|
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
21 public class TagRemover implements HtmlDOMVisitor{
22
23 private Set tagsToRemove = new HashSet ();
24
25 public TagRemover(){}
26
27
33 public void init(int siteId){
34 }
35
36 public void addTag(String tag){
37 if ( tag != null && !this.tagsToRemove.contains(tag) ){
38 this.tagsToRemove.add(tag);
39 }
40 }
41
42 public void addTags(String [] tags){
43 if ( tags != null ){
44 int size = tags.length;
45 for ( int i=0; i<size; i++ ){
46 this.addTag(tags[i]);
47 }
48 }
49 }
50
51
57 public Document parseDOM(Document doc){
58 if ( doc != null && this.tagsToRemove.size()>0 ){
59 removeMarkups(doc.getDocumentElement(),this.tagsToRemove);
60 }
61 return doc;
62 }
63
64
70 private void removeMarkups(Node node, Set markups){
71
72 if ( markups == null || node == null ){
73 return;
74 }
75
76 boolean toRemove = matchMarkup(node.getNodeName(),markups);
77 Node parentNode = node.getParentNode();
78
79 NodeList childs = node.getChildNodes();
80 ArrayList al = new ArrayList ();
81
82 int count = childs.getLength();
83 for ( int i=0 ; i<count ; i++ ){
84 Node child = (Node )childs.item(i);
85 al.add(child);
86 }
87 for ( int i=0; i<count ; i++ ){
88 Node child = (Node )al.get(i);
89 removeMarkups(child,markups);
90 }
91
92 if ( toRemove ){
93 parentNode.removeChild(node);
94 } else{
95 Node refNode = node;
96 while ( matchMarkup(parentNode.getNodeName(),markups) ){
97 refNode = parentNode;
98 parentNode = parentNode.getParentNode();
99 }
100
101 if ( !refNode.equals(node) ){
102 parentNode.insertBefore(node,refNode);
103 }
104 }
105 }
106
107 private boolean matchMarkup(String name, Set markups){
108 if ( name == null || markups == null || markups.size()==0 ){
109 return false;
110 }
111 Iterator iterator = markups.iterator();
112 String tag = null;
113 while ( iterator.hasNext() ){
114 tag = (String )iterator.next();
115 if ( tag.toLowerCase().equals(name.toLowerCase()) ){
116 return true;
117 }
118 }
119 return false;
120 }
121 }
122 |
Popular Tags |
Java API By Example, From Geeks To Geeks.