markdown::Char(' '), Markdown::Char('('), Markdown::Char('s'), Markdown::Char('e'), Markdown::Char('e'), Markdown::Char(':'), Markdown::Char(' ')
			// link (balanced parens)
			, Markdown::Link { text: "something".into(), url: "https://en.wikipedia.org/wiki/Text_(literary_theory)".into(), }
			// closing parens, and following text
			, Markdown::Char(')'), Markdown::Char(' '), Markdown::Char('b'), Markdown::Char('a'), Markdown::Char('r')
			])));

		// trailing parens, balanced and recognized
		assert_eq!(markdown("foo (see: [something](https://en.wikipedia.org/wiki/Text()) bar"), Ok(("",
			vec!
			// opening text, parens, and more text
			[ Markdown::Char('f'), Markdown::Char('o'), Markdown::Char('o'), Markdown::Char(' '), Markdown::Char('('), Markdown::Char('s'), Markdown::Char('e'), Markdown::Char('e'), Markdown::Char(':'), Markdown::Char(' ')
			// link (balanced parens)
			, Markdown::Link { text: "something".into(), url: "https://en.wikipedia.org/wiki/Text()".into(), }
			// closing parens, and following text
			, Markdown::Char(' '), Markdown::Char('b'), Markdown::Char('a'), Markdown::Char('r')
			])));

		// trailing parens, unbalanced and ignored
		assert_eq!(markdown("[something](https://en.wikipedia.org/wiki/Halting_problem))"), Ok(("",
			vec!
			// link (balanced parens, meaning no parens)
			[ Markdown::Link { text: "something".into(), url: "https://en.wikipedia.org/wiki/Halting_problem".into(), }
			// closing parens which doesn't match the link
			, Markdown::Char(')')
			])));

		// spaces in link
		assert_eq!(markdown("[something](https://en.wikipedia.org/wiki/Among Us)"), Ok(("",
			vec!
			// link (balanced parens, meaning no parens), does include space
			[ Markdown::Link { text: "something".into(), url: "https://en.wikipedia.org/wiki/Among Us".into(), }
			])));

		// spaces and unbalanced parens
		assert_eq!(markdown("[something](https://en.wikipedia.org/wiki/Amogus) Us)"), Ok(("",
			vec!
			// link (balanced parens, meaning no parens), does not include space
			[ Markdown::Link { text: "something".into(), url: "https://en.wikipedia.org/wiki/Amogus".into(), }
			// trailing " Us)" that doesn't belong to the link
			, Markdown::Char(' '), Markdown::Char('U'), Markdown::Char('s'), Markdown::Char(')')
			])));

		// parens and spaces
		assert_eq!(markdown("[something](https://en.wikipedia.org/wiki/Amogus (meme)) Us)"), Ok(("",
			vec!
			// link (balanced parens, meaning it contains the "(meme)"), does include space
			[ Markdown::Link { text: "something".into(), url: "https://en.wikipedia.org/wiki/Amogus (meme)".into(), }
			// trailing " Us)" that doesn't belong to the link
			, Markdown::Char(' '), Markdown::Char('U'), Markdown::Char('s'), Markdown::Char(')')
			])));

		// the failing `(` Wikipedia article
		assert_eq!(markdown("[something](https://en.wikipedia.org/wiki/()"), Ok(("",
			vec!
			// turns into plaintext, the link name first
			[ Markdown::Char('['), Markdown::Char('s'), Markdown::Char('o'), Markdown::Char('m'), Markdown::Char('e'), Markdown::Char('t'), Markdown::Char('h'), Markdown::Char('i'), Markdown::Char('n'), Markdown::Char('g'), Markdown::Char(']')
			// link ref next
			, Markdown::Char('('), Markdown::Char('h'), Markdown::Char('t'), Markdown::Char('t'), Markdown::Char('p'), Markdown::Char('s'), Markdown::Char(':'), Markdown::Char('/'), Markdown::Char('/'), Markdown::Char('e'), Markdown::Char('n'), Markdown::Char('.'), Markdown::Char('w'), Markdown::Char('i'), Markdown::Char('k'), Markdown::Char('i'), Markdown::Char('p'), Markdown::Char('e'), Markdown::Char('d'), Markdown::Char('i'), Markdown::Char('a'), Markdown::Char('.'), Markdown::Char('o'), Markdown::Char('r'), Markdown::Char('g'), Markdown::Char('/'), Markdown::Char('w'), Markdown::Char('i'), Markdown::Char('k'), Markdown::Char('i'), Markdown::Char('/'), Markdown::Char('('), Markdown::Char(')'),
			])));
	}
}