mirror of
https://github.com/ankitects/anki.git
synced 2025-09-18 14:02:21 -04:00
Compare commits
4 commits
1d53ead2a4
...
255f614dd2
Author | SHA1 | Date | |
---|---|---|---|
![]() |
255f614dd2 | ||
![]() |
db695b1ad8 | ||
![]() |
f8c0a40681 | ||
![]() |
91fad11528 |
2 changed files with 42 additions and 10 deletions
|
@ -242,6 +242,7 @@ Lee Doughty <32392044+leedoughty@users.noreply.github.com>
|
|||
memchr <memchr@proton.me>
|
||||
Max Romanowski <maxr777@proton.me>
|
||||
Aldlss <ayaldlss@gmail.com>
|
||||
Matbe766 <matildabergstrom01@gmail.com>
|
||||
|
||||
********************
|
||||
|
||||
|
|
|
@ -64,19 +64,22 @@ pub fn extract_latex(text: &str, svg: bool) -> (Cow<'_, str>, Vec<ExtractedLatex
|
|||
let mut extracted = vec![];
|
||||
|
||||
let new_text = LATEX.replace_all(text, |caps: &Captures| {
|
||||
// [latex] blocks preserve newlines between <br>/<div>
|
||||
// [$] and [$$] blocks remove <br>/<div> completely so that
|
||||
// no blank lines are introduced inside math environments.
|
||||
let latex = match (caps.get(1), caps.get(2), caps.get(3)) {
|
||||
(Some(m), _, _) => m.as_str().into(),
|
||||
(_, Some(m), _) => format!("${}$", m.as_str()),
|
||||
(_, _, Some(m)) => format!(r"\begin{{displaymath}}{}\end{{displaymath}}", m.as_str()),
|
||||
(Some(m), _, _) => strip_html_for_latex(m.as_str()).into(),
|
||||
(_, Some(m), _) => format!("${}$", strip_html_for_latex_math(m.as_str())),
|
||||
(_, _, Some(m)) => format!(
|
||||
r"\begin{{displaymath}}{}\end{{displaymath}}",
|
||||
strip_html_for_latex_math(m.as_str())
|
||||
),
|
||||
_ => unreachable!(),
|
||||
};
|
||||
let latex_text = strip_html_for_latex(&latex);
|
||||
let fname = fname_for_latex(&latex_text, svg);
|
||||
let img_link = image_link_for_fname(&latex_text, &fname);
|
||||
extracted.push(ExtractedLatex {
|
||||
fname,
|
||||
latex: latex_text.into(),
|
||||
});
|
||||
|
||||
let fname = fname_for_latex(&latex, svg);
|
||||
let img_link = image_link_for_fname(&latex, &fname);
|
||||
extracted.push(ExtractedLatex { fname, latex });
|
||||
|
||||
img_link
|
||||
});
|
||||
|
@ -96,6 +99,20 @@ fn strip_html_for_latex(html: &str) -> Cow<'_, str> {
|
|||
out
|
||||
}
|
||||
|
||||
/// Removes HTML breaks (<br>, <div>) from math blocks instead of
|
||||
/// converting them to newlines. This prevents LaTeX environments
|
||||
/// from being broken by unintended blank lines.
|
||||
fn strip_html_for_latex_math(html: &str) -> Cow<'_, str> {
|
||||
let mut out: Cow<str> = html.into();
|
||||
if let Cow::Owned(o) = LATEX_NEWLINES.replace_all(html, "") {
|
||||
out = o.into();
|
||||
}
|
||||
if let Cow::Owned(o) = strip_html(out.as_ref()) {
|
||||
out = o.into();
|
||||
}
|
||||
out
|
||||
}
|
||||
|
||||
fn fname_for_latex(latex: &str, svg: bool) -> String {
|
||||
let ext = if svg { "svg" } else { "png" };
|
||||
let csum = hex::encode(sha1_of_data(latex.as_bytes()));
|
||||
|
@ -146,4 +163,18 @@ mod test {
|
|||
}]
|
||||
);
|
||||
}
|
||||
|
||||
/// Ensures that math blocks do not contain unintended blank lines
|
||||
/// when <br> or <div> elements are present in the HTML.
|
||||
#[test]
|
||||
fn no_blank_lines_in_math() {
|
||||
let input = "[$$]\\begin{tikzcd}x & y\\end{tikzcd}[/$$]";
|
||||
let (_, extracts) = extract_latex(input, false);
|
||||
let extracted = &extracts[0].latex;
|
||||
|
||||
assert!(
|
||||
!extracted.contains("\n\n"),
|
||||
"Should not contain blank lines inside math"
|
||||
);
|
||||
}
|
||||
}
|
||||
|
|
Loading…
Reference in a new issue