aleleba/pdfcraft
1
0
Fork 0
mirror of https://github.com/storytold/pdfcraft.git synced 2026-10-30 13:08:48 -06:00
Code Issues Packages Projects Releases Wiki Activity

fix(render): avoid emitting full ActualText replacement when span is partially cropped (#2838)

Browse Source
This commit is contained in:
jonathanwxh-cell
2026-10-18 22:39:15 +08:00
parent 76d1bea652
commit c6bf3d0dbd
1 changed files with 15 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
crates/render/src/text.rs
+15 -4
View File
@@ -363,6 +363,7 @@ struct TextDevice {
filter: TextFilter,
excluded_depth: Option<usize>,
actual_filtered: bool,
actual_cropped: bool,
partial_replacement: bool,
}
@@ -437,6 +438,7 @@ impl TextDevice {
filter: TextFilter::default(),
excluded_depth: None,
actual_filtered: false,
actual_cropped: false,
partial_replacement: false,
}
}
@@ -451,12 +453,17 @@ impl TextDevice {
/// Unicode). An empty one (a soft hyphen at a line break, say) stands for no text at all.
fn end_actual_text(&mut self) {
let Some((_, text, drawn)) = self.actual.take() else { return };
let cropped = std::mem::take(&mut self.actual_cropped);
if std::mem::take(&mut self.actual_filtered) {
// ActualText has no mapping from replacement words to individual glyphs. A mixed
// span cannot be filtered honestly; a wholly excluded span can simply be omitted.
self.partial_replacement |= !drawn.is_empty();
return;
}
if cropped {
self.glyphs.extend(drawn.into_iter().filter(|d| !d.2.text.chars().all(|c| c.is_control())));
return;
}
let Some(glyph) = merge(&drawn, text.clone()) else {
self.glyphs.extend(drawn.into_iter().filter(|d| !d.2.text.chars().all(|c| c.is_control())));
return;
@@ -622,12 +629,14 @@ impl<'a> Device<'a> for TextDevice {
let b = corners.iter().fold(Rect::new(f64::INFINITY, f64::INFINITY, f64::NEG_INFINITY, f64::NEG_INFINITY), |b, p| {
Rect::new(b.x0.min(p.x), b.y0.min(p.y), b.x1.max(p.x), b.y1.max(p.y))
});
let outside_crop = b.x0.max(self.page.x0) >= b.x1.min(self.page.x1)
|| b.y0.max(self.page.y0) >= b.y1.min(self.page.y1);
if !(b.x0.is_finite() && b.y0.is_finite() && b.x1.is_finite() && b.y1.is_finite())
|| b.width() > 10_000.0
|| b.height() > 10_000.0
|| b.x0.max(self.page.x0) >= b.x1.min(self.page.x1)
|| b.y0.max(self.page.y0) >= b.y1.min(self.page.y1)
|| outside_crop
{
self.actual_cropped |= replaced && outside_crop;
return;
}
let length = dx.hypot(dy);
@@ -2812,6 +2821,7 @@ mod tests {
let ops = [
"BT /F1 12 Tf 1 0 0 1 100 200 Tm (Inside crop) Tj ET",
"BT /F1 12 Tf 1 0 0 1 100 150 Tm /Span <</ActualText (ActualIn)>> BDC (in) Tj EMC ET",
"BT /F1 12 Tf 1 0 0 1 5 125 Tm /Span <</ActualText (CroppedSpan KeptSpan)>> BDC (OutSpan) Tj 1 0 0 1 100 125 Tm (KeptSpan) Tj EMC ET",
"BT /F1 12 Tf 1 0 0 1 5 200 Tm (LeftOut) Tj ET",
"BT /F1 12 Tf 1 0 0 1 360 200 Tm (RightOut) Tj ET",
"BT /F1 12 Tf 1 0 0 1 100 20 Tm (BelowOut) Tj ET",
@@ -2843,10 +2853,11 @@ mod tests {
doc.save_to(&mut bytes).unwrap();
let pdf = Pdf::new(std::sync::Arc::new(bytes)).unwrap();
let text = extract_page(&pdf, 0, &InterpreterSettings::default()).unwrap();
assert_eq!(text.plain_text(), "Inside crop\nActualIn\nA", "rotate={rotate} unit={unit}");
assert_eq!(text.plain_text(), "Inside crop\nActualIn\nKeptSpan\nA", "rotate={rotate} unit={unit}");
assert!(!text.find("Inside").is_empty(), "rotate={rotate} unit={unit}");
assert!(!text.find("ActualIn").is_empty(), "rotate={rotate} unit={unit}");
for hidden in ["LeftOut", "RightOut", "BelowOut", "AboveOut", "ActualOut"] {
assert!(!text.find("KeptSpan").is_empty(), "rotate={rotate} unit={unit}");
for hidden in ["LeftOut", "RightOut", "BelowOut", "AboveOut", "ActualOut", "CroppedSpan", "OutSpan"] {
assert!(text.find(hidden).is_empty(), "{hidden} must be clipped at rotate={rotate} unit={unit}");
}
}