|
1 package org.jahia.services.htmlparser;
2
3 import java.util.ArrayList  ;
4 import java.util.HashSet ;
5 import java.util.Iterator ;
6 import java.util.Set ;
7
8 import org.w3c.dom.Document ;
9 import org.w3c.dom.Node ;
10 import org.w3c.dom.NodeList ;
11
12 import org.jahia.registries.ServicesRegistry;
13
14
23 public class RemoveUnwantedMarkupVisitor implements HtmlDOMVisitor{
24
25 private Set htmlMarkupSettings = new HashSet ();
26 private Set unwantedMarkups = new HashSet ();
27
28 public RemoveUnwantedMarkupVisitor(){}
29
30
36 public void init(int siteId){
37 this.htmlMarkupSettings = new HashSet ();
38
39 Iterator iterator = ServicesRegistry.getInstance()
40 .getHtmlParserService()
41 .getMarkupSettings(siteId);
42
43 while ( iterator.hasNext() ){
44 MarkupSetting setting = (MarkupSetting)iterator.next();
45 this.htmlMarkupSettings.add(setting);
46 if ( setting.getSettingType().equals(ToRemoveMarkupSetting.SETTING_TYPE) ){
47 MarkupDefinition markupDef =
49 ServicesRegistry.getInstance().getHtmlParserService()
50 .getMarkupDefinition(setting.getMarkupDefId());
51 if ( markupDef != null ){
52 this.unwantedMarkups.add(markupDef);
53 }
54 }
55 }
56 }
57
58
64 public Document parseDOM(Document doc){
65 if ( doc != null && this.unwantedMarkups.size()>0 ){
66 removeMarkups(doc.getDocumentElement(),this.unwantedMarkups);
67 }
68 return doc;
69 }
70
71
77 private void removeMarkups(Node node, Set markups){
78
79 if ( markups == null || node == null ){
80 return;
81 }
82
83 boolean toRemove = matchMarkup(node.getNodeName(),markups);
84 Node parentNode = node.getParentNode();
85
86 NodeList childs = node.getChildNodes();
87 ArrayList al = new ArrayList ();
88
89 int count = childs.getLength();
90 for ( int i=0 ; i<count ; i++ ){
91 Node child = (Node )childs.item(i);
92 al.add(child);
93 }
94 for ( int i=0; i<count ; i++ ){
95 Node child = (Node )al.get(i);
96 removeMarkups(child,markups);
97 }
98
99 if ( toRemove ){
100 parentNode.removeChild(node);
101 } else{
102 Node refNode = node;
103 while ( matchMarkup(parentNode.getNodeName(),markups) ){
104 refNode = parentNode;
105 parentNode = parentNode.getParentNode();
106 }
107
108 if ( !refNode.equals(node) ){
109 parentNode.insertBefore(node,refNode);
110 }
111 }
112 }
113
114 private boolean matchMarkup(String name, Set markups){
115 if ( name == null || markups == null || markups.size()==0 ){
116 return false;
117 }
118 Iterator iterator = markups.iterator();
119 while ( iterator.hasNext() ){
120 MarkupDefinition def = (MarkupDefinition)iterator.next();
121 if ( !def.isCaseSensitive() ){
122 if ( def.getName().toLowerCase().equals(name.toLowerCase()) ){
123 return true;
124 }
125 } else if ( def.getName().equals(name) ){
126 return true;
127 }
128 }
129 return false;
130 }
131 } |
Popular Tags |
Java API By Example, From Geeks To Geeks.