KickJava   Java API By Example, From Geeks To Geeks.

Java > Open Source Codes > org > apache > slide > extractor > MSPowerPointExtractor


1 /*
2  * $Header: /home/cvs/jakarta-slide/src/share/org/apache/slide/extractor/MSPowerPointExtractor.java,v 1.2.2.1 2004/09/29 15:01:26 unico Exp $
3  * $Revision: 1.2.2.1 $
4  * $Date: 2004/09/29 15:01:26 $
5  *
6  * ====================================================================
7  *
8  * Copyright 2004 The Apache Software Foundation
9  *
10  * Licensed under the Apache License, Version 2.0 (the "License");
11  * you may not use this file except in compliance with the License.
12  * You may obtain a copy of the License at
13  *
14  * http://www.apache.org/licenses/LICENSE-2.0
15  *
16  * Unless required by applicable law or agreed to in writing, software
17  * distributed under the License is distributed on an "AS IS" BASIS,
18  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
19  * See the License for the specific language governing permissions and
20  * limitations under the License.
21  *
22  */
23
24 package org.apache.slide.extractor;
25
26 import org.apache.poi.util.LittleEndian;
27 import org.apache.poi.poifs.eventfilesystem.POIFSReaderListener;
28 import org.apache.poi.poifs.eventfilesystem.POIFSReaderEvent;
29 import org.apache.poi.poifs.eventfilesystem.POIFSReader;
30 import org.apache.poi.poifs.filesystem.DocumentInputStream;
31
32 import java.io.*;
33
34 /**
35  * Author: Ryan Rhodes
36  * Date: Jun 27, 2004
37  * Time: 3:45:39 AM
38  */
39 public class MSPowerPointExtractor extends AbstractContentExtractor implements POIFSReaderListener
40 {
41     private ByteArrayOutputStream writer = new ByteArrayOutputStream();
42
43     public MSPowerPointExtractor(String JavaDoc uri, String JavaDoc contentType, String JavaDoc namespace) {
44         super(uri, contentType, namespace);
45     }
46
47     public Reader extract(InputStream content) throws ExtractorException {
48         try {
49             POIFSReader reader = new POIFSReader();
50             reader.registerListener(this);
51             reader.read(content);
52
53             return new InputStreamReader(new ByteArrayInputStream(writer.toByteArray()));
54         }
55         catch(Exception JavaDoc e) {
56                 throw new ExtractorException(e.getMessage());
57         }
58     }
59
60     public void processPOIFSReaderEvent(POIFSReaderEvent event)
61     {
62         try{
63             if(!event.getName().equalsIgnoreCase("PowerPoint Document"))
64                 return;
65
66             DocumentInputStream input = event.getStream();
67
68             byte[] buffer = new byte[input.available()];
69             input.read(buffer, 0, input.available());
70
71             for(int i=0; i<buffer.length-20; i++)
72             {
73                 long type = LittleEndian.getUShort(buffer,i+2);
74                 long size = LittleEndian.getUInt(buffer,i+4);
75
76                 if(type==4008)
77                 {
78                     writer.write(buffer, i + 4 + 1, (int) size +3);
79                     i = i + 4 + 1 + (int) size - 1;
80
81                 }
82             }
83         }
84         catch (Exception JavaDoc e)
85         {
86
87         }
88     }
89
90     public static void main(String JavaDoc[] args) throws Exception JavaDoc
91     {
92         FileInputStream in = new FileInputStream(args[0]);
93
94         MSPowerPointExtractor ex = new MSPowerPointExtractor(null, null, null);
95
96         Reader reader = ex.extract(in);
97
98         int c;
99         do
100         {
101             c = reader.read();
102
103             System.out.print((char)c);
104         }
105         while( c != -1 );
106     }
107 }
Popular Tags