gwt getInnerText -> javascript innerText or textContent

src/com/dom_distiller/client/IEReadingViewParser.java

 // Copyright 2014 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 package com.dom_distiller.client;
 
 import java.util.ArrayList;
 import java.util.EnumMap;
 import java.util.List;
 import java.util.Map;
@@ skipping 115 matching lines @@
             }
         }
     }
 
     private void findDate() {
         mDate = "";
 
         // Get date from any element that includes the "dateline" class.
         Element elem = DomUtil.getFirstElementWithClassName(mRoot, "dateline");
         if (elem != null) {
-            mDate = elem.getInnerText();
+            // Use javascript textContent (instead of javascript innerText) to include invisible
+            // text.
+            mDate = DomUtil.javascriptTextContent(elem);
         } else {  // Otherwise, get date from meta tag with "displaydate" as name.
             for (int i = 0; i < mAllMeta.getLength(); i++) {
                 MetaElement meta = MetaElement.as(mAllMeta.getItem(i));
                 if (meta.getName().equalsIgnoreCase("displaydate")) {
                     mDate = meta.getContent();
                     break;
                 }
             }
         }
     }
 
     private void findAuthor() {
         mAuthor = "";
 
         // Get author from the first element that includes the "byline-name" class.
         // Note that we ignore the order of this element for now.
         Element elem = DomUtil.getFirstElementWithClassName(mRoot, "byline-name");
-        if (elem != null) mAuthor = elem.getInnerText();
+        // Use javascript textContent (instead of javascript innerText) to include invisible text.
+        if (elem != null) mAuthor = DomUtil.javascriptTextContent(elem);
     }
 
     private void findPublisher() {
         mPublisher = "";
 
         // Look for "publisher" or "source_organization" attribute in any html tag.
         NodeList<Element> allElems = mRoot.getElementsByTagName("*");
         for (int i = 0; i < allElems.getLength() && mPublisher.isEmpty(); i++) {
             Element e = allElems.getItem(i);
             mPublisher = e.getAttribute("publisher");
@@ skipping 56 matching lines @@
         double aspectRatio = (double) width / (double) image.getHeight();
         return aspectRatio >= 1.3 && aspectRatio <= 3.0;
     }
 
     private static String getCaption(ImageElement image) {
         // If |image| is a child of <figure>, then get the <figcaption> elements.
         Element parent = image.getParentElement();
         if (!parent.hasTagName("FIGURE")) return "";
         NodeList<Element> captions = parent.getElementsByTagName("FIGCAPTION");
         int numCaptions = captions.getLength();
-        if (numCaptions > 0 && numCaptions <= 2)
-            return captions.getItem(0).getInnerText();  // Just use the first one.
-        return "";
+        String caption = "";
+        if (numCaptions > 0 && numCaptions <= 2) {
+            // Use javascript innerText (instead of javascript textContent) to get only visible
+            // captions.
+            for (int i = 0; i < numCaptions && caption.isEmpty(); i++) {
+                caption = DomUtil.getInnerText(captions.getItem(i));
+            }
+        }
+        return caption;
     }
 
     private static boolean isTextInBody(Element root, String text) {
         String lowerText = text.toLowerCase();
         NodeList<Element> bodies = root.getElementsByTagName("BODY");
         for (int i = 0; i < bodies.getLength(); i++) {
-            if (bodies.getItem(i).getInnerText().toLowerCase().contains(lowerText)) return true;
+            // Use javascript textContent (instead of javascript innerText) to include invisible
+            // text.
+            if (DomUtil.javascriptTextContent(
+                    bodies.getItem(i)).toLowerCase().contains(lowerText)) {
+                return true;
+            }
         }
         return false;
     }
 }