Merge pull request #1101 from p0lyg0n/fix/pdf-japanese-spacing

PDF/.ai import: glyphs spaced each their own way keep their place

crates/pdf/src/import_text.rs

@@ -136,6 +136,8 @@ pub(crate) struct TextLine {
     /// The spaces read from gaps between glyphs: their byte offset in the text, and the gap from
     /// the end of the glyph before to the start of the next (points).
     gaps: Vec<(usize, f64)>,
+    /// Each glyph's advance in the file's font (points), with [`Self::baseline`].
+    advances: Vec<f64>,
 }
 
 /// What a finished line of type tells beyond its text object (see [`crate::import_lines`]).
@@ -168,6 +170,7 @@ impl TextLine {
             column: false,
             ink: vec![],
             gaps: vec![],
+            advances: vec![],
         }
     }
 
@@ -228,7 +231,8 @@ impl TextLine {
                 && at.dir.dot(self.at.dir) > 0.9995
                 && (at.slant - self.at.slant).abs() < 1e-3
                 && (at.origin - self.at.origin).dot(self.at.up()).abs() < size * 0.15
-                && gap > -size * 0.3
+                // Proportional widths close CJK marks up by up to half an em.
+                && gap > -size * 0.6
                 && gap < size * 3.0;
             // Or on a curve: turned a little from the glyph before it, and starting about where
             // that one ends.
@@ -263,6 +267,7 @@ impl TextLine {
         self.last = at.origin;
         self.last_dir = at.dir;
         self.baseline.push(at.origin);
+        self.advances.push(advance);
         true
     }
 
@@ -329,6 +334,14 @@ impl TextLine {
         // The glyphs' tops lean along the baseline (+x), up being -y in the type's own space.
         t.xf = Affine::translate(self.at.origin.to_vec2()) * Affine::rotate(angle) * Affine::skew(-self.at.slant, 0.0);
         let db = vectorcraft_text::FontDb::global();
+        // Glyphs spaced each their own way (character tightening, optical kerning) keep their
+        // place: each character its own tracking.
+        let own = (self.upright.is_none() && t.para.direction.is_none())
+            .then(|| own_spacing(&t, &self.baseline, &self.advances, path.as_ref(), self.at.dir, self.at.size))
+            .flatten();
+        if let Some(own) = &own {
+            t.runs = tracked_each(std::mem::take(&mut t.runs), own, self.at.size);
+        }
         if let Some(path) = path {
             t.kind = TextKind::OnPath { path: vectorcraft_geom::PathData::from_bezpath(&path), start: 0.0, end: None };
             t.xf = Affine::IDENTITY;
@@ -358,7 +371,7 @@ impl TextLine {
         let chars: usize = t.runs.iter().map(|r| r.text.chars().count()).sum();
         let size = self.at.size;
         let mut bounds = laid.bounds;
-        if chars > 1 && length.is_finite() && (length - natural).abs() > size * 0.01 {
+        if own.is_none() && chars > 1 && length.is_finite() && (length - natural).abs() > size * 0.01 {
             // The spaces read from gaps keep their own place only in text left in the file's order.
             let total: usize = t.runs.iter().map(|r| r.text.len()).sum();
             let gaps: Vec<(usize, f64)> =
@@ -373,6 +386,59 @@ impl TextLine {
     }
 }
 
+/// The space each glyph of `t` (laid out untracked) leaves after its advance to put the next one
+/// where the file did (points; the last glyph's is the one before it), when that varies by more
+/// than a fiftieth of an em: one tracking value can't place them. Glyph origins are `baseline`,
+/// measured along `path` (type on a path: one curve per glyph) or along `dir`. None when the
+/// glyphs can't be matched to the characters one to one (spaces, ligatures), or when the font
+/// laid out isn't the file's (its advances differ from `advances`): spacing measured against a
+/// stand-in's widths would keep the stand-in's differences once the font is installed.
+fn own_spacing(t: &TextObject, baseline: &[Point], advances: &[f64], path: Option<&BezPath>, dir: Vec2, size: f64) -> Option<Vec<f64>> {
+    let text: String = t.runs.iter().map(|r| r.text.as_str()).collect();
+    let chars = text.chars().count();
+    if chars < 3 || text.contains(' ') || baseline.len() != chars || advances.len() != chars {
+        return None;
+    }
+    let laid = vectorcraft_text::layout(vectorcraft_text::FontDb::global(), t);
+    let advs: Vec<f64> = laid.glyphs.iter().map(|g| g.advance).collect();
+    if advs.len() != chars || advs.iter().zip(advances).any(|(a, b)| (a - b).abs() > size * 0.01) {
+        return None;
+    }
+    let dist: Vec<f64> = match path {
+        Some(bp) => bp.segments().take(chars - 1).map(|seg| kurbo::ParamCurveArclen::arclen(&seg, 1e-3)).collect(),
+        None => baseline.windows(2).map(|w| (w[1] - w[0]).dot(dir)).collect(),
+    };
+    if dist.len() != chars - 1 {
+        return None;
+    }
+    let mut extra: Vec<f64> = dist.iter().zip(&advs).map(|(d, a)| d - a).collect();
+    let mean = extra.iter().sum::<f64>() / extra.len() as f64;
+    if extra.iter().all(|e| (e - mean).abs() <= size * 0.02) {
+        return None;
+    }
+    extra.push(*extra.last()?);
+    Some(extra)
+}
+
+/// `runs` with each character's tracking from `own` (points, one per character), characters with
+/// the same tracking kept in one run.
+fn tracked_each(runs: Vec<TextRun>, own: &[f64], size: f64) -> Vec<TextRun> {
+    let em = |pts: f64| (pts / size * 1000.0).clamp(-1000.0, 1000.0).round();
+    let mut out: Vec<TextRun> = vec![];
+    let mut k = 0;
+    for r in runs {
+        for c in r.text.chars() {
+            let style = CharStyle { tracking: own.get(k).map_or(0.0, |e| em(*e)), ..r.style.clone() };
+            k += 1;
+            match out.last_mut() {
+                Some(last) if last.style == style && last.inline == r.inline => last.text.push(c),
+                _ => out.push(TextRun { text: c.to_string(), style, inline: r.inline.clone() }),
+            }
+        }
+    }
+    out
+}
+
 /// The tracking (thousandths of an em) that sets a line of `chars` characters, laid out untracked
 /// as `laid`, `extra` points longer, as the file set it: the same after every character, except
 /// the spaces read from gaps (`gaps`: byte offset, gap in points) where that would put the glyphs

crates/pdf/src/tests_import_fidelity.rs

@@ -464,3 +464,12 @@ fn art_past_the_page_clip_keeps_the_clip() {
     assert!(!all(&r.document).iter().any(|n| matches!(n.kind, NodeKind::Group { clip: true, .. })));
     assert!(!r.warnings.iter().any(|w| w == crate::import::PAST_PAGE_NOTE), "{:?}", r.warnings);
 }
+
+/// Glyphs closed up by up to half an em (the proportional widths of CJK punctuation) still
+/// continue their line.
+#[test]
+fn glyphs_closed_up_by_half_an_em_stay_one_line() {
+    let t = texts(&open(&one_page("BT /F1 20 Tf 10 40 Td [(T) 400 (i) 450 (g) 0 (h) 0 (t)] TJ ET", HELVETICA, &[])));
+    assert_eq!(t.len(), 1, "one line: {:?}", t.iter().map(|t| t.plain_text()).collect::<Vec<_>>());
+    assert_eq!(t[0].plain_text(), "Tight");
+}

crates/pdf/src/tests_realtext.rs

@@ -329,3 +329,46 @@ fn type_in_another_installed_version_reopens_in_that_version() {
         assert_eq!(found, [("Hamburgefonstiv".to_string(), version.map(String::from))], "{version:?}");
     }
 }
+
+/// The glyph origins of the type objects of `d`, in document space, in order.
+fn origins(d: &Document) -> Vec<Point> {
+    let mut out = vec![];
+    d.walk(|n| {
+        if let NodeKind::Text(t) = &n.kind {
+            out.extend(vectorcraft_text::layout(FontDb::global(), t).glyphs.iter().map(|g| t.xf * g.origin));
+        }
+    });
+    out
+}
+
+/// Type spaced glyph by glyph (character tightening, optical kerning) comes back with each glyph
+/// in its place, not evened out by one tracking value; evenly spaced type keeps one run.
+#[test]
+fn glyphs_spaced_each_their_own_way_keep_their_place() {
+    let runs = [("T", 0.0), ("i", -120.0), ("g", 160.0), ("h", -60.0), ("tl", 0.0), ("y", 0.0)];
+    let mut t = TextObject::point(Point::new(20.0, 100.0), "", style(40.0));
+    t.runs = runs.iter().map(|(s, tr)| TextRun { text: s.to_string(), style: CharStyle { tracking: *tr, ..style(40.0) }, inline: None }).collect();
+    let d = doc(vec![t]);
+    let want = origins(&d);
+    let back = import_as(&pdf(&d, false).bytes, TextAs::default());
+    let got = origins(&back);
+    assert_eq!(got.len(), want.len(), "{:?}", texts(&back));
+    for (g, w) in got.iter().zip(&want) {
+        assert!((g.x - w.x).abs() < 0.1 && (g.y - w.y).abs() < 0.1, "{got:?}\nvs {want:?}");
+    }
+    let mut runs_back = 0;
+    back.walk(|n| {
+        if let NodeKind::Text(t) = &n.kind {
+            runs_back += t.runs.len();
+        }
+    });
+    assert!(runs_back > 1, "one tracking can't place them");
+    // Evenly tracked: one run.
+    let even = doc(vec![TextObject::point(Point::new(20.0, 100.0), "Tightly", CharStyle { tracking: -50.0, ..style(40.0) })]);
+    let back = import_as(&pdf(&even, false).bytes, TextAs::default());
+    back.walk(|n| {
+        if let NodeKind::Text(t) = &n.kind {
+            assert_eq!(t.runs.len(), 1, "{:?}", t.runs.iter().map(|r| r.style.tracking).collect::<Vec<_>>());
+        }
+    });
+}