organic/src/parser/paragraph.rs

use nom::branch::alt;
use nom::combinator::eof;
use nom::combinator::recognize;
use nom::combinator::verify;
use nom::multi::many0;
use nom::multi::many1;
use nom::multi::many_till;
use nom::sequence::tuple;

use super::element_parser::detect_element;
use super::keyword::affiliated_keyword;
use super::org_source::OrgSource;
use super::util::blank_line;
use super::util::get_consumed;
use super::util::get_has_affiliated_keyword;
use super::util::get_name;
use super::util::maybe_consume_trailing_whitespace_if_not_exiting;
use crate::context::parser_with_context;
use crate::context::ContextElement;
use crate::context::ExitClass;
use crate::context::ExitMatcherNode;
use crate::context::HasAffiliatedKeywordInner;
use crate::context::RefContext;
use crate::error::CustomError;
use crate::error::MyError;
use crate::error::Res;
use crate::parser::object_parser::standard_set_object;
use crate::parser::util::exit_matcher_parser;
use crate::parser::util::start_of_line;
use crate::types::Paragraph;

#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
pub(crate) fn paragraph<'b, 'g, 'r, 's>(
    context: RefContext<'b, 'g, 'r, 's>,
    input: OrgSource<'s>,
) -> Res<OrgSource<'s>, Paragraph<'s>> {
    let (remaining, affiliated_keywords) = many0(affiliated_keyword)(input)?;
    let contexts = [
        ContextElement::HasAffiliatedKeyword(HasAffiliatedKeywordInner {
            start_after_affiliated_keywords: remaining,
            keywords: &affiliated_keywords,
        }),
        ContextElement::ExitMatcherNode(ExitMatcherNode {
            class: ExitClass::Gamma,
            exit_matcher: &paragraph_end,
        }),
    ];
    let parser_context = context.with_additional_node(&contexts[0]);
    let parser_context = parser_context.with_additional_node(&contexts[1]);
    let standard_set_object_matcher = parser_with_context!(standard_set_object)(&parser_context);
    let exit_matcher = parser_with_context!(exit_matcher_parser)(&parser_context);

    let (remaining, (children, _exit_contents)) = verify(
        many_till(standard_set_object_matcher, exit_matcher),
        |(children, _exit_contents)| !children.is_empty(),
    )(remaining)?;

    // Not checking parent exit matcher because if there are any children matched then we have a valid paragraph.

    let (remaining, _trailing_ws) =
        maybe_consume_trailing_whitespace_if_not_exiting(context, remaining)?;
    let source = get_consumed(input, remaining);

    Ok((
        remaining,
        Paragraph {
            source: source.into(),
            name: get_name(&affiliated_keywords),
            children,
        },
    ))
}

#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
fn paragraph_end<'b, 'g, 'r, 's>(
    context: RefContext<'b, 'g, 'r, 's>,
    input: OrgSource<'s>,
) -> Res<OrgSource<'s>, OrgSource<'s>> {
    let regular_end = recognize(tuple((start_of_line, many1(blank_line))))(input);
    if regular_end.is_ok() {
        return regular_end;
    }
    match get_has_affiliated_keyword(context) {
        Some(start_post_affiliated_keywords) if input == start_post_affiliated_keywords => {
            return Err(nom::Err::Error(CustomError::MyError(MyError(
                "No exit due to affiliated keywords.",
            ))));
        }
        _ => {}
    }
    // Check to see if input is the start of a HasAffiliatedKeyword
    alt((
        recognize(parser_with_context!(detect_element(false))(context)),
        eof,
    ))(input)
}

#[cfg(test)]
mod tests {
    use crate::context::parser_with_context;
    use crate::context::Context;
    use crate::context::ContextElement;
    use crate::context::GlobalSettings;
    use crate::context::List;
    use crate::parser::element_parser::element;
    use crate::parser::org_source::OrgSource;
    use crate::types::GetStandardProperties;

    #[test]
    fn two_paragraphs() {
        let input = OrgSource::new("foo bar baz\n\nlorem ipsum");
        let global_settings = GlobalSettings::default();
        let initial_context = ContextElement::document_context();
        let initial_context = Context::new(&global_settings, List::new(&initial_context));
        let paragraph_matcher = parser_with_context!(element(true))(&initial_context);
        let (remaining, first_paragraph) = paragraph_matcher(input).expect("Parse first paragraph");
        let (remaining, second_paragraph) =
            paragraph_matcher(remaining).expect("Parse second paragraph.");
        assert_eq!(Into::<&str>::into(remaining), "");
        assert_eq!(
            first_paragraph.get_standard_properties().get_source(),
            "foo bar baz\n\n"
        );
        assert_eq!(
            second_paragraph.get_standard_properties().get_source(),
            "lorem ipsum"
        );
    }
}