Merge PR #654 from NikitkaRa: PPTX: shape text anchors at the top when the file gives no anchor; write ctr for centred shapes

crates/pptx/src/read/mod.rs

@@ -657,6 +657,7 @@ pub fn import(bytes: &[u8]) -> Result<Presentation, PptxError> {
     p.kept_parts = kept_parts(&pkg, &pres_path);
 
     inherit_placeholder_geometry(&mut p);
+    default_text_anchor(&mut p);
     p.header_footer = header_footer(&p);
     p.embedded_fonts = read_embedded_fonts(&mut imp, &pres_part, pres_el);
     p.media = std::mem::take(&mut imp.media);
@@ -748,6 +749,53 @@ pub fn color_map(cm: &El) -> Vec<(String, String)> {
 /// geometry of the layout or master placeholder they inherit from, as PowerPoint draws them: a
 /// layout's picture placeholder cut into a shape keeps that shape on the slide. Anything left
 /// is a rectangle.
+/// The anchor the theme's shape defaults (`a:objectDefaults/a:spDef/a:bodyPr`) give new shapes
+/// text, if any.
+fn theme_shape_anchor(t: &deckcraft_model::Theme) -> Option<deckcraft_model::text::Anchor> {
+    let raw = t.raw_extra.as_deref()?;
+    let sp = &raw[raw.find("spDef")?..];
+    let sp = &sp[..sp.find("/a:spDef").or_else(|| sp.find("txDef")).unwrap_or(sp.len())];
+    let b = &sp[sp.find("bodyPr")?..];
+    let b = &b[..b.find('>').unwrap_or(b.len())];
+    let v = &b[b.find("anchor=\"")? + 8..];
+    deckcraft_model::text::Anchor::from_xml(&v[..v.find('"')?])
+}
+
+/// A shape's text without an anchor of its own sits at the top, as the standard has it (or where
+/// the theme's shape defaults put it). The model's own default for shapes is the middle, so a
+/// top or bottom anchor is made explicit here and a middle one left unset (placeholders and text
+/// boxes keep inheriting).
+fn default_text_anchor(p: &mut Presentation) {
+    fn fix(shapes: &mut [deckcraft_model::Shape], a: deckcraft_model::text::Anchor) {
+        walk_shapes_mut(shapes, &mut |s| {
+            if s.ph.is_none()
+                && !s.text_box
+                && !matches!(s.kind, deckcraft_model::ShapeKind::Table(_))
+                && let Some(t) = &mut s.text
+            {
+                // Middle is the model's own default for shapes, so it is kept as unset.
+                let got = t.body.anchor.unwrap_or(a);
+                t.body.anchor = (got != deckcraft_model::text::Anchor::Middle).then_some(got);
+            }
+        });
+    }
+    let top = deckcraft_model::text::Anchor::Top;
+    let anchors: Vec<_> = p.masters.iter().map(|m| (m.id, theme_shape_anchor(&m.theme).unwrap_or(top))).collect();
+    for m in &mut p.masters {
+        let a = anchors.iter().find(|(id, _)| *id == m.id).map(|x| x.1).unwrap_or(top);
+        let m = Arc::make_mut(m);
+        fix(&mut m.shapes, a);
+        for l in &mut m.layouts {
+            fix(&mut l.shapes, a);
+        }
+    }
+    let per_slide: Vec<_> =
+        p.slides.iter().map(|s| p.master_for(s).and_then(|(m, _)| anchors.iter().find(|(id, _)| *id == m.id)).map(|x| x.1).unwrap_or(top)).collect();
+    for (s, a) in p.slides.iter_mut().zip(per_slide) {
+        fix(&mut Arc::make_mut(s).shapes, a);
+    }
+}
+
 fn inherit_placeholder_geometry(p: &mut Presentation) {
     let unset = |g: &Geom| matches!(g, Geom::Custom { paths } if paths.is_empty());
     // Masters first, then layouts (from their masters), then slides (from layouts and masters).

crates/pptx/src/write/shapes.rs

@@ -646,7 +646,15 @@ fn sp(w: &mut W, x: &mut Exp, o: &mut Out, id: u32, s: &Shape, math: bool) {
         style(w, st);
     }
     if let Some(tb) = &s.text {
-        dml::text_body(w, x, o, "p:txBody", tb, math);
+        // A shape's text without its own anchor sits in the middle in the model; a file without
+        // one would put it at the top, so say so.
+        if s.ph.is_none() && !s.text_box && tb.body.anchor.is_none() {
+            let mut tb = tb.clone();
+            tb.body.anchor = Some(deckcraft_model::text::Anchor::Middle);
+            dml::text_body(w, x, o, "p:txBody", &tb, math);
+        } else {
+            dml::text_body(w, x, o, "p:txBody", tb, math);
+        }
     }
     w.close("p:sp");
 }

crates/pptx/tests/anchor.rs

@@ -0,0 +1,74 @@
+//! Text anchoring of shapes without an `anchor` of their own: the model centres a shape's text,
+//! a file without one puts it at the top (or where the theme's shape defaults say).
+
+use std::io::{Cursor, Read, Write};
+
+use deckcraft_model::text::Anchor;
+use deckcraft_model::{Presentation, ShapeKind};
+use zip::write::SimpleFileOptions;
+
+fn sample() -> Presentation {
+    let mut s = deckcraft_engine::Session::new();
+    deckcraft_engine::sample::open_sample(&mut s).expect("sample");
+    s.doc().expect("doc").doc.as_ref().clone()
+}
+
+/// Every part of `bytes` with `edit` applied to the XML ones.
+fn rewrite(bytes: &[u8], edit: impl Fn(&str, String) -> String) -> Vec<u8> {
+    let mut src = zip::ZipArchive::new(Cursor::new(bytes)).expect("zip");
+    let mut out = zip::ZipWriter::new(Cursor::new(Vec::new()));
+    for i in 0..src.len() {
+        let mut f = src.by_index(i).expect("entry");
+        let name = f.name().to_string();
+        let mut data = Vec::new();
+        f.read_to_end(&mut data).expect("read");
+        if name.ends_with(".xml") {
+            data = edit(&name, String::from_utf8(data).expect("utf8")).into_bytes();
+        }
+        out.start_file(name, SimpleFileOptions::default()).expect("start");
+        out.write_all(&data).expect("write");
+    }
+    out.finish().expect("finish").into_inner()
+}
+
+/// Anchors of the shapes with text that aren't placeholders, text boxes or tables.
+fn anchors(p: &Presentation) -> Vec<Option<Anchor>> {
+    p.slides
+        .iter()
+        .flat_map(|s| s.shapes.iter())
+        .filter(|s| s.ph.is_none() && !s.text_box && !matches!(s.kind, ShapeKind::Table(_)))
+        .filter_map(|s| s.text.as_ref().map(|t| t.body.anchor))
+        .collect()
+}
+
+#[test]
+fn shapes_keep_their_middle_anchor_and_files_without_one_anchor_at_the_top() {
+    let p = sample();
+    assert!(anchors(&p).contains(&None), "the sample has shapes with the model's default anchor");
+    let bytes = deckcraft_pptx::export(&p).expect("export");
+
+    // Written explicitly, so PowerPoint (which reads a missing anchor as top) centres them too.
+    let back = deckcraft_pptx::import(&bytes).expect("import");
+    let a = anchors(&back);
+    assert_eq!(a, anchors(&p), "the middle anchor survives (as the model's default)");
+
+    // The same file without the anchors: their text sits at the top.
+    let bare = rewrite(&bytes, |n, x| if n.starts_with("ppt/slides/slide") { x.replace(r#" anchor="ctr""#, "") } else { x });
+    let a = anchors(&deckcraft_pptx::import(&bare).expect("import"));
+    assert!(a.iter().all(|x| *x == Some(Anchor::Top)), "{a:?}");
+
+    // ...unless the theme's shape defaults centre them, as Office's themes do.
+    let themed = rewrite(&bare, |n, x| {
+        if n.starts_with("ppt/theme/") {
+            x.replace(
+                "<a:objectDefaults/>",
+                r#"<a:objectDefaults><a:spDef><a:spPr/><a:bodyPr rtlCol="0" anchor="ctr"/><a:lstStyle/></a:spDef></a:objectDefaults>"#,
+            )
+        } else {
+            x
+        }
+    });
+    let a = anchors(&deckcraft_pptx::import(&themed).expect("import"));
+    assert!(a.iter().all(|x| x.is_none() || *x == Some(Anchor::Middle)), "{a:?}");
+    assert!(a.iter().filter(|x| x.is_none()).count() >= 3, "{a:?}");
+}