rust-parser-combinator/src/lib.rs

#![feature(assert_matches)]
#![allow(dead_code)] //TODO eventually turn this off
mod bnf;
mod parser;
mod primitives;
mod sequence;

use parser::{ParseResult, Parser};

fn map<P, F, I, O1, O2, E>(parser: P, map_fn: F) -> impl Parser<I, O2, E>
where
    P: Parser<I, O1, E>,
    F: Fn(O1) -> O2,
{
    move |input| {
        parser
            .parse(input)
            .map(|(result, rest)| (map_fn(result), rest))
    }
}

fn pred<P, F, I, O>(parser: P, pred_fn: F) -> impl Parser<I, O, I>
where
    P: Parser<I, O, I>,
    F: Fn(&O) -> bool,
{
    move |input| {
        parser.parse(input).and_then(|(result, rest)| {
            if pred_fn(&result) {
                Ok((result, rest))
            } else {
                Err(rest)
            }
        })
    }
}

fn zero_or_more<P, I, O>(parser: P) -> impl Parser<I, Vec<O>, I>
where
    P: Parser<I, O, I>,
    I: Copy,
{
    move |mut input| {
        let mut results = Vec::new();

        while let Ok((item, rest)) = parser.parse(input) {
            results.push(item);
            input = rest;
        }

        Ok((results, input))
    }
}

fn one_or_more<P, I, O>(parser: P) -> impl Parser<I, Vec<O>, I>
where
    P: Parser<I, O, I> + 'static,
    I: Copy + 'static,
    O: 'static,
{
    let parser = std::rc::Rc::new(parser);
    parser
        .clone()
        .then(zero_or_more(parser))
        .map(|(first, rest)| {
            let mut output = vec![first];
            output.extend(rest.into_iter());
            output
        })
}

fn any_char(input: &str) -> ParseResult<&str, char, &str> {
    match input.chars().next() {
        Some(ch) => Ok((ch, &input[ch.len_utf8()..])),
        None => Err(input),
    }
}

fn choice<P1, P2, I, O, E>(parser1: P1, parser2: P2) -> impl Parser<I, O, E>
where
    P1: Parser<I, O, E>,
    P2: Parser<I, O, E>,
    I: Copy,
{
    move |input| match parser1.parse(input) {
        ok @ Ok(..) => ok,
        Err(_e) => parser2.parse(input),
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use crate::primitives::literal;
    use std::assert_matches::assert_matches;
    use std::collections::HashMap;

    #[test]
    fn test_parsing() {
        let output = literal("a")("a yolo");
        assert_matches!(output.unwrap(), ("a", " yolo"));
    }

    #[test]
    fn test_map() {
        let lit_a = literal("a");
        let output = lit_a.map(|s| s.to_uppercase()).parse("a yolo");
        assert_matches!(output.unwrap(), (s, " yolo") if s == "A");
    }

    #[test]
    fn test_one_or_more() {
        let p = one_or_more(literal("bongo "));
        let input = "bongo bongo bongo bongo bongo ";
        assert_matches!(p.parse(input), Ok((v, "")) if v.len() == 5);
        let input = "bongo ecks";
        assert_matches!(p.parse(input), Ok((v, "ecks")) if v.len() == 1);
    }

    #[test]
    fn test_choice() {
        let p = choice(literal("gnostika").to(1), one_or_more(literal(" ")).to(2));
        assert_eq!(p.parse("gnostika twentynine"), Ok((1, " twentynine")));
    }

    /*
     * JSON BNF
     * <JSON>     ::= <value>
    <value>    ::= <object> | <array> | <boolean> | <string> | <number> | <null>
    <array>    ::= "[" [<value>] {"," <value>}* "]"
    <object>   ::= "{" [<property>] {"," <property>}* "}"
    <property> ::= <string> ":" <value>
    */
    #[derive(Debug, Clone)]
    enum JsonValue {
        Null,
        Bool(bool),
        Str(String),
        Num(f64),
        Array(Vec<JsonValue>),
        Object(HashMap<String, JsonValue>),
    }

    #[test]
    fn parse_json() {
        let json_null = literal("null").to(JsonValue::Null);
        let json_true = literal("true").to(JsonValue::Bool(true));
        let json_false = literal("false").to(JsonValue::Bool(false));

        let json_value = choice(json_null, choice(json_true, json_false));

        assert_matches!(json_value.parse("true"), Ok((JsonValue::Bool(true), "")));
    }
}