organic/src/parser/lesser_block.rs

434 lines
17 KiB
Rust
Raw Normal View History

2023-04-21 20:57:49 +00:00
use nom::branch::alt;
use nom::bytes::complete::is_not;
use nom::bytes::complete::tag;
2023-04-21 20:57:49 +00:00
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;
2023-04-21 20:57:49 +00:00
use nom::combinator::opt;
use nom::combinator::recognize;
2023-04-21 20:57:49 +00:00
use nom::combinator::verify;
use nom::multi::many_till;
use nom::multi::separated_list1;
2023-04-21 20:57:49 +00:00
use nom::sequence::tuple;
use super::org_source::OrgSource;
2023-09-03 15:05:34 +00:00
use crate::context::parser_with_context;
use crate::context::ContextElement;
2023-09-03 16:28:45 +00:00
use crate::context::ContextMatcher;
2023-09-03 15:05:34 +00:00
use crate::context::ExitClass;
use crate::context::ExitMatcherNode;
use crate::context::RefContext;
use crate::error::Res;
2023-04-21 20:57:49 +00:00
use crate::parser::object_parser::standard_set_object;
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;
use crate::parser::util::text_until_exit;
2023-09-03 15:05:34 +00:00
use crate::types::CommentBlock;
use crate::types::ExampleBlock;
use crate::types::ExportBlock;
use crate::types::LineNumber;
2023-09-03 15:05:34 +00:00
use crate::types::Object;
use crate::types::PlainText;
use crate::types::SrcBlock;
use crate::types::SwitchNumberLines;
2023-09-03 15:05:34 +00:00
use crate::types::VerseBlock;
2023-08-11 00:04:59 +00:00
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
2023-09-11 17:13:28 +00:00
pub(crate) fn verse_block<'b, 'g, 'r, 's>(
context: RefContext<'b, 'g, 'r, 's>,
input: OrgSource<'s>,
) -> Res<OrgSource<'s>, VerseBlock<'s>> {
2023-04-21 22:01:06 +00:00
let (remaining, name) = lesser_block_begin("verse")(context, input)?;
2023-04-21 20:57:49 +00:00
let (remaining, parameters) = opt(tuple((space1, data)))(remaining)?;
let (remaining, _nl) = recognize(tuple((space0, line_ending)))(remaining)?;
2023-04-21 22:01:06 +00:00
let lesser_block_end_specialized = lesser_block_end("verse");
2023-09-03 15:05:34 +00:00
let contexts = [
ContextElement::ConsumeTrailingWhitespace(true),
ContextElement::Context("lesser block"),
ContextElement::ExitMatcherNode(ExitMatcherNode {
class: ExitClass::Alpha,
exit_matcher: &lesser_block_end_specialized,
2023-09-03 15:05:34 +00:00
}),
];
2023-09-03 16:28:45 +00:00
let parser_context = context.with_additional_node(&contexts[0]);
let parser_context = parser_context.with_additional_node(&contexts[1]);
let parser_context = parser_context.with_additional_node(&contexts[2]);
2023-04-21 20:57:49 +00:00
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) {
2023-04-21 22:01:06 +00:00
Ok((remaining, (whitespace, (_children, _exit_contents)))) => (
remaining,
vec![Object::PlainText(PlainText {
source: whitespace.into(),
})],
2023-04-21 22:01:06 +00:00
),
2023-04-21 20:57:49 +00:00
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)?;
2023-04-21 20:57:49 +00:00
let source = get_consumed(input, remaining);
Ok((
remaining,
VerseBlock {
source: source.into(),
name: name.into(),
data: parameters.map(|parameters| Into::<&str>::into(parameters)),
2023-04-21 20:57:49 +00:00
children,
},
))
}
2023-08-11 00:04:59 +00:00
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
2023-09-11 17:13:28 +00:00
pub(crate) fn comment_block<'b, 'g, 'r, 's>(
context: RefContext<'b, 'g, 'r, 's>,
input: OrgSource<'s>,
) -> Res<OrgSource<'s>, CommentBlock<'s>> {
let (remaining, name) = lesser_block_begin("comment")(context, input)?;
let (remaining, _parameters) = opt(tuple((space1, data)))(remaining)?;
let (remaining, _nl) = recognize(tuple((space0, line_ending)))(remaining)?;
let lesser_block_end_specialized = lesser_block_end("comment");
2023-09-03 15:05:34 +00:00
let contexts = [
ContextElement::ConsumeTrailingWhitespace(true),
ContextElement::Context("lesser block"),
ContextElement::ExitMatcherNode(ExitMatcherNode {
class: ExitClass::Alpha,
exit_matcher: &lesser_block_end_specialized,
2023-09-03 15:05:34 +00:00
}),
];
2023-09-03 16:28:45 +00:00
let parser_context = context.with_additional_node(&contexts[0]);
let parser_context = parser_context.with_additional_node(&contexts[1]);
let parser_context = parser_context.with_additional_node(&contexts[2]);
let (remaining, contents) = parser_with_context!(text_until_exit)(&parser_context)(remaining)?;
let (remaining, _end) = lesser_block_end_specialized(&parser_context, remaining)?;
let source = get_consumed(input, remaining);
Ok((
remaining,
CommentBlock {
source: source.into(),
name: name.into(),
contents: contents.into(),
},
))
}
2023-08-11 00:04:59 +00:00
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
2023-09-11 17:13:28 +00:00
pub(crate) fn example_block<'b, 'g, 'r, 's>(
context: RefContext<'b, 'g, 'r, 's>,
input: OrgSource<'s>,
) -> Res<OrgSource<'s>, ExampleBlock<'s>> {
2023-08-24 22:39:55 +00:00
let (remaining, _name) = lesser_block_begin("example")(context, input)?;
let (remaining, parameters) = opt(tuple((space1, example_switches)))(remaining)?;
let (remaining, _nl) = recognize(tuple((space0, line_ending)))(remaining)?;
let lesser_block_end_specialized = lesser_block_end("example");
2023-09-03 16:28:45 +00:00
let contexts = [
ContextElement::ConsumeTrailingWhitespace(true),
ContextElement::Context("lesser block"),
ContextElement::ExitMatcherNode(ExitMatcherNode {
class: ExitClass::Alpha,
exit_matcher: &lesser_block_end_specialized,
2023-09-03 16:28:45 +00:00
}),
];
let parser_context = context.with_additional_node(&contexts[0]);
let parser_context = parser_context.with_additional_node(&contexts[1]);
let parser_context = parser_context.with_additional_node(&contexts[2]);
let parameters = parameters.map(|(_, parameters)| parameters);
let (remaining, contents) = parser_with_context!(text_until_exit)(&parser_context)(remaining)?;
let (remaining, _end) = lesser_block_end_specialized(&parser_context, remaining)?;
let source = get_consumed(input, remaining);
let (switches, number_lines, retain_labels, use_labels, label_format) = {
if let Some(parameters) = parameters {
(
Some(parameters.source),
parameters.number_lines,
parameters.retain_labels,
parameters.use_labels,
parameters.label_format,
)
} else {
(None, None, true, true, None)
}
};
Ok((
remaining,
ExampleBlock {
source: source.into(),
name: source.into(),
switches,
number_lines,
retain_labels,
use_labels,
label_format,
contents: contents.into(),
},
))
}
2023-08-11 00:04:59 +00:00
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
2023-09-11 17:13:28 +00:00
pub(crate) fn export_block<'b, 'g, 'r, 's>(
context: RefContext<'b, 'g, 'r, 's>,
input: OrgSource<'s>,
) -> Res<OrgSource<'s>, ExportBlock<'s>> {
let (remaining, name) = lesser_block_begin("export")(context, input)?;
2023-04-21 22:11:26 +00:00
// 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) = recognize(tuple((space0, line_ending)))(remaining)?;
let lesser_block_end_specialized = lesser_block_end("export");
2023-09-03 16:28:45 +00:00
let contexts = [
ContextElement::ConsumeTrailingWhitespace(true),
ContextElement::Context("lesser block"),
ContextElement::ExitMatcherNode(ExitMatcherNode {
class: ExitClass::Alpha,
exit_matcher: &lesser_block_end_specialized,
2023-09-03 16:28:45 +00:00
}),
];
let parser_context = context.with_additional_node(&contexts[0]);
let parser_context = parser_context.with_additional_node(&contexts[1]);
let parser_context = parser_context.with_additional_node(&contexts[2]);
let parameters = match parameters {
Some((_ws, parameters)) => Some(parameters),
None => None,
};
let (remaining, contents) = parser_with_context!(text_until_exit)(&parser_context)(remaining)?;
let (remaining, _end) = lesser_block_end_specialized(&parser_context, remaining)?;
let source = get_consumed(input, remaining);
Ok((
remaining,
ExportBlock {
source: source.into(),
name: name.into(),
data: parameters.map(|parameters| Into::<&str>::into(parameters)),
contents: contents.into(),
},
))
}
2023-08-11 00:04:59 +00:00
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
2023-09-11 17:13:28 +00:00
pub(crate) fn src_block<'b, 'g, 'r, 's>(
context: RefContext<'b, 'g, 'r, 's>,
input: OrgSource<'s>,
) -> Res<OrgSource<'s>, SrcBlock<'s>> {
let (remaining, name) = lesser_block_begin("src")(context, input)?;
2023-04-21 22:13:23 +00:00
// 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) = recognize(tuple((space0, line_ending)))(remaining)?;
let lesser_block_end_specialized = lesser_block_end("src");
2023-09-03 16:28:45 +00:00
let contexts = [
ContextElement::ConsumeTrailingWhitespace(true),
ContextElement::Context("lesser block"),
ContextElement::ExitMatcherNode(ExitMatcherNode {
class: ExitClass::Alpha,
exit_matcher: &lesser_block_end_specialized,
2023-09-03 16:28:45 +00:00
}),
];
let parser_context = context.with_additional_node(&contexts[0]);
let parser_context = parser_context.with_additional_node(&contexts[1]);
let parser_context = parser_context.with_additional_node(&contexts[2]);
let parameters = match parameters {
Some((_ws, parameters)) => Some(parameters),
None => None,
};
let (remaining, contents) = parser_with_context!(text_until_exit)(&parser_context)(remaining)?;
let (remaining, _end) = lesser_block_end_specialized(&parser_context, remaining)?;
let source = get_consumed(input, remaining);
Ok((
remaining,
SrcBlock {
source: source.into(),
name: name.into(),
switches: parameters.map(|parameters| Into::<&str>::into(parameters)),
contents: contents.into(),
},
))
}
2023-08-11 00:04:59 +00:00
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
fn name<'s>(input: OrgSource<'s>) -> Res<OrgSource<'s>, OrgSource<'s>> {
2023-04-21 20:57:49 +00:00
is_not(" \t\r\n")(input)
}
2023-08-11 00:04:59 +00:00
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
fn data<'s>(input: OrgSource<'s>) -> Res<OrgSource<'s>, OrgSource<'s>> {
2023-04-21 20:57:49 +00:00
is_not("\r\n")(input)
}
fn lesser_block_end<'c>(current_name: &'c str) -> impl ContextMatcher + 'c {
// Since the lesser block names are statically defined in code, we can simply assert that the name is lowercase instead of causing an allocation by converting to lowercase.
debug_assert!(current_name == current_name.to_lowercase());
move |context, input: OrgSource<'_>| _lesser_block_end(context, input, current_name)
}
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
fn _lesser_block_end<'b, 'g, 'r, 's, 'c>(
_context: RefContext<'b, 'g, 'r, 's>,
input: OrgSource<'s>,
current_name_lower: &'c str,
) -> Res<OrgSource<'s>, OrgSource<'s>> {
start_of_line(input)?;
let (remaining, _leading_whitespace) = space0(input)?;
let (remaining, (_begin, _name, _ws, _ending)) = tuple((
tag_no_case("#+end_"),
tag_no_case(current_name_lower),
space0,
alt((eof, line_ending)),
))(remaining)?;
let source = get_consumed(input, remaining);
Ok((remaining, source))
}
/// Parser for the beginning of a lesser block
///
/// current_name MUST be lowercase. We do not do the conversion ourselves because it is not allowed in a const fn.
2023-09-03 16:28:45 +00:00
const fn lesser_block_begin<'c>(current_name: &'c str) -> impl ContextMatcher + 'c {
// TODO: Since this is a const fn, is there ANY way to "generate" functions at compile time?
2023-09-03 15:05:34 +00:00
move |context, input: OrgSource<'_>| _lesser_block_begin(context, input, current_name)
}
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
fn _lesser_block_begin<'b, 'g, 'r, 's, 'c>(
_context: RefContext<'b, 'g, 'r, 's>,
input: OrgSource<'s>,
current_name_lower: &'c str,
) -> Res<OrgSource<'s>, OrgSource<'s>> {
start_of_line(input)?;
let (remaining, _leading_whitespace) = space0(input)?;
let (remaining, (_begin, name)) = tuple((
tag_no_case("#+begin_"),
verify(name, |name: &OrgSource<'_>| {
Into::<&str>::into(name).to_lowercase().as_str() == current_name_lower
}),
))(remaining)?;
Ok((remaining, name))
}
struct ExampleSwitches<'s> {
source: &'s str,
number_lines: Option<SwitchNumberLines>,
retain_labels: bool,
use_labels: bool,
label_format: Option<&'s str>,
}
enum SwitchState {
Normal,
NewLineNumber,
ContinuedLineNumber,
LabelFormat,
}
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
fn example_switches<'s>(input: OrgSource<'s>) -> Res<OrgSource<'s>, ExampleSwitches<'s>> {
let mut number_lines = None;
let mut retain_labels = true;
let mut use_labels = true;
let mut label_format = None;
let (remaining, (source, (words, _))) = consumed(tuple((
separated_list1(space1, map(switch_word, |val| Into::<&str>::into(val))),
space0,
)))(input)?;
let mut state = SwitchState::Normal;
for word in words {
loop {
match (&state, word) {
(SwitchState::Normal, "-n") => {
state = SwitchState::NewLineNumber;
}
(SwitchState::Normal, "+n") => {
state = SwitchState::ContinuedLineNumber;
}
(SwitchState::Normal, "-r") => {
retain_labels = false;
use_labels = false;
}
(SwitchState::Normal, "-l") => {
state = SwitchState::LabelFormat;
}
(SwitchState::NewLineNumber, _) => {
let val = word.parse::<LineNumber>();
if let Ok(val) = val {
if val < 0 {
number_lines = Some(SwitchNumberLines::New(0));
} else {
// Note that this can result in a negative 1 if the val is originally 0.
number_lines = Some(SwitchNumberLines::New(val - 1));
}
state = SwitchState::Normal;
} else {
number_lines = Some(SwitchNumberLines::New(0));
state = SwitchState::Normal;
continue; // Re-processes the word
}
}
(SwitchState::ContinuedLineNumber, _) => {
let val = word.parse::<LineNumber>();
if let Ok(val) = val {
if val < 0 {
number_lines = Some(SwitchNumberLines::Continued(0));
} else {
// Note that this can result in a negative 1 if the val is originally 0.
number_lines = Some(SwitchNumberLines::Continued(val - 1));
}
state = SwitchState::Normal;
} else {
number_lines = Some(SwitchNumberLines::Continued(0));
state = SwitchState::Normal;
continue; // Re-processes the word
}
}
(SwitchState::LabelFormat, _) => {
label_format = Some(word);
state = SwitchState::Normal;
}
(SwitchState::Normal, _) => {}
};
break;
}
}
Ok((
remaining,
ExampleSwitches {
source: Into::<&str>::into(source),
number_lines,
retain_labels,
use_labels,
label_format,
},
))
}
#[cfg_attr(feature = "tracing", tracing::instrument(ret, level = "debug"))]
fn switch_word<'s>(input: OrgSource<'s>) -> Res<OrgSource<'s>, OrgSource<'s>> {
alt((
map(
tuple((tag(r#"""#), is_not("\"\r\n"), tag(r#"""#))),
|(_, contents, _)| contents,
),
is_not(" \t\r\n"),
))(input)
}