Extract Lang attribute for marked contents

src/core/evaluator.js

@@ -2317,9 +2317,19 @@ class PartialEvaluator {
               return;
             }
             // Other marked content types aren't supported yet.
+            let props = null;
+            if (args[1] instanceof Dict) {
+              const lang = args[1].get("Lang");
+              if (typeof lang === "string") {
+                props = Object.create(null);
+                props.lang = stringToPDFString(lang);
+              }
+            }
+
             args = [
               args[0].name,
               args[1] instanceof Dict ? args[1].get("MCID") : null,
+              props,
             ];
 
             break;
@@ -3505,15 +3515,21 @@ class PartialEvaluator {
               markedContentData.level++;
 
               let mcid = null;
+              let itemLang = null;
               if (args[1] instanceof Dict) {
                 mcid = args[1].get("MCID");
+                const langString = args[1].get("Lang");
+                if (typeof langString === "string") {
+                  itemLang = stringToPDFString(langString);
+                }
               }
               textContent.items.push({
                 type: "beginMarkedContentProps",
                 id: Number.isInteger(mcid)
                   ? `${self.idFactory.getPageObjId()}_mc${mcid}`
                   : null,
                 tag: args[0] instanceof Name ? args[0].name : null,
+                lang: itemLang,
               });
             }
             break;

src/display/api.js

@@ -1166,6 +1166,8 @@ class PDFDocumentProxy {
  *   'beginMarkedContentProps', or 'endMarkedContent'.
  * @property {string} id - The marked content identifier. Only used for type
  *   'beginMarkedContentProps'.
+ * @property {string|null} tag - The marked content tag.
+ * @property {string|null} lang - The lang attribute for the marked content.
  */
 
 /**

src/display/text_layer.js

@@ -293,6 +293,9 @@ class TextLayer {
           if (item.id) {
             this.#container.setAttribute("id", `${item.id}`);
           }
+          if (item.lang) {
+            this.#container.setAttribute("lang", item.lang);
+          }
           parent.append(this.#container);
         } else if (item.type === "endMarkedContent") {
           this.#container = this.#container.parentNode;

test/pdfs/.gitignore

@@ -752,3 +752,4 @@
 !bug1937438_af_from_latex.pdf
 !bug1937438_from_word.pdf
 !bug1937438_mml_from_latex.pdf
+!marked_content_lang.pdf

test/unit/api_spec.js

@@ -4488,6 +4488,23 @@ Caron Broadcasting, Inc., an Ohio corporation (“Lessee”).`)
       await loadingTask.destroy();
     });
 
+    it("gets operatorList, with marked content lang", async function () {
+      const loadingTask = getDocument(
+        buildGetDocumentParams("marked_content_lang.pdf")
+      );
+      const pdfDoc = await loadingTask.promise;
+      const pdfPage = await pdfDoc.getPage(1);
+      const opList = await pdfPage.getOperatorList({
+        annotationMode: AnnotationMode.DISABLE,
+      });
+      expect(opList.fnArray[0]).toEqual(OPS.beginMarkedContentProps);
+      expect(opList.argsArray[0][0]).toEqual("P");
+      expect(opList.argsArray[0][2]?.lang).toEqual("en-US");
+      expect(opList.fnArray[10]).toEqual(OPS.beginMarkedContentProps);
+      expect(opList.argsArray[10][0]).toEqual("P");
+      expect(opList.argsArray[10][2]?.lang).toEqual("es-ES");
+    });
+
     it("gets operatorList, with page resources containing corrupt /CCITTFaxDecode data", async function () {
       const loadingTask = getDocument(
         buildGetDocumentParams("poppler-90-0-fuzzed.pdf")

test/unit/text_layer_spec.js

@@ -250,4 +250,29 @@ describe("textLayer", function () {
 
     await loadingTask.destroy();
   });
+
+  it("handles lang attribute for marked content", async function () {
+    if (isNodeJS) {
+      pending("document.createElement is not supported in Node.js.");
+    }
+    const loadingTask = getDocument(
+      buildGetDocumentParams("marked_content_lang.pdf")
+    );
+    const pdfDocument = await loadingTask.promise;
+    const page = await pdfDocument.getPage(1);
+
+    const container = document.createElement("div");
+    const textLayer = new TextLayer({
+      textContentSource: page.streamTextContent({
+        includeMarkedContent: true,
+      }),
+      container,
+      viewport: page.getViewport({ scale: 1 }),
+    });
+    await textLayer.render();
+
+    const span = container.querySelector("#p17R_mc1");
+    expect(span.getAttribute("lang")).toEqual("es-ES");
+    expect(span.textContent).toEqual("Esto es español");
+  });
 });