rust-parser-combinator/src/lib.rs

#![allow(dead_code)] //TODO eventually turn this off

type ParseResult<I, O, E> = Result<(O, I), E>;

trait Parser<I, O, E> {
    fn parse(&self, input: I) -> ParseResult<I, O, E>;
}

impl<I, O, E, F> Parser<I, O, E> for F
where
    F: Fn(I) -> ParseResult<I, O, E>,
{
    fn parse(&self, input: I) -> ParseResult<I, O, E> {
        self(input)
    }
}

fn literal(expected: &'static str) -> impl Fn(&str) -> ParseResult<&str, &str, &str> {
    move |input| match input.get(0..expected.len()) {
        Some(next) if next == expected => Ok((next, &input[expected.len()..])),
        _ => Err(input),
    }
}

fn sequence<I, O1, O2, E>(
    first: impl Parser<I, O1, E>,
    second: impl Parser<I, O2, E>,
) -> impl Parser<I, (O1, O2), E> {
    move |input| -> ParseResult<I, (O1, O2), E> {
        first.parse(input).and_then(|(result1, rest)| {
            second
                .parse(rest)
                .map(|(result2, rest2)| ((result1, result2), rest2))
        })
    }
}

fn choice<'a, I, O, E>(parsers: &'a [&'a dyn Parser<I, O, E>]) -> impl Parser<I, O, E> + 'a
where
    I: Clone,
{
    move |input: I| {
        //TODO need a more principled way to return an error when no choices work
        let mut err = None;

        for parser in parsers.iter() {
            match parser.parse(input.clone()) {
                Ok(res) => return Ok(res),
                Err(e) => {
                    err = Some(e);
                }
            }
        }
        Err(err.unwrap())
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn parsing() {
        let (parsed, rest) = literal("a")("a yolo").unwrap();
        assert_eq!(parsed, "a");
        assert_eq!(rest, " yolo");
    }

    #[test]
    fn test_sequence() {
        let parser = sequence(literal("bongo"), sequence(literal(" "), literal("jonzzz")));
        let output = parser.parse("bongo jonzzz").unwrap();
        assert_eq!(output.0 .0, "bongo");
        assert_eq!(output.0 .1, (" ", "jonzzz"));
        assert_eq!(output.1, "");
    }
}