|
1
16 package org.apache.cocoon.components.search;
17
18 import org.apache.lucene.document.Document;
19 import org.apache.lucene.document.Field;
20 import org.xml.sax.Attributes  ;
21 import org.xml.sax.ContentHandler ;
22 import org.xml.sax.Locator ;
23 import org.xml.sax.helpers.AttributesImpl ;
24
25 import java.util.ArrayList ;
26 import java.util.HashSet ;
27 import java.util.Iterator ;
28 import java.util.List ;
29 import java.util.Stack ;
30
31
42 public class LuceneIndexContentHandler implements ContentHandler
43 {
44 public static final String LUCENE_URI = "http://apache.org/cocoon/lucene/1.0";
45
46
51 public static final String LUCENE_ATTR_TO_TEXT_ATTRIBUTE = "text-attr";
52
53 StringBuffer bodyText;
54 private List documents;
55 private Document bodyDocument;
56 private Stack elementStack;
57 private HashSet fieldTags;
58
59
62 public LuceneIndexContentHandler() {
63 this.bodyText = new StringBuffer ();
64 this.bodyDocument = new Document();
65 this.documents = new ArrayList ();
66 this.documents.add(this.bodyDocument);
67 this.elementStack = new Stack ();
68 this.fieldTags = new HashSet ();
69 }
70
71
76 public void setFieldTags(HashSet fieldTags) {
77 this.fieldTags = fieldTags;
78 }
79
80
85 public void setDocumentLocator(Locator locator) { }
86
87 public List allDocuments() {
88 return documents;
89 }
90
91 public Iterator iterator() {
92 return documents.iterator();
93 }
94
95 public void characters(char[] ch, int start, int length) {
96 if (ch.length > 0 && start >= 0 && length > 1) {
97 if (elementStack.size() > 0) {
98 IndexHelperField tos = (IndexHelperField) elementStack.peek();
99 tos.appendText(ch, start, length);
100 }
101 bodyText.append(' ');
102 bodyText.append(ch, start, length);
103 }
104 }
105
106 public void endDocument() {
107 bodyDocument.add(Field.UnStored(LuceneXMLIndexer.BODY_FIELD, bodyText.toString()));
108 }
109
110 public void endElement(String namespaceURI, String localName, String qName) {
111 IndexHelperField tos = (IndexHelperField) elementStack.pop();
112 String lname = tos.getLocalFieldName();
113 StringBuffer text = tos.getText();
114
115 Attributes atts = tos.getAttributes();
117 boolean attributesToText = atts.getIndex(LUCENE_URI, LUCENE_ATTR_TO_TEXT_ATTRIBUTE) != -1;
118 for (int i = 0; i < atts.getLength(); i++) {
119 if (LUCENE_URI.equals(atts.getURI(i))) continue;
120
121 String atts_lname = atts.getLocalName(i);
122 String atts_value = atts.getValue(i);
123 bodyDocument.add(Field.UnStored(lname + "@" + atts_lname, atts_value));
124 if (attributesToText) {
125 text.append(' ');
126 text.append(atts_value);
127 bodyText.append(' ');
128 bodyText.append(atts_value);
129 }
130 }
131
132 if (text != null && text.length() > 0) {
133 if (isFieldTag(lname)) {
134 bodyDocument.add(Field.UnIndexed(lname, text.toString()));
135 }
136 bodyDocument.add(Field.UnStored(lname, text.toString()));
137 }
138 }
139
140 public void endPrefixMapping(String prefix) { }
141
142 public void ignorableWhitespace(char[] ch, int start, int length) { }
143
144 public void processingInstruction(String target, String data) { }
145
146 public void skippedEntity(String name) { }
147
148 public void startDocument() { }
149
150 public void startElement(String namespaceURI, String localName, String qName, Attributes atts) {
151 IndexHelperField ihf = new IndexHelperField(localName, qName, new AttributesImpl (atts));
152 elementStack.push(ihf);
153 }
154
155 public void startPrefixMapping(String prefix, String uri) { }
156
157
163 private boolean isFieldTag(String tag) {
164 if (fieldTags == null) {
166 return false;
167 }
168 Iterator i = fieldTags.iterator();
169 while (i.hasNext()) {
170 if (tag.equals(i.next())) {
171 return true;
172 }
173 }
174 return false;
175 }
176 }
177 |
Popular Tags |
Java API By Example, From Geeks To Geeks.