OLD | NEW |
1 // Copyright 2014 The Chromium Authors. All rights reserved. | 1 // Copyright 2014 The Chromium Authors. All rights reserved. |
2 // Use of this source code is governed by a BSD-style license that can be | 2 // Use of this source code is governed by a BSD-style license that can be |
3 // found in the LICENSE file. | 3 // found in the LICENSE file. |
4 | 4 |
5 package org.chromium.distiller; | 5 package org.chromium.distiller; |
6 | 6 |
7 import com.google.gwt.core.client.JsArray; | 7 import com.google.gwt.core.client.JsArray; |
8 import com.google.gwt.core.client.JsArrayString; | 8 import com.google.gwt.core.client.JsArrayString; |
9 import com.google.gwt.dom.client.AnchorElement; | 9 import com.google.gwt.dom.client.AnchorElement; |
10 import com.google.gwt.dom.client.Document; | 10 import com.google.gwt.dom.client.Document; |
(...skipping 109 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
120 public static Element getArticleElement(Element root) { | 120 public static Element getArticleElement(Element root) { |
121 NodeList<Element> allArticles = root.getElementsByTagName("ARTICLE"); | 121 NodeList<Element> allArticles = root.getElementsByTagName("ARTICLE"); |
122 List<Element> visibleElements = getVisibleElements(allArticles); | 122 List<Element> visibleElements = getVisibleElements(allArticles); |
123 // Having multiple article elements usually indicates a bad case for thi
s shortcut. | 123 // Having multiple article elements usually indicates a bad case for thi
s shortcut. |
124 // TODO(wychen): some sites exclude things like title and author in arti
cle element. | 124 // TODO(wychen): some sites exclude things like title and author in arti
cle element. |
125 if (visibleElements.size() == 1) { | 125 if (visibleElements.size() == 1) { |
126 return visibleElements.get(0); | 126 return visibleElements.get(0); |
127 } | 127 } |
128 // Note that the CSS property matching is case sensitive, and "Article"
is the correct | 128 // Note that the CSS property matching is case sensitive, and "Article"
is the correct |
129 // capitalization. | 129 // capitalization. |
130 String query = "[itemscope][itemtype*=\"Article\"],[itemscope][itemtype*
=\"Post\"]"; | 130 String query = "[itemscope][itemtype*=\"Article\"],[itemscope][itemtype*
=\"Posting\"]"; |
131 allArticles = DomUtil.querySelectorAll(root, query); | 131 allArticles = DomUtil.querySelectorAll(root, query); |
132 visibleElements = getVisibleElements(allArticles); | 132 visibleElements = getVisibleElements(allArticles); |
133 // It is commonly seen that the article is wrapped separately or in mult
iple layers. | 133 // It is commonly seen that the article is wrapped separately or in mult
iple layers. |
134 if (visibleElements.size() > 0) { | 134 if (visibleElements.size() > 0) { |
135 return Element.as(DomUtil.getNearestCommonAncestor(visibleElements))
; | 135 return Element.as(DomUtil.getNearestCommonAncestor(visibleElements))
; |
136 } | 136 } |
137 return null; | 137 return null; |
138 } | 138 } |
139 | 139 |
140 /** | 140 /** |
(...skipping 383 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
524 }-*/; | 524 }-*/; |
525 | 525 |
526 public static native Document createHTMLDocument(Document doc) /*-{ | 526 public static native Document createHTMLDocument(Document doc) /*-{ |
527 return doc.implementation.createHTMLDocument(); | 527 return doc.implementation.createHTMLDocument(); |
528 }-*/; | 528 }-*/; |
529 | 529 |
530 public static native Element getFirstElementChild(Document document) /*-{ | 530 public static native Element getFirstElementChild(Document document) /*-{ |
531 return document.firstElementChild; | 531 return document.firstElementChild; |
532 }-*/; | 532 }-*/; |
533 } | 533 } |
OLD | NEW |