organic/src/parser/lesser_block.rs

use crate::error::Res;
use nom::branch::alt;
use nom::bytes::complete::is_not;
use nom::bytes::complete::tag_no_case;
use nom::character::complete::line_ending;
use nom::character::complete::space0;
use nom::character::complete::space1;
use nom::combinator::consumed;
use nom::combinator::eof;
use nom::combinator::map;
use nom::combinator::opt;
use nom::combinator::verify;
use nom::multi::many_till;
use nom::sequence::tuple;

use super::Context;
use crate::parser::exiting::ExitClass;
use crate::parser::lesser_element::CommentBlock;
use crate::parser::lesser_element::ExampleBlock;
use crate::parser::lesser_element::ExportBlock;
use crate::parser::lesser_element::SrcBlock;
use crate::parser::lesser_element::VerseBlock;
use crate::parser::object::Object;
use crate::parser::object::PlainText;
use crate::parser::object_parser::standard_set_object;
use crate::parser::parser_context::ContextElement;
use crate::parser::parser_context::ExitMatcherNode;
use crate::parser::parser_with_context::parser_with_context;
use crate::parser::plain_text::plain_text;
use crate::parser::util::blank_line;
use crate::parser::util::exit_matcher_parser;
use crate::parser::util::get_consumed;

use crate::parser::util::start_of_line;

#[tracing::instrument(ret, level = "debug")]
pub fn verse_block<'r, 's>(
    context: Context<'r, 's>,
    input: &'s str,
) -> Res<&'s str, VerseBlock<'s>> {
    let (remaining, name) = lesser_block_begin("verse")(context, input)?;
    let (remaining, parameters) = opt(tuple((space1, data)))(remaining)?;
    let (remaining, _nl) = line_ending(remaining)?;
    let lesser_block_end_specialized = lesser_block_end("verse");
    let parser_context = context
        .with_additional_node(ContextElement::ConsumeTrailingWhitespace(true))
        .with_additional_node(ContextElement::Context("lesser block"))
        .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
            class: ExitClass::Beta,
            exit_matcher: &lesser_block_end_specialized,
        }));
    let parameters = match parameters {
        Some((_ws, parameters)) => Some(parameters),
        None => None,
    };

    let object_matcher = parser_with_context!(standard_set_object)(&parser_context);
    let exit_matcher = parser_with_context!(exit_matcher_parser)(&parser_context);
    // Check for a completely empty block
    let (remaining, children) = match consumed(many_till(blank_line, exit_matcher))(remaining) {
        Ok((remaining, (whitespace, (_children, _exit_contents)))) => (
            remaining,
            vec![Object::PlainText(PlainText { source: whitespace })],
        ),
        Err(_) => {
            let (remaining, (children, _exit_contents)) =
                many_till(object_matcher, exit_matcher)(remaining)?;
            (remaining, children)
        }
    };
    let (remaining, _end) = lesser_block_end_specialized(&parser_context, remaining)?;

    let source = get_consumed(input, remaining);
    Ok((
        remaining,
        VerseBlock {
            source,
            name,
            data: parameters,
            children,
        },
    ))
}

#[tracing::instrument(ret, level = "debug")]
pub fn comment_block<'r, 's>(
    context: Context<'r, 's>,
    input: &'s str,
) -> Res<&'s str, CommentBlock<'s>> {
    let (remaining, name) = lesser_block_begin("comment")(context, input)?;
    let (remaining, parameters) = opt(tuple((space1, data)))(remaining)?;
    let (remaining, _nl) = line_ending(remaining)?;
    let lesser_block_end_specialized = lesser_block_end("comment");
    let parser_context = context
        .with_additional_node(ContextElement::ConsumeTrailingWhitespace(true))
        .with_additional_node(ContextElement::Context("lesser block"))
        .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
            class: ExitClass::Beta,
            exit_matcher: &lesser_block_end_specialized,
        }));
    let parameters = match parameters {
        Some((_ws, parameters)) => Some(parameters),
        None => None,
    };

    let (remaining, contents) = map(parser_with_context!(plain_text)(&parser_context), |obj| {
        obj.source
    })(remaining)?;
    let (remaining, _end) = lesser_block_end_specialized(&parser_context, remaining)?;

    let source = get_consumed(input, remaining);
    Ok((
        remaining,
        CommentBlock {
            source,
            name,
            data: parameters,
            contents,
        },
    ))
}

#[tracing::instrument(ret, level = "debug")]
pub fn example_block<'r, 's>(
    context: Context<'r, 's>,
    input: &'s str,
) -> Res<&'s str, ExampleBlock<'s>> {
    let (remaining, name) = lesser_block_begin("example")(context, input)?;
    let (remaining, parameters) = opt(tuple((space1, data)))(remaining)?;
    let (remaining, _nl) = line_ending(remaining)?;
    let lesser_block_end_specialized = lesser_block_end("example");
    let parser_context = context
        .with_additional_node(ContextElement::ConsumeTrailingWhitespace(true))
        .with_additional_node(ContextElement::Context("lesser block"))
        .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
            class: ExitClass::Beta,
            exit_matcher: &lesser_block_end_specialized,
        }));
    let parameters = match parameters {
        Some((_ws, parameters)) => Some(parameters),
        None => None,
    };

    let (remaining, contents) = map(parser_with_context!(plain_text)(&parser_context), |obj| {
        obj.source
    })(remaining)?;
    let (remaining, _end) = lesser_block_end_specialized(&parser_context, remaining)?;

    let source = get_consumed(input, remaining);
    Ok((
        remaining,
        ExampleBlock {
            source,
            name,
            data: parameters,
            contents,
        },
    ))
}

#[tracing::instrument(ret, level = "debug")]
pub fn export_block<'r, 's>(
    context: Context<'r, 's>,
    input: &'s str,
) -> Res<&'s str, ExportBlock<'s>> {
    let (remaining, name) = lesser_block_begin("export")(context, input)?;
    // https://orgmode.org/worg/org-syntax.html#Blocks claims that export blocks must have a single word for data but testing shows no data and multi-word data still parses as an export block.
    let (remaining, parameters) = opt(tuple((space1, data)))(remaining)?;
    let (remaining, _nl) = line_ending(remaining)?;
    let lesser_block_end_specialized = lesser_block_end("export");
    let parser_context = context
        .with_additional_node(ContextElement::ConsumeTrailingWhitespace(true))
        .with_additional_node(ContextElement::Context("lesser block"))
        .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
            class: ExitClass::Beta,
            exit_matcher: &lesser_block_end_specialized,
        }));
    let parameters = match parameters {
        Some((_ws, parameters)) => Some(parameters),
        None => None,
    };

    let (remaining, contents) = map(parser_with_context!(plain_text)(&parser_context), |obj| {
        obj.source
    })(remaining)?;
    let (remaining, _end) = lesser_block_end_specialized(&parser_context, remaining)?;

    let source = get_consumed(input, remaining);
    Ok((
        remaining,
        ExportBlock {
            source,
            name,
            data: parameters,
            contents,
        },
    ))
}

#[tracing::instrument(ret, level = "debug")]
pub fn src_block<'r, 's>(context: Context<'r, 's>, input: &'s str) -> Res<&'s str, SrcBlock<'s>> {
    let (remaining, name) = lesser_block_begin("src")(context, input)?;
    // https://orgmode.org/worg/org-syntax.html#Blocks claims that data is mandatory and must follow the LANGUAGE SWITCHES ARGUMENTS pattern but testing has shown that no data and incorrect data here will still parse to a src block.
    let (remaining, parameters) = opt(tuple((space1, data)))(remaining)?;
    let (remaining, _nl) = line_ending(remaining)?;
    let lesser_block_end_specialized = lesser_block_end("src");
    let parser_context = context
        .with_additional_node(ContextElement::ConsumeTrailingWhitespace(true))
        .with_additional_node(ContextElement::Context("lesser block"))
        .with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
            class: ExitClass::Beta,
            exit_matcher: &lesser_block_end_specialized,
        }));
    let parameters = match parameters {
        Some((_ws, parameters)) => Some(parameters),
        None => None,
    };

    let (remaining, contents) = map(parser_with_context!(plain_text)(&parser_context), |obj| {
        obj.source
    })(remaining)?;
    let (remaining, _end) = lesser_block_end_specialized(&parser_context, remaining)?;

    let source = get_consumed(input, remaining);
    Ok((
        remaining,
        SrcBlock {
            source,
            name,
            data: parameters,
            contents,
        },
    ))
}

#[tracing::instrument(ret, level = "debug")]
fn name<'s>(input: &'s str) -> Res<&'s str, &'s str> {
    is_not(" \t\r\n")(input)
}

#[tracing::instrument(ret, level = "debug")]
fn data<'s>(input: &'s str) -> Res<&'s str, &'s str> {
    is_not("\r\n")(input)
}

fn lesser_block_end(
    current_name: &str,
) -> impl for<'r, 's> Fn(Context<'r, 's>, &'s str) -> Res<&'s str, &'s str> {
    let current_name_lower = current_name.to_lowercase();
    move |context: Context, input: &str| {
        start_of_line(context, input)?;
        let (remaining, _leading_whitespace) = space0(input)?;
        let (remaining, (_begin, _name, _ws)) = tuple((
            tag_no_case("#+end_"),
            tag_no_case(current_name_lower.as_str()),
            alt((eof, line_ending)),
        ))(remaining)?;
        let source = get_consumed(input, remaining);
        Ok((remaining, source))
    }
}

fn lesser_block_begin(
    current_name: &str,
) -> impl for<'r, 's> Fn(Context<'r, 's>, &'s str) -> Res<&'s str, &'s str> {
    let current_name_lower = current_name.to_lowercase();
    move |context: Context, input: &str| {
        start_of_line(context, input)?;
        let (remaining, _leading_whitespace) = space0(input)?;
        let (remaining, (_begin, name)) = tuple((
            tag_no_case("#+begin_"),
            verify(name, |name: &str| {
                name.to_lowercase().as_str() == current_name_lower
            }),
        ))(remaining)?;
        Ok((remaining, name))
    }
}