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 com.dom_distiller.client; | 5 package com.dom_distiller.client; |
6 | 6 |
7 import com.google.gwt.dom.client.Element; | 7 import com.google.gwt.dom.client.Element; |
8 import com.google.gwt.dom.client.Node; | 8 import com.google.gwt.dom.client.Node; |
9 | 9 |
10 import java.util.ArrayList; | 10 import java.util.ArrayList; |
(...skipping 37 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
48 * Finds the relevant elements rooted at |root|. | 48 * Finds the relevant elements rooted at |root|. |
49 */ | 49 */ |
50 private void find(final Node root) { | 50 private void find(final Node root) { |
51 DomVisitor domVisitor = new DomVisitor(); | 51 DomVisitor domVisitor = new DomVisitor(); |
52 new DomWalker(domVisitor).walk(root); | 52 new DomWalker(domVisitor).walk(root); |
53 } | 53 } |
54 | 54 |
55 private static final Set<String> sRelevantTags; | 55 private static final Set<String> sRelevantTags; |
56 static { | 56 static { |
57 sRelevantTags = new HashSet<String>(); | 57 sRelevantTags = new HashSet<String>(); |
| 58 sRelevantTags.add("BR"); |
| 59 sRelevantTags.add("FIGURE"); |
58 sRelevantTags.add("IMG"); | 60 sRelevantTags.add("IMG"); |
59 sRelevantTags.add("TABLE"); | 61 sRelevantTags.add("TABLE"); |
| 62 sRelevantTags.add("VIDEO"); |
60 } | 63 } |
61 | 64 |
62 /** | 65 /** |
63 * This class traverses the root element pre-orderly and determines if a vis
ible element is | 66 * This class traverses the root element pre-orderly and determines if a vis
ible element is |
64 * relevant with respect to Boilerpipe's content nodes. A relevant element
is then possibly | 67 * relevant with respect to Boilerpipe's content nodes. A relevant element
is then possibly |
65 * extracted by ElementVisitor. The final output is a combined ordered list
of the content | 68 * extracted by ElementVisitor. The final output is a combined ordered list
of the content |
66 * nodes and extracted elements. | 69 * nodes and extracted elements. |
67 */ | 70 */ |
68 private class DomVisitor implements DomWalker.Visitor { | 71 private class DomVisitor implements DomWalker.Visitor { |
69 private boolean inContent; | 72 private boolean inContent; |
(...skipping 114 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
184 case Node.DOCUMENT_NODE: | 187 case Node.DOCUMENT_NODE: |
185 default: | 188 default: |
186 return false; // Don't recurse into comments or sub-documen
ts. | 189 return false; // Don't recurse into comments or sub-documen
ts. |
187 } | 190 } |
188 } | 191 } |
189 | 192 |
190 @Override | 193 @Override |
191 public void exit(Node n) {} | 194 public void exit(Node n) {} |
192 } // ElementVisitor | 195 } // ElementVisitor |
193 } | 196 } |
OLD | NEW |