001 /*
002 * Copyright (c) 2006 Henri Sivonen
003 *
004 * Permission is hereby granted, free of charge, to any person obtaining a
005 * copy of this software and associated documentation files (the "Software"),
006 * to deal in the Software without restriction, including without limitation
007 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
008 * and/or sell copies of the Software, and to permit persons to whom the
009 * Software is furnished to do so, subject to the following conditions:
010 *
011 * The above copyright notice and this permission notice shall be included in
012 * all copies or substantial portions of the Software.
013 *
014 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
015 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
016 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
017 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
018 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
019 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
020 * DEALINGS IN THE SOFTWARE.
021 */
022
023 package fi.iki.hsivonen.xml;
024
025 import java.util.Arrays;
026
027 import org.xml.sax.Attributes;
028 import org.xml.sax.SAXException;
029
030 import gnu.xml.pipeline.EventConsumer;
031 import gnu.xml.pipeline.EventFilter;
032
033 public class BlockPassFilter extends EventFilter {
034 /**
035 * The XHTML namespace URI
036 */
037 private final static String XHTML_NS = "http://www.w3.org/1999/xhtml";
038
039 private final static String[] BLOCK_ELEMENTS = { "address", "blockquote",
040 "div", "dl", "h1", "h2", "h3", "h4", "h5", "h6", "ol", "p",
041 "table", "ul" };
042
043 private int depth;
044
045 public BlockPassFilter(EventConsumer next) {
046 super(next);
047 setContentHandler(this);
048 }
049
050 private static boolean isBlock(String name) {
051 return Arrays.binarySearch(BLOCK_ELEMENTS, name) > -1;
052 }
053
054 /**
055 * @see gnu.xml.pipeline.EventFilter#characters(char[], int, int)
056 */
057 public void characters(char[] arg0, int arg1, int arg2) throws SAXException {
058 if (depth > 0) {
059 super.characters(arg0, arg1, arg2);
060 }
061 }
062
063 /**
064 * @see gnu.xml.pipeline.EventFilter#endElement(java.lang.String,
065 * java.lang.String, java.lang.String)
066 */
067 public void endElement(String arg0, String arg1, String arg2)
068 throws SAXException {
069 if (depth > 0) {
070 super.endElement(arg0, arg1, arg2);
071 }
072 depth--;
073 }
074
075 /**
076 * @see gnu.xml.pipeline.EventFilter#ignorableWhitespace(char[], int, int)
077 */
078 public void ignorableWhitespace(char[] arg0, int arg1, int arg2)
079 throws SAXException {
080 if (depth > 0) {
081 super.ignorableWhitespace(arg0, arg1, arg2);
082 }
083 }
084
085 /**
086 * @see gnu.xml.pipeline.EventFilter#processingInstruction(java.lang.String,
087 * java.lang.String)
088 */
089 public void processingInstruction(String arg0, String arg1)
090 throws SAXException {
091 if (depth > 0) {
092 super.processingInstruction(arg0, arg1);
093 }
094 }
095
096 /**
097 * @see gnu.xml.pipeline.EventFilter#startDocument()
098 */
099 public void startDocument() throws SAXException {
100 depth = -1;
101 super.startDocument();
102 }
103
104 /**
105 * @see gnu.xml.pipeline.EventFilter#startElement(java.lang.String,
106 * java.lang.String, java.lang.String, org.xml.sax.Attributes)
107 */
108 public void startElement(String namespaceURI, String localName,
109 String qName, Attributes atts) throws SAXException {
110 if (depth > 0) {
111 super.startElement(namespaceURI, localName, qName, atts);
112 depth++;
113 } else if (XHTML_NS.equals(namespaceURI) && isBlock(localName)) {
114 super.startElement(namespaceURI, localName, qName, atts);
115 depth = 1;
116 }
117 }
118
119 }