organic/src/parser/text_element_parser.rs

212 lines
7.4 KiB
Rust
Raw Normal View History

2022-07-16 03:26:49 +00:00
//! A single element of text.
2022-07-17 01:32:23 +00:00
use crate::parser::parser_with_context::parser_with_context;
use crate::parser::text::paragraph_end;
use super::combinator::context_many1;
use super::combinator::context_many_till;
use super::error::CustomError;
use super::error::MyError;
use super::nom_context::ChainBehavior;
use super::nom_context::ContextElement;
use super::nom_context::ContextTree;
2022-12-16 02:24:53 +00:00
use super::nom_context::ExitMatcherNode;
2022-12-04 05:53:59 +00:00
use super::nom_context::PreviousElementNode;
2022-07-17 01:55:33 +00:00
use super::text::bold_end;
use super::text::bold_start;
2022-07-16 03:26:49 +00:00
use super::text::line_break;
use super::text::link_end;
use super::text::link_start;
2022-07-16 03:26:49 +00:00
use super::text::space;
use super::text::span;
use super::text::symbol;
2022-07-17 01:55:33 +00:00
use super::text::Bold;
use super::text::Link;
use super::text::Res;
2022-07-16 03:26:49 +00:00
use super::text::TextElement;
2022-12-04 04:53:52 +00:00
use super::token::Token;
use super::Context;
2022-07-16 03:26:49 +00:00
use nom::branch::alt;
use nom::bytes::complete::tag;
2022-12-11 07:07:12 +00:00
use nom::bytes::complete::take;
use nom::combinator::cond;
2022-11-26 23:25:53 +00:00
use nom::combinator::eof;
2022-07-16 03:26:49 +00:00
use nom::combinator::map;
use nom::combinator::not;
use nom::combinator::peek;
2022-07-17 01:55:33 +00:00
use nom::combinator::recognize;
use nom::error::ErrorKind;
use nom::error::ParseError;
2022-12-16 03:38:28 +00:00
use nom::error::VerboseError;
use nom::multi::many1;
2022-07-17 01:32:23 +00:00
use nom::multi::many_till;
2022-07-17 01:55:33 +00:00
use nom::sequence::tuple;
2022-07-16 03:26:49 +00:00
use nom::IResult;
2022-11-27 00:46:59 +00:00
use nom::InputLength;
2022-10-15 00:17:48 +00:00
type UnboundMatcher<'r, 's, I, O, E> = dyn Fn(Context<'r, 's>, I) -> IResult<I, O, E>;
pub fn document(input: &str) -> Res<&str, Vec<(Vec<TextElement>, &str)>> {
let initial_context: ContextTree<'_, '_> = ContextTree::new();
let ret = context_many1(&initial_context, paragraph)(input);
ret
}
2022-12-04 04:13:21 +00:00
pub fn context_paragraph_end<'s, 'r>(
context: Context<'r, 's>,
2022-12-04 04:13:21 +00:00
input: &'s str,
) -> Res<&'s str, &'s str> {
2022-12-04 04:07:16 +00:00
paragraph_end(input)
}
2022-12-11 07:07:12 +00:00
fn can_start_bold<'s, 'r>(context: Context<'r, 's>) -> bool {
2022-12-16 02:57:21 +00:00
_preceded_by_whitespace(context) && !_in_section(context, "bold")
}
fn _in_section<'s, 'r, 'x>(context: Context<'r, 's>, section_name: &'x str) -> bool {
for thing in context.iter() {
match thing.get_data() {
ContextElement::ExitMatcherNode(_) => {}
ContextElement::PreviousElementNode(_) => {}
ContextElement::Context(name) if *name == section_name => return true,
ContextElement::Context(_) => {}
ContextElement::StartOfParagraph => {} // TODO: If we specialize this to bold then this would be a good spot to stop scanning
}
}
false
}
fn _preceded_by_whitespace<'s, 'r>(context: Context<'r, 's>) -> bool {
2022-12-11 07:07:12 +00:00
let mut context_iterator = context.iter().enumerate();
loop {
if let Some((i, ctx)) = context_iterator.next() {
match ctx.get_data() {
2022-12-16 02:24:53 +00:00
ContextElement::ExitMatcherNode(_) => {}
2022-12-11 07:07:12 +00:00
ContextElement::PreviousElementNode(previous_element_node) => {
match &previous_element_node.element {
Token::TextElement(text_element) => {
match text_element {
TextElement::Span(_) => return false,
TextElement::Space(_) => return true,
TextElement::LineBreak(_) => return true,
TextElement::Symbol(_) => return false,
TextElement::Bold(_) => return false,
TextElement::Link(_) => return false,
};
2022-12-16 01:32:00 +00:00
}
2022-12-11 07:07:12 +00:00
};
2022-12-16 01:32:00 +00:00
}
2022-12-11 07:07:12 +00:00
ContextElement::StartOfParagraph => {
return true;
2022-12-16 01:32:00 +00:00
}
2022-12-16 02:57:21 +00:00
ContextElement::Context(_) => {}
2022-12-11 07:07:12 +00:00
}
} else {
break;
}
}
false
}
pub fn context_bold_start<'s, 'r>(
context: Context<'r, 's>,
input: &'s str,
) -> Res<&'s str, &'s str> {
2022-12-11 07:07:12 +00:00
if can_start_bold(context) {
recognize(bold_start)(input)
} else {
// TODO: Make this a specific error instead of just a generic MyError
return Err(nom::Err::Error(CustomError::MyError(MyError(
"Cannot start bold",
))));
2022-12-11 07:07:12 +00:00
}
}
2022-12-11 05:33:40 +00:00
pub fn context_bold_end<'s, 'r>(context: Context<'r, 's>, input: &'s str) -> Res<&'s str, &'s str> {
let (remaining, actual_match) = recognize(bold_end)(input)?;
peek(alt((
2022-12-16 02:24:53 +00:00
// Must have whitespace after the end asterisk or it must be the end of that section (as checked by the exit matcher)
tag(" "),
tag("\t"),
tag("\n"),
2022-12-16 02:24:53 +00:00
|i| context.check_exit_matcher(i),
)))(remaining)?;
Ok((remaining, actual_match))
2022-12-11 05:33:40 +00:00
}
pub fn paragraph<'s, 'r>(
context: Context<'r, 's>,
2022-11-27 00:26:48 +00:00
i: &'s str,
) -> Res<&'s str, (Vec<TextElement<'s>>, &'s str)> {
// Add a not(eof) check because many_till cannot match a zero-length string
not(eof)(i)?;
2022-12-11 05:21:30 +00:00
let paragraph_context = context
.with_additional_node(ContextElement::StartOfParagraph)
2022-12-16 02:24:53 +00:00
.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
exit_matcher: ChainBehavior::AndParent(Some(&context_paragraph_end)),
2022-12-04 02:11:39 +00:00
}));
2022-12-11 04:57:39 +00:00
let (remaining, (many, till)) =
context_many_till(&paragraph_context, flat_text_element, context_paragraph_end)(i)?;
let many = many
.into_iter()
.filter_map(|token| match token {
Token::TextElement(text_element) => Some(text_element),
})
.collect();
Ok((remaining, (many, till)))
}
fn flat_text_element<'s, 'r>(
context: Context<'r, 's>,
i: &'s str,
) -> Res<&'s str, TextElement<'s>> {
2022-12-16 02:24:53 +00:00
not(|i| context.check_exit_matcher(i))(i)?;
let bold_matcher = parser_with_context!(flat_bold)(&context);
let link_matcher = parser_with_context!(flat_link)(&context);
alt((
2022-11-25 23:55:56 +00:00
map(bold_matcher, TextElement::Bold),
map(link_matcher, TextElement::Link),
map(span, TextElement::Span),
map(symbol("*"), TextElement::Symbol),
map(symbol("["), TextElement::Symbol),
map(symbol("]"), TextElement::Symbol),
map(space, TextElement::Space),
map(line_break, TextElement::LineBreak),
))(i)
2022-10-15 00:17:48 +00:00
}
2022-10-15 18:16:52 +00:00
fn flat_bold<'s, 'r>(context: Context<'r, 's>, i: &'s str) -> Res<&'s str, Bold<'s>> {
let bold_start = parser_with_context!(context_bold_start)(&context);
2022-12-16 02:57:21 +00:00
let nom_context = context
.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
2022-12-16 02:24:53 +00:00
exit_matcher: ChainBehavior::AndParent(Some(&context_bold_end)),
2022-12-16 02:57:21 +00:00
}))
.with_additional_node(ContextElement::Context("bold"));
let (remaining, captured) = recognize(tuple((bold_start, |i| {
context_many_till(&nom_context, flat_text_element, context_bold_end)(i)
})))(i)?;
2022-11-25 23:55:56 +00:00
let ret = Bold { contents: captured };
Ok((remaining, ret))
2022-10-15 18:28:24 +00:00
}
2022-12-16 01:32:00 +00:00
fn recognize_link_end<'s, 'r>(context: Context<'r, 's>, input: &'s str) -> Res<&'s str, &'s str> {
recognize(link_end)(input)
}
fn flat_link<'s, 'r>(context: Context<'r, 's>, i: &'s str) -> Res<&'s str, Link<'s>> {
let nom_context =
2022-12-16 02:24:53 +00:00
context.with_additional_node(ContextElement::ExitMatcherNode(ExitMatcherNode {
exit_matcher: ChainBehavior::AndParent(Some(&recognize_link_end)),
2022-12-04 02:11:39 +00:00
}));
2022-12-16 02:24:53 +00:00
// let nom_context = context.with_additional_exit_matcher(&recognize_link_end);
let text_element_parser = parser_with_context!(flat_text_element)(&nom_context);
let (remaining, captured) = recognize(tuple((
link_start,
many_till(text_element_parser, link_end),
)))(i)?;
2022-11-25 23:55:56 +00:00
let ret = Link { contents: captured };
Ok((remaining, ret))
}