Support <picture> in image extraction

java/org/chromium/distiller/DomUtil.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 org.chromium.distiller;
 
 import com.google.gwt.core.client.JsArray;
 import com.google.gwt.core.client.JsArrayString;
 import com.google.gwt.dom.client.AnchorElement;
 import com.google.gwt.dom.client.Document;
@@ skipping 21 matching lines @@
     public static native JsArray<Node> getAttributes(Element elem) /*-{
         return elem.attributes;
     }-*/;
 
     // Returns the first element with |className| in the tree rooted at |root|, null if none is
     // found.
     public static native Element getFirstElementWithClassName(Element root, String className) /*-{
         return root.querySelector("." + className);
     }-*/;
 
+    // Returns the first element with |tagName| in the tree rooted at |root|, including root.
+    // null if none is found.
+    public static Element getFirstElementByTagNameInc(Element e, String tagName) {
+        if (e.getTagName() == tagName) {
+            return e;
+        }
+        return getFirstElementByTagName(e, tagName);
+    }
+
+    // Returns the first element with |tagName| in the tree rooted at |root|.
+    // null if none is found.
+    public static Element getFirstElementByTagName(Element e, String tagName) {
+        NodeList<Element> elements = e.getElementsByTagName(tagName);
+        if (elements.getLength() > 0) {
+            return elements.getItem(0);
+        }
+        return null;
+    }
+
     public static native boolean hasClassName(Element elem, String className) /*-{
         return elem.classList.contains(className);
     }-*/;
 
     public static native JsArrayString getClassList(Element elem) /*-{
         return elem.classList;
     }-*/;
 
     /**
      * Check to see if a provided URL has the specified root domain (ex. http://a.b.c/foo/bar has
@@ skipping 214 matching lines @@
         stripTargetAttributes(clonedSubtree);
         stripFontColorAttributes(clonedSubtree);
         stripTableBackgroundColorAttributes(clonedSubtree);
         stripStyleAttributes(clonedSubtree);
         stripImageElements(clonedSubtree);
 
         return (Element) clonedSubtree;
     }
 
     /**
-     * Makes all anchors and video posters absolute. This calls "makeAllSrcAttributesAbsolute".
+     * Makes all anchors and video posters absolute. This calls "makeAllSrcAttributesAbsolute"
+     * and "makeAllSrcSetAbsolute".
      * @param rootNode The root Node to look through.
      */
     public static void makeAllLinksAbsolute(Node rootNode) {
         Element root = Element.as(rootNode);
 
         // AnchorElement.getHref() and ImageElement.getSrc() both return the
         // absolute URI, so simply set them as the respective attributes.
 
         if ("A".equals(root.getTagName())) {
             AnchorElement link = AnchorElement.as(root);
@@ skipping 16 matching lines @@
         }
         NodeList<Element> videoTags = root.getElementsByTagName("VIDEO");
         for (int i = 0; i < videoTags.getLength(); i++) {
             VideoElement video = (VideoElement) videoTags.getItem(i);
             if (!video.getPoster().isEmpty()) {
                 video.setPoster(video.getPoster());
             }
         }
         makeAllSrcAttributesAbsolute(root);
 
-        makeSrcSetAbsolute(root);
+        makeAllSrcSetAbsolute(root);
     }
 
-    private static void makeSrcSetAbsolute(Element root) {
-        if (root.getTagName().equals("IMG")) {
-            makeSrcSetAbsolute(ImageElement.as(root));
+    public static void makeAllSrcSetAbsolute(Element root) {
+        if (root.hasAttribute("srcset")) {
+            makeSrcSetAbsolute(root);
         }
-        NodeList<Element> imgs = DomUtil.querySelectorAll(root, "IMG[SRCSET]");
-        for (int i = 0; i < imgs.getLength(); i++) {
-            makeSrcSetAbsolute(ImageElement.as(imgs.getItem(i)));
+        NodeList<Element> es = DomUtil.querySelectorAll(root, "[SRCSET]");
+        for (int i = 0; i < es.getLength(); i++) {
+            makeSrcSetAbsolute(es.getItem(i));
         }
     }
 
-    public static void makeSrcSetAbsolute(ImageElement ie) {
+    private static void makeSrcSetAbsolute(Element ie) {
         String srcset = ie.getAttribute("srcset");
         if (srcset.isEmpty()) {
             ie.removeAttribute("srcset");
             return;
         }
 
-        String oldsrc = ie.getSrc();
+        ImageElement holder = Document.get().createImageElement();
         String[] sizes = StringUtil.jsSplit(srcset, ",");
         for(int i = 0; i < sizes.length; i++) {
             String size = StringUtil.jsTrim(sizes[i]);
             if (size.isEmpty()) continue;
             String[] comp = size.split(" ");
-            ie.setSrc(comp[0]);
-            comp[0] = ie.getSrc();
+            holder.setSrc(comp[0]);
+            comp[0] = holder.getSrc();
             sizes[i] = StringUtil.join(comp, " ");
         }
         ie.setAttribute("srcset", StringUtil.join(sizes, ", "));
-        ie.setSrc(oldsrc);
     }
 
-    public static List<String> getSrcSetUrls(ImageElement ie) {
+    public static List<String> getAllSrcSetUrls(Element root) {
         List<String> list = new ArrayList<>();
-        String srcset = ie.getAttribute("srcset");
+        if (root.hasAttribute("srcset")) {
+            list.addAll(getSrcSetUrls(root));
+        }
+        NodeList<Element> es = DomUtil.querySelectorAll(root, "[SRCSET]");
+        for (int i = 0; i < es.getLength(); i++) {
+            list.addAll(getSrcSetUrls(es.getItem(i)));
+        }
+        return list;
+    }
+
+    public static List<String> getSrcSetUrls(Element e) {
+        List<String> list = new ArrayList<>();
+        String srcset = e.getAttribute("srcset");
         if (srcset.isEmpty()) {
             return list;
         }
 
         String[] sizes = StringUtil.jsSplit(srcset, ",");
         for(int i = 0; i < sizes.length; i++) {
             String size = StringUtil.jsTrim(sizes[i]);
             if (size.isEmpty()) continue;
             String[] comp = size.split(" ");
             list.add(comp[0]);
@@ skipping 203 matching lines @@
     }-*/;
 
     public static native Document createHTMLDocument(Document doc) /*-{
         return doc.implementation.createHTMLDocument();
     }-*/;
 
     public static native Element getFirstElementChild(Document document) /*-{
         return document.firstElementChild;
     }-*/;
 }