wip on pandoc parser

Cargo.lock

@@ -463,6 +463,12 @@ dependencies = [
  "dtoa",
 ]
 
+[[package]]
+name = "either"
+version = "1.15.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "48c757948c5ede0e46177b7add2e67155f70e33c07fea8284df6576da70b3719"
+
 [[package]]
 name = "elasticlunr-rs"
 version = "3.0.2"
@@ -1039,6 +1045,15 @@ version = "1.70.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7943c866cc5cd64cbc25b2e01621d07fa8eb2a1a23160ee81ce38704e97b8ecf"
 
+[[package]]
+name = "itertools"
+version = "0.12.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ba291022dbbd398a455acf126c1e341954079855bc60dfdda641363bd6922569"
+dependencies = [
+ "either",
+]
+
 [[package]]
 name = "itoa"
 version = "1.0.15"
@@ -1226,6 +1241,8 @@ dependencies = [
  "log",
  "mdbook",
  "once_cell",
+ "pandoc",
+ "pandoc_ast",
  "path-slash",
  "pretty_assertions",
  "pulldown-cmark 0.13.0",
@@ -1388,6 +1405,26 @@ dependencies = [
  "windows-sys 0.59.0",
 ]
 
+[[package]]
+name = "pandoc"
+version = "0.8.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "463d53d1a77a4291203dbf9d461365609e6857c95bd7d807098bffdc0a02a65c"
+dependencies = [
+ "itertools",
+]
+
+[[package]]
+name = "pandoc_ast"
+version = "0.8.6"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e1ff7a92f9fb2cac7cb4b8b697efaf3a30abcb64003d4cd34a67cc1ae35194ce"
+dependencies = [
+ "serde",
+ "serde_derive",
+ "serde_json",
+]
+
 [[package]]
 name = "parking_lot"
 version = "0.12.4"

Cargo.toml

@@ -35,8 +35,10 @@ env_logger = { version = "0.11", default-features = false, optional = true }
 log = "0.4.21"
 mdbook = "0.4.40"
 once_cell = "1.19.0"
+pandoc = { version = "0.8", optional = true }
+pandoc_ast = { version = "0.8", optional = true }
 path-slash = "0.2.1"
-pulldown-cmark = "0.13"
+pulldown-cmark = { version = "0.13", optional = true }
 regex = "1.10.5"
 semver = "1.0.23"
 serde = { version = "1.0.203", features = ["derive"] }
@@ -49,7 +51,9 @@ hex_color = { version = "3.0.0", features = ["serde"] }
 pretty_assertions = "1.4.0"
 
 [features]
-default = ["cli", "cli-install"]
+default = ["cli", "cli-install", "parser-markdown"]
+parser-markdown = ["pulldown-cmark"]
+parser-pandoc = ["pandoc", "pandoc_ast"]
 
 # Enable the command line binary
 cli = ["clap", "env_logger"]

src/lib.rs

@@ -6,7 +6,10 @@ mod book_config;
 mod config;
 #[doc(hidden)]
 pub mod custom;
+#[cfg(feature = "parser-markdown")]
 mod markdown;
+#[cfg(feature = "parser-pandoc")]
+mod pandoc;
 mod parse;
 mod preprocessor;
 mod render;

src/pandoc.rs

@@ -0,0 +1,78 @@
+use mdbook::errors::Result as MdbookResult;
+
+use crate::{
+    book_config::OnFailure,
+    types::{Overrides, RenderTextMode},
+};
+
+pub(crate) fn preprocess(
+    content: &str,
+    on_failure: OnFailure,
+    overrides: &Overrides,
+    render_text_mode: RenderTextMode,
+) -> MdbookResult<String> {
+    if render_text_mode == RenderTextMode::Strip {
+        // We don't support stripping in pandoc - they just show up as divs.
+        return Ok(content.to_owned());
+    }
+
+    let ADMONISH_KIND = "note".to_owned();
+
+    let mut pandoc = pandoc::new();
+    pandoc.set_input(pandoc::InputKind::Pipe(content.into()));
+    pandoc.set_input_format(pandoc::InputFormat::CommonmarkX, Vec::new());
+    pandoc.set_output_format(pandoc::OutputFormat::CommonmarkX, Vec::new());
+    pandoc.set_output(pandoc::OutputKind::Pipe);
+    pandoc.add_filter(|json| {
+        pandoc_ast::filter(json, |mut pandoc| {
+            for block in &mut pandoc.blocks {
+                println!("{:?}", &block);
+                use pandoc_ast::Block;
+                use pandoc_ast::Inline;
+                match block {
+                    Block::Div((identifier, classes, kv), content) => {
+                        if let Some(ADMONISH_KIND) = classes.first() {
+                            // TODO: generate values for below by reading the pandoc input
+                            // decide on a syntax for kv and classes, admonition type, and plug
+                            // that in to our existing structs which deal with links and overrides
+                            // etc.
+
+                            *block = Block::Div(
+                                (
+                                    // TODO: get from kv
+                                    "admonition-title".to_owned(),
+                                    // TODO: compute from input above
+                                    vec!["admonition".to_owned(), "admonish-note".to_owned()],
+                                    vec![
+                                        ("role".to_owned(), "note".to_owned()),
+                                        (
+                                            "aria-labelledby".to_owned(),
+                                            "admonition-title-title".to_owned(),
+                                        ),
+                                    ],
+                                ),
+                                vec![
+                                    Block::Div(
+                                        // TODO: add anchor link elements (a, href, etc)
+                                        ("admonition-title-title".to_owned(), vec![], vec![]),
+                                        vec![Block::Plain(vec![Inline::Str("Title".to_owned())])],
+                                    ),
+                                    Block::Div(("".to_owned(), vec![], vec![]), content.clone()),
+                                ],
+                            )
+                        }
+                    }
+                    _ => {}
+                }
+            }
+            pandoc
+        })
+    });
+    let result = pandoc.execute().unwrap();
+    use pandoc::PandocOutput::*;
+    let output = match result {
+        ToBuffer(output) => output,
+        ToFile(_) | ToBufferRaw(_) => panic!("expected buffer output"),
+    };
+    Ok(output)
+}

src/preprocessor.rs

@@ -7,10 +7,14 @@ use mdbook::{
 
 use crate::{
     book_config::{admonish_config_from_context, Config, RenderMode},
-    markdown::preprocess,
     types::{Overrides, RenderTextMode},
 };
 
+#[cfg(feature = "parser-markdown")]
+use crate::markdown::preprocess;
+#[cfg(feature = "parser-pandoc")]
+use crate::pandoc::preprocess;
+
 pub struct Admonish;
 
 impl Preprocessor for Admonish {
@@ -166,8 +170,9 @@ mod test {
         serde_json::from_value(value).unwrap()
     }
 
+    #[cfg(feature = "parser-markdown")]
     #[test]
-    fn run_html() {
+    fn run_markdown_to_html() {
         let content = r#"
 ````admonish title="Title"
 ```rust
@@ -210,6 +215,43 @@ x = 20;
         assert_eq!(Admonish.run(&ctx, book).unwrap(), expected_book)
     }
 
+    #[cfg(feature = "parser-pandoc")]
+    #[test]
+    fn run_pandoc_to_html() {
+        let content = r#"
+:::{.note title="Title"}
+```rust
+let x = 10;
+x = 20;
+```
+:::
+"#;
+        let expected_content = r##"::::: {#admonition-title .admonition .admonish-note role="note" aria-labelledby="admonition-title-title"}
+::: {#admonition-title-title}
+Title
+:::
+
+::: {}
+``` rust
+let x = 10;
+x = 20;
+```
+:::
+:::::
+"##;
+
+        let ctx = mock_context(
+            &json!({
+                "assets_version": "3.0.0"
+            }),
+            "html",
+        );
+        let book = mock_book(content);
+        let expected_book = mock_book(expected_content);
+
+        assert_eq!(Admonish.run(&ctx, book).unwrap(), expected_book)
+    }
+
     #[test]
     fn run_test_preserves_by_default() {
         let content = r#"
@@ -232,8 +274,9 @@ x = 20;
         assert_eq!(Admonish.run(&ctx, book).unwrap(), expected_book)
     }
 
+    #[cfg(feature = "parser-markdown")]
     #[test]
-    fn run_test_can_strip() {
+    fn run_test_can_strip_markdown() {
         let content = r#"
 ````admonish title="Title"
 ```rust