Skip to content

fix(html2md): preserve visible labels and literal preformatted text #1591

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
lanyue-llk wants to merge 1 commit into from
Closed

fix(html2md): preserve visible labels and literal preformatted text #1591

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
98 changes: 95 additions & 3 deletions moli-html2md/src/machine.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,12 +9,13 @@ enum Task<'a, Id> {
Children(Option<Id>, usize),
Boundary,
PopStyle,
EndLink(usize),
EndLink(usize, Option<&'a str>),
EndQuote,
EndHeading(usize),
EndList(usize),
EndItem(String),
RawChildren(Option<Id>, usize, bool),
EndPreLink(&'a str, usize),
EndRawBlock,
EndCode,
EndPre(Option<&'a str>),
Expand All @@ -37,6 +38,7 @@ struct Machine<'a, D: Dom + ?Sized> {
lists: Vec<List>,
tables: Vec<Table<D::NodeId>>,
raw: String,
pre_links: Vec<(String, &'a str)>,
serial: usize,
}

Expand All @@ -49,6 +51,7 @@ pub(crate) fn convert<D: Dom + ?Sized>(dom: &D, root: D::NodeId, options: &Optio
lists: Vec::new(),
tables: Vec::new(),
raw: String::new(),
pre_links: Vec::new(),
serial: 0,
};
machine.run();
Expand All @@ -70,7 +73,9 @@ impl<'a, D: Dom + ?Sized> Machine<'a, D> {
Task::Children(None, _) | Task::RawChildren(None, _, _) => {}
Task::Boundary => self.writer().boundary(2),
Task::PopStyle => self.writer().pop_style(),
Task::EndLink(serial) => self.writer().end_link(serial),
Task::EndLink(serial, fallback_label) => {
self.writer().end_link(serial, fallback_label)
}
Task::EndQuote => {
let content = self.take_writer().into_string();
let mut quote = String::new();
Expand Down Expand Up @@ -132,6 +137,9 @@ impl<'a, D: Dom + ?Sized> Machine<'a, D> {
self.writer().block(item.into(), separation, separation);
}
Task::RawChildren(Some(node), depth, inline) => self.raw_node(node, depth, inline),
Task::EndPreLink(href, start) => {
self.pre_links.push((self.raw[start..].to_owned(), href));
}
Task::EndRawBlock => {
if !self.raw.is_empty() && !self.raw.ends_with('\n') {
self.raw.push('\n');
Expand Down Expand Up @@ -166,6 +174,7 @@ impl<'a, D: Dom + ?Sized> Machine<'a, D> {
let writer = self.writers.pop().expect("capture has an output");
if let Some(parent) = self.writers.last_mut() {
parent.last_link = parent.last_link.max(writer.last_link);
parent.last_link_content = parent.last_link_content.max(writer.last_link_content);
}
writer.finish()
}
Expand Down Expand Up @@ -232,7 +241,10 @@ impl<'a, D: Dom + ?Sized> Machine<'a, D> {
title: self.dom.attribute(node, "title"),
};
if self.writer().push_style(style) {
self.tasks.push(Task::EndLink(serial));
self.tasks.push(Task::EndLink(
serial,
self.dom.attribute(node, "aria-label"),
Comment thread
lanyue-llk marked this conversation as resolved.
));
}
}
}
Expand All @@ -248,8 +260,35 @@ impl<'a, D: Dom + ?Sized> Machine<'a, D> {
}
return;
}
"video" | "audio" => {
Comment thread
lanyue-llk marked this conversation as resolved.
self.writer().boundary(2);
if tag == "video"
&& let Some(poster) = self.dom.attribute(node, "poster").and_then(resource_url)
{
self.writer().image("Video poster", poster, None);
self.writer().boundary(1);
}
if let Some(src) = self.dom.attribute(node, "src").and_then(resource_url) {
self.resource_link(src, if tag == "video" { "Video" } else { "Audio" });
}
self.tasks.push(Task::Boundary);
}
"source" | "track" => {
if let Some(src) = self.dom.attribute(node, "src").and_then(resource_url) {
self.resource_link(
src,
if tag == "source" {
"Media source"
} else {
"Media track"
},
);
}
return;
}
"code" | "pre" => {
self.raw.clear();
self.pre_links.clear();
let end = if tag == "pre" {
Task::EndPre(self.language(node))
} else {
Expand Down Expand Up @@ -315,6 +354,20 @@ impl<'a, D: Dom + ?Sized> Machine<'a, D> {
}
}

fn resource_link(&mut self, href: &'a str, label: &str) {
self.serial += 1;
let serial = self.serial;
if self.writer().push_style(Style::Link {
Comment thread
lanyue-llk marked this conversation as resolved.
serial,
href,
title: None,
}) {
self.writer().text(label);
self.writer().end_link(serial, None);
}
self.writer().boundary(1);
}

fn start_list(&mut self, node: D::NodeId, ordered: bool) {
let before = if self
.lists
Expand Down Expand Up @@ -401,6 +454,12 @@ impl<'a, D: Dom + ?Sized> Machine<'a, D> {
}
self.tasks.push(Task::EndRawBlock);
}
if !inline
&& let NodeKind::Element("a") = self.dom.node_kind(node)
&& let Some(href) = self.dom.attribute(node, "href").and_then(resource_url)
{
self.tasks.push(Task::EndPreLink(href, self.raw.len()));
}
self.tasks.push(Task::RawChildren(
self.dom.first_child(node),
depth + 1,
Expand All @@ -415,6 +474,7 @@ impl<'a, D: Dom + ?Sized> Machine<'a, D> {
let text = std::mem::take(&mut self.raw);
if text.trim().is_empty() {
self.writer().boundary(2);
self.write_pre_links();
return;
}
// CommonMark code blocks normalize CRLF/CR and require a closing fence
Expand Down Expand Up @@ -446,9 +506,41 @@ impl<'a, D: Dom + ?Sized> Machine<'a, D> {
}
block.push_str(&fence);
self.writer().block(block.into(), 2, 2);
self.write_pre_links();
}

fn write_pre_links(&mut self) {
// A preformatted block may contain either source or a directory listing.
// Keep its literal text intact and expose destinations after the fence.
for (label, href) in std::mem::take(&mut self.pre_links) {
Comment thread
lanyue-llk marked this conversation as resolved.
let label = label.trim();
self.resource_link(href, if label.is_empty() { href } else { label });
}
}
}

fn resource_url(value: &str) -> Option<&str> {
Comment thread
lanyue-llk marked this conversation as resolved.
let value = value.trim();
if value.is_empty() || value.starts_with('#') || value.eq_ignore_ascii_case("about:blank") {
return None;
}
let normalized = value
.chars()
.filter(|ch| !ch.is_ascii_control())
.collect::<String>();
let Some((scheme, _)) = normalized.split_once(':') else {
return Some(value);
};
if scheme.contains(['/', '#', '?']) {
return Some(value);
}
matches!(
scheme.to_ascii_lowercase().as_str(),
"http" | "https" | "mailto" | "tel" | "ftp"
)
.then_some(value)
}

fn class_language(class: Option<&str>) -> Option<&str> {
class?
.split_ascii_whitespace()
Expand Down
24 changes: 20 additions & 4 deletions moli-html2md/src/writer.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,6 +40,7 @@ pub(crate) struct Writer<'a> {
heading: bool,
single_line_attributes: bool,
pub(crate) last_link: usize,
pub(crate) last_link_content: usize,
// Block events in this output buffer, used when finalizing a list item.
pub(crate) has_blocks: bool,
}
Expand Down Expand Up @@ -86,12 +87,21 @@ impl<'a> Writer<'a> {
self.desired.pop();
}

pub(crate) fn end_link(&mut self, serial: usize) {
pub(crate) fn end_link(&mut self, serial: usize, fallback_label: Option<&str>) {
self.flush_code();
if self.last_link < serial {
// Empty anchors still carry a destination. Materialize the link
// after its leading whitespace, even without a text event.
self.prepare_inline('[');
if let Some(label) = fallback_label
.filter(|label| self.last_link_content < serial && !label.trim().is_empty())
{
// An empty anchor can still have an author-provided accessible
// name. Preserve that name inside the link instead of emitting
// an unusable empty label. Visible descendants remain primary.
self.text(label);
} else {
// Empty anchors still carry a destination. Materialize the link
// after its leading whitespace, even without a text event.
self.prepare_inline('[');
}
}
self.pop_style();
}
Expand Down Expand Up @@ -250,6 +260,11 @@ impl<'a> Writer<'a> {
self.boundary(before.max(after));
return;
}
for style in &self.desired {
if let Style::Link { serial, .. } = style {
self.last_link_content = self.last_link_content.max(*serial);
Comment thread
lanyue-llk marked this conversation as resolved.
}
}
// Closing styles before detaching the tail keeps all inline edit
// offsets local to the current String, never inside a finished block.
self.boundary(before);
Expand Down Expand Up @@ -320,6 +335,7 @@ impl<'a> Writer<'a> {
Style::Strike => self.output.push_str(if html { "<del>" } else { "~~" }),
Style::Link { serial, .. } => {
self.last_link = self.last_link.max(serial);
self.last_link_content = self.last_link_content.max(serial);
if self.output.ends_with('!') {
self.output.insert(self.output.len() - 1, '\\');
}
Expand Down
Loading
Loading