Discard hidden articles when using fast path

java/org/chromium/distiller/ContentExtractor.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 org.chromium.distiller.document.TextDocument;
 import org.chromium.distiller.document.TextDocumentStatistics;
 import org.chromium.distiller.extractors.ArticleExtractor;
 import org.chromium.distiller.proto.DomDistillerProtos.StatisticsInfo;
@@ skipping 144 matching lines @@
 
     /**
      * Get a list of the content image URLs in the provided document.
      * @return A list of image URLs.
      */
     public List<String> getImageUrls() {
         return imageUrls;
     }
 
     /**
-     * Get the element of the main article, if any.
-     * @return An element of article (not necessarily the html5 article element).
-     */
-    private Element getArticleElement(Element root) {
-        NodeList<Element> allArticles = root.getElementsByTagName("ARTICLE");
-        // Having multiple article elements usually indicates a bad case for this shortcut.
-        // TODO(wychen): some sites exclude things like title and author in article element.
-        if (allArticles.getLength() == 1) {
-            return allArticles.getItem(0);
-        }
-        // Note that the CSS property matching is case sensitive, and "Article" is the correct
-        // capitalization.
-        String query = "[itemscope][itemtype*=\"Article\"],[itemscope][itemtype*=\"Post\"]";
-        allArticles = DomUtil.querySelectorAll(root, query);
-        // It is commonly seen that the article is wrapped separately or in multiple layers.
-        if (allArticles.getLength() > 0) {
-            return Element.as(DomUtil.getNearestCommonAncestor(allArticles));
-        }
-        return null;
-    }
-
-    /**
      * Converts the original HTML page into a WebDocument for analysis.
      */
     private WebDocumentInfo createWebDocumentInfoFromPage() {
         WebDocumentInfo info = new WebDocumentInfo();
         WebDocumentBuilder documentBuilder = new WebDocumentBuilder();
         DomConverter converter = new DomConverter(documentBuilder);
-        Element walkerRoot = getArticleElement(documentElement);
+        Element walkerRoot = DomUtil.getArticleElement(documentElement);
         if (walkerRoot == null) {
             walkerRoot = documentElement;
         }
         new DomWalker(converter).walk(walkerRoot);
         info.document = documentBuilder.toWebDocument();
         ensureTitleInitialized();
         info.hiddenElements = converter.getHiddenElements();
 
         return info;
     }
 
     /**
      * Implements the actual analysis of the page content, identifying the core elements of the
      * page.
      *
      * @param document the WebDocument representation of the page extracted from the DOM.
      */
     private void processDocument(WebDocument document) {
         TextDocument textDocument = document.createTextDocumentView();
         ArticleExtractor.INSTANCE.process(textDocument, candidateTitles);
         mStatisticsInfo.setWordCount(TextDocumentStatistics.countWordsInContent(textDocument));
         textDocument.applyToModel();
     }
 }