Handle indirect and dangling structure tree root kids

Preserve indirect root kids and marked content from removed link annotations. Ignore dangling kid references instead of throwing.

src/core/editor/pdf_editor.js

@@ -530,6 +530,14 @@ class PDFEditor {
     return ref;
   }
 
+  async #resolveStructKids(rawKids, xref) {
+    if (rawKids instanceof Ref) {
+      const fetched = await xref.fetchAsync(rawKids);
+      return Array.isArray(fetched) ? fetched : [rawKids];
+    }
+    return Array.isArray(rawKids) ? rawKids : [rawKids];
+  }
+
   async #cloneStructTreeNode(
     parentStructRef,
     node,
@@ -547,19 +555,11 @@ class PDFEditor {
     if (pg instanceof Ref && !pagesMap.has(pg)) {
       return null;
     }
-    let kids;
+    const k = node.getRaw("K");
-    const k = (kids = node.getRaw("K"));
+    if (k instanceof Ref && visited.has(k)) {
-    if (k instanceof Ref) {
+      return null;
-      // We're only interested by ref referencing nodes and not an array.
-      if (visited.has(k)) {
-        return null;
-      }
-      kids = await xref.fetchAsync(k);
-      if (!Array.isArray(kids)) {
-        kids = [k];
-      }
     }
-    kids = Array.isArray(kids) ? kids : [kids];
+    const kids = await this.#resolveStructKids(k, xref);
     const newKids = [];
     const structElemIndices = [];
     for (let kid of kids) {
@@ -1523,17 +1523,24 @@ class PDFEditor {
       }
 
       // Get the kids.
-      let kids = structTreeRoot.dict.get("K");
+      const rawKids = structTreeRoot.dict.getRaw("K");
-      if (!kids) {
+      if (!rawKids) {
         continue;
       }
-      kids = Array.isArray(kids) ? kids : [kids];
+      const kids = await this.#resolveStructKids(rawKids, xref);
       for (let kid of kids) {
         const kidRef = kid instanceof Ref ? kid : null;
-        if (kidRef && removedStructElements.has(kidRef)) {
+        kid = await xref.fetchIfRefAsync(kid);
+        if (!(kid instanceof Dict)) {
           continue;
         }
-        kid = await xref.fetchIfRefAsync(kid);
+        let setAsSpan = false;
+        if (kidRef && removedStructElements.has(kidRef)) {
+          if (!isName(kid.get("S"), "Link")) {
+            continue;
+          }
+          setAsSpan = true;
+        }
         const newKidRef = await this.#cloneStructTreeNode(
           kidRef,
           kid,
@@ -1545,6 +1552,12 @@ class PDFEditor {
         );
         if (newKidRef) {
           structTreeKids.push(newKidRef);
+          if (kidRef) {
+            oldRefMapping.put(kidRef, newKidRef);
+          }
+          if (setAsSpan) {
+            this.xref[newKidRef.num].setIfName("S", "Span");
+          }
         }
       }
 

test/unit/api_spec.js

@@ -6622,6 +6622,102 @@ small scripts as well as for`);
     });
 
     describe("Struct trees", function () {
+      it("preserves an indirect StructTreeRoot kid", async function () {
+        const objects = [
+          "1 0 obj\n<< /Type /Catalog /Pages 2 0 R " +
+            "/StructTreeRoot 4 0 R >>\nendobj\n",
+          "2 0 obj\n<< /Type /Pages /Kids [3 0 R] /Count 1 >>\nendobj\n",
+          "3 0 obj\n<< /Type /Page /Parent 2 0 R /StructParents 0 " +
+            "/MediaBox [0 0 100 100] >>\nendobj\n",
+          "4 0 obj\n<< /Type /StructTreeRoot /K 5 0 R " +
+            "/ParentTree 6 0 R /ParentTreeNextKey 1 >>\nendobj\n",
+          "5 0 obj\n<< /Type /StructElem /S /P /P 4 0 R /Pg 3 0 R " +
+            "/K 0 >>\nendobj\n",
+          "6 0 obj\n<< /Nums [0 [5 0 R]] >>\nendobj\n",
+        ];
+
+        let loadingTask = getDocument({ data: assemblePdf(objects) });
+        let pdfDoc = await loadingTask.promise;
+        let page = await pdfDoc.getPage(1);
+        expect((await page.getStructTree()).children[0].role).toEqual("P");
+        const data = await pdfDoc.extractPages([{ document: null }]);
+        await loadingTask.destroy();
+
+        loadingTask = getDocument({ data });
+        pdfDoc = await loadingTask.promise;
+        page = await pdfDoc.getPage(1);
+        const tree = await page.getStructTree();
+        expect(tree.children.length).toEqual(1);
+        expect(tree.children[0].role).toEqual("P");
+        expect(tree.children[0].children[0].type).toEqual("content");
+        await loadingTask.destroy();
+      });
+
+      it("extracts pages when the StructTreeRoot /K is dangling", async function () {
+        const objects = [
+          "1 0 obj\n<< /Type /Catalog /Pages 2 0 R " +
+            "/StructTreeRoot 4 0 R >>\nendobj\n",
+          "2 0 obj\n<< /Type /Pages /Kids [3 0 R] /Count 1 >>\nendobj\n",
+          "3 0 obj\n<< /Type /Page /Parent 2 0 R " +
+            "/MediaBox [0 0 100 100] >>\nendobj\n",
+          "4 0 obj\n<< /Type /StructTreeRoot /K 99 0 R >>\nendobj\n",
+        ];
+
+        let loadingTask = getDocument({ data: assemblePdf(objects) });
+        let pdfDoc = await loadingTask.promise;
+        const data = await pdfDoc.extractPages([{ document: null }]);
+        expect(data).not.toBeNull();
+        await loadingTask.destroy();
+
+        loadingTask = getDocument({ data });
+        pdfDoc = await loadingTask.promise;
+        expect(pdfDoc.numPages).toEqual(1);
+        const page = await pdfDoc.getPage(1);
+        expect((await page.getStructTree())?.children.length ?? 0).toEqual(0);
+        await loadingTask.destroy();
+      });
+
+      it("preserves a root Link when its annotation is removed", async function () {
+        const objects = [
+          "1 0 obj\n<< /Type /Catalog /Pages 2 0 R " +
+            "/StructTreeRoot 6 0 R >>\nendobj\n",
+          "2 0 obj\n<< /Type /Pages /Kids [3 0 R 4 0 R] " +
+            "/Count 2 >>\nendobj\n",
+          "3 0 obj\n<< /Type /Page /Parent 2 0 R /StructParents 0 " +
+            "/Annots [5 0 R] /MediaBox [0 0 100 100] >>\nendobj\n",
+          "4 0 obj\n<< /Type /Page /Parent 2 0 R " +
+            "/MediaBox [0 0 100 100] >>\nendobj\n",
+          "5 0 obj\n<< /Type /Annot /Subtype /Link /Rect [0 0 10 10] " +
+            "/StructParent 1 /Dest [4 0 R /Fit] >>\nendobj\n",
+          "6 0 obj\n<< /Type /StructTreeRoot /K 7 0 R " +
+            "/ParentTree 8 0 R /ParentTreeNextKey 2 >>\nendobj\n",
+          "7 0 obj\n<< /Type /StructElem /S /Link /P 6 0 R /Pg 3 0 R " +
+            "/K 0 >>\nendobj\n",
+          "8 0 obj\n<< /Nums [0 [7 0 R] 1 7 0 R] >>\nendobj\n",
+        ];
+
+        let loadingTask = getDocument({ data: assemblePdf(objects) });
+        let pdfDoc = await loadingTask.promise;
+        let page = await pdfDoc.getPage(1);
+        expect((await page.getStructTree()).children[0].role).toEqual("Link");
+
+        const data = await pdfDoc.extractPages([
+          { document: null, excludePages: [1] },
+        ]);
+        await loadingTask.destroy();
+
+        loadingTask = getDocument({ data });
+        pdfDoc = await loadingTask.promise;
+        expect(pdfDoc.numPages).toEqual(1);
+        page = await pdfDoc.getPage(1);
+        expect((await page.getAnnotations()).length).toEqual(0);
+        const tree = await page.getStructTree();
+        expect(tree.children.length).toEqual(1);
+        expect(tree.children[0].role).toEqual("Span");
+        expect(tree.children[0].children[0].type).toEqual("content");
+        await loadingTask.destroy();
+      });
+
       it("extract pages and merge struct trees", async function () {
         let loadingTask = getDocument(
           buildGetDocumentParams("two_paragraphs.pdf")