Index: java/org/chromium/distiller/ContentExtractor.java |
diff --git a/java/org/chromium/distiller/ContentExtractor.java b/java/org/chromium/distiller/ContentExtractor.java |
index 4a8f8bd8dbee9acff98fbbc9084f7f5d8b4626fe..3a08dd44cac51952c2bcb68d9a5f0fa52a08ae73 100644 |
--- a/java/org/chromium/distiller/ContentExtractor.java |
+++ b/java/org/chromium/distiller/ContentExtractor.java |
@@ -167,23 +167,40 @@ public class ContentExtractor { |
*/ |
private Element getArticleElement(Element root) { |
NodeList<Element> allArticles = root.getElementsByTagName("ARTICLE"); |
+ List<Element> visibleElements = getVisibleElements(allArticles); |
// 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); |
+ if (visibleElements.size() == 1) { |
+ return visibleElements.get(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); |
+ visibleElements = getVisibleElements(allArticles); |
// It is commonly seen that the article is wrapped separately or in multiple layers. |
- if (allArticles.getLength() > 0) { |
- return Element.as(DomUtil.getNearestCommonAncestor(allArticles)); |
+ if (visibleElements.size() > 0) { |
+ return Element.as(DomUtil.getNearestCommonAncestor(visibleElements)); |
} |
return null; |
} |
/** |
+ * Get a list of visible elements. |
+ * @return A list of visible elements. |
+ */ |
+ private List<Element> getVisibleElements(NodeList<Element> nodeList) { |
+ List<Element> visibleElements = new ArrayList<>(); |
+ for (int i = 0; i < nodeList.getLength(); i ++) { |
mdjones
2015/10/21 18:15:02
nit: i++
|
+ Element element = nodeList.getItem(i); |
+ if (DomUtil.isVisible(element)) { |
+ visibleElements.add(element); |
+ } |
+ } |
+ return visibleElements; |
+ } |
+ |
+ /** |
* Converts the original HTML page into a WebDocument for analysis. |
*/ |
private WebDocumentInfo createWebDocumentInfoFromPage() { |