organic/src/parser/plain_list.rs

use nom::branch::alt;
use nom::bytes::complete::tag;
use nom::character::complete::anychar;
use nom::character::complete::digit1;
use nom::character::complete::line_ending;
use nom::character::complete::one_of;
use nom::character::complete::space0;
use nom::combinator::consumed;
use nom::combinator::not;
use nom::combinator::opt;
use nom::combinator::peek;
use nom::combinator::recognize;
use nom::combinator::verify;
use nom::multi::many0_count;
use nom::multi::many1;
use nom::sequence::tuple;

use super::combinator::context_many_till;
use super::error::CustomError;
use super::error::MyError;
use super::error::Res;
use super::paragraph::paragraph_end;
use super::parser_context::ContextElement;
use super::parser_with_context::parser_with_context;
use super::text::space;
use super::text::text_element;
use super::token::ListItem;
use super::token::PlainList;
use super::token::Token;
use super::Context;

pub fn plain_list<'r, 's>(context: Context<'r, 's>, i: &'s str) -> Res<&'s str, PlainList<'s>> {
    // todo
    todo!()
}

pub fn item<'r, 's>(context: Context<'r, 's>, i: &'s str) -> Res<&'s str, ListItem<'s>> {
    let (remaining, leading_whitespace) = space0(i)?;
    let indent_level = leading_whitespace.len();
    let list_item_context = context.with_additional_node(ContextElement::ListItem(indent_level));
    let (remaining, (bul, countset, check, tg, sp, (contents, end))) = tuple((
        bullet,
        opt(tuple((space, counter_set))),
        opt(tuple((space, check_box))),
        opt(tuple((space, item_tag))),
        space,
        context_many_till(&list_item_context, text_element, item_end),
    ))(remaining)?;

    let elements = contents
        .into_iter()
        .filter_map(|token| match token {
            Token::TextElement(text_element) => Some(text_element),
            Token::Paragraph(_) => panic!("There should only be text elements in items."),
        })
        .collect();

    let source = {
        let offset = remaining.as_ptr() as usize - i.as_ptr() as usize;
        &i[..offset]
    };

    let ret = ListItem {
        source,
        leading_whitespace,
        bullet: bul,
        counter_set: countset.map(|(_spc, count)| count),
        check_box: check.map(|(_spc, check)| check),
        item_tag: tg.map(|(_spc, tg)| tg),
        contents: elements,
    };
    Ok((remaining, ret))
}

fn counter<'s>(i: &'s str) -> Res<&'s str, &'s str> {
    alt((recognize(one_of("abcdefghijklmnopqrstuvwxyz")), digit1))(i)
}

fn bullet<'s>(i: &'s str) -> Res<&'s str, &'s str> {
    alt((
        tag("*"),
        tag("-"),
        tag("+"),
        recognize(tuple((counter, alt((tag("."), tag(")")))))),
    ))(i)
}

fn counter_set<'s>(i: &'s str) -> Res<&'s str, &'s str> {
    recognize(tuple((tag("[@"), counter, tag("]"))))(i)
}

fn check_box<'s>(i: &'s str) -> Res<&'s str, &'s str> {
    recognize(alt((tag("[ ]"), tag("[X]"), tag("[-]"))))(i)
}

fn item_tag<'s>(i: &'s str) -> Res<&'s str, &'s str> {
    recognize(tuple((tag_text, tag_separator)))(i)
}

fn tag_text<'s>(i: &'s str) -> Res<&'s str, &'s str> {
    recognize(many1(tag_text_character))(i)
}

fn tag_text_character<'s>(i: &'s str) -> Res<&'s str, &'s str> {
    not(alt((tag_separator, line_ending)))(i)?;
    recognize(anychar)(i)
}

fn tag_separator<'s>(i: &'s str) -> Res<&'s str, &'s str> {
    tag(" :: ")(i)
}

pub fn item_end<'r, 's>(context: Context<'r, 's>, i: &'s str) -> Res<&'s str, &'s str> {
    let item_matcher = parser_with_context!(item)(&context);
    let line_indented_matcher = parser_with_context!(line_indented_lte)(&context);
    alt((
        paragraph_end,
        recognize(tuple((line_ending, peek(line_indented_matcher)))),
        // TODO: Do we still need the item_matcher entry here? If we remove it, then child items should become part of the body of the parent item which would match the description on https://orgmode.org/worg/org-syntax.html
        recognize(tuple((line_ending, peek(item_matcher)))),
    ))(i)
}

fn line_indented_lte<'r, 's>(context: Context<'r, 's>, i: &'s str) -> Res<&'s str, &'s str> {
    let current_item_indent_level: &usize = get_context_item_indent(context).ok_or(
        nom::Err::Error(CustomError::MyError(MyError("NotInPlainListItem"))),
    )?;

    let matched = recognize(verify(
        tuple((space0::<&str, _>, anychar)),
        |(_space0, _anychar)| _space0.len() <= *current_item_indent_level,
    ))(i)?;

    Ok(matched)
}

fn get_context_item_indent<'r, 's>(context: Context<'r, 's>) -> Option<&'r usize> {
    for thing in context.iter() {
        match thing.get_data() {
            ContextElement::ListItem(depth) => return Some(depth),
            _ => {}
        };
    }
    None
}