Merge pull request #13221 from Snuffleupagus/structTree-fixes

[api-minor] A couple of smaller `PDFPageProxy.getStructTree` fixes (PR 13171 follow-up)

src/core/document.js

@@ -454,6 +454,13 @@ class Page {
     const structTreeRoot = await this.pdfManager.ensureCatalog(
       "structTreeRoot"
     );
+    return this.pdfManager.ensure(this, "_parseStructTree", [structTreeRoot]);
+  }
+
+  /**
+   * @private
+   */
+  _parseStructTree(structTreeRoot) {
     const tree = new StructTreePage(structTreeRoot, this.pageDict);
     tree.parse();
     return tree;

src/core/struct_tree.js

@@ -328,6 +328,10 @@ class StructTreePage {
       }
       nodeToSerializable(child, root);
     }
+
+    if (root.children.length === 0) {
+      return null;
+    }
     return root;
   }
 }

src/display/api.js

@@ -1522,7 +1522,8 @@ class PDFPageProxy {
 
   /**
    * @returns {Promise<StructTreeNode>} A promise that is resolved with a
-   *   {@link StructTreeNode} object that represents the page's structure tree.
+   *   {@link StructTreeNode} object that represents the page's structure tree,
+   *   or `null` when no structure tree is present for the current page.
    */
   getStructTree() {
     return (this._structTreePromise ||= this._transport.getStructTree(

test/unit/api_spec.js

@@ -1702,6 +1702,69 @@ describe("api", function () {
         .catch(done.fail);
     });
 
+    it("gets empty structure tree", async function () {
+      const tree = await page.getStructTree();
+
+      expect(tree).toEqual(null);
+    });
+    it("gets simple structure tree", async function () {
+      const loadingTask = getDocument(
+        buildGetDocumentParams("structure_simple.pdf")
+      );
+      const pdfDoc = await loadingTask.promise;
+      const pdfPage = await pdfDoc.getPage(1);
+      const tree = await pdfPage.getStructTree();
+
+      expect(tree).toEqual({
+        role: "Root",
+        children: [
+          {
+            role: "Document",
+            children: [
+              {
+                role: "H1",
+                children: [
+                  {
+                    role: "NonStruct",
+                    children: [{ type: "content", id: "page2R_mcid0" }],
+                  },
+                ],
+              },
+              {
+                role: "P",
+                children: [
+                  {
+                    role: "NonStruct",
+                    children: [{ type: "content", id: "page2R_mcid1" }],
+                  },
+                ],
+              },
+              {
+                role: "H2",
+                children: [
+                  {
+                    role: "NonStruct",
+                    children: [{ type: "content", id: "page2R_mcid2" }],
+                  },
+                ],
+              },
+              {
+                role: "P",
+                children: [
+                  {
+                    role: "NonStruct",
+                    children: [{ type: "content", id: "page2R_mcid3" }],
+                  },
+                ],
+              },
+            ],
+          },
+        ],
+      });
+
+      await loadingTask.destroy();
+    });
+
     it("gets operator list", function (done) {
       const promise = page.getOperatorList();
       promise

web/pdf_page_view.js

@@ -619,6 +619,9 @@ class PDFPageView {
         this.eventBus._off("textlayerrendered", this._onTextLayerRendered);
         this._onTextLayerRendered = null;
         this.pdfPage.getStructTree().then(tree => {
+          if (!tree) {
+            return;
+          }
           const treeDom = this.structTreeLayer.render(tree);
           treeDom.classList.add("structTree");
           this.canvas.appendChild(treeDom);