Java And OpenDocument Files

[Arthur Buliva]

Hi Friends,

Now, lets say you have a file, lets call it document.odt

I have noticed that .odt files are not "stand alone" files like .txt or .dat files. Rather they are like zipped files with other defining files in there.

Before I get out of track, let me go to my question.

document.odt contains some text, say "Hello World". I want to edit this text through Java and save the resulting text. How can this be achieved?

Thanks and regards.

[Joe Ess]

OpenOffice has a Java API.

[Arthur Buliva]

And how can I use the API to open/edit files through a TextArea on my Java application?

[Ulf Dittmer]

Accessing ODF files through the OO Java API and using the data in GUI elements are separate activities. You'll need to dig into the API to figure out how to access the data that you need for the GUI. The AccessingFileFormats wiki page links to a number of articles about the OO Java API and the ODF file format.

[Arthur Buliva]

Dear Dittmer,

I am simply lost in the array of suggestions in those pages...

[Jesper de Jong]

The OpenOffice Java API

OpenOffice can read a number of file formats, and makes them accessible through its API. A starting point might be this article and of course the OO developer site

Some introductory information about the OO file format can be found here and here - basic Java code for reading OO files is here

Reading an OpenOffice file is not as simple as reading a plain text file, simply because OpenOffice contains a lot more features than a plain text editor.
[ October 31, 2007: Message edited by: Jesper Young ]

[Arthur Buliva]

So whats your suggestion on how to open them?

[Ulf Dittmer]

Either use the OO Java API to convert the file to some other format that you do know how to open, or -based on the articles linked above- write code that opens and processes the files.

From my cursory look at the articles OO files appear to be zipped-up XML files. Since Java has APIs for dealing with ZIP and XML files, getting at the actual contents shouldn't be too hard. Making sense of those is a different matter, of course - I'd recommend to start with a simple document, to see if you can manage to extract whatever information you need.

[Arthur Buliva]

I guess this leads to the next question, my good people.

How do I use the Java API to zip/unzip file packages?

[Ulf Dittmer]

Examples of using the ZIP API -as well as just about all other java.* classes- can be found at the Developer's Almanac: http://www.exampledepot.com/egs/java.util.zip/pkg.html

[Arthur Buliva]

public class Zipper
{
public static void main(String[] args)
{
try
 {
        // Open the ZIP file
        ZipFile zf = new ZipFile("filename.odt");
    
        // Enumerate each entry
        for (Enumeration entries = zf.entries(); entries.hasMoreElements() :wink: 
{
            // Get the entry name
            String zipEntryName = ((ZipEntry)entries.nextElement()).getName();
System.out.println(zipEntryName);
        }
    } 
catch (IOException e)
{ }
}
}

from the reference you gave me has solved my most immediate problem. Thanks!

[Arthur Buliva]

This is a sample output:

 
F:\>java Zipper
mimetype
Configurations2/statusbar/
Configurations2/accelerator/current.xml
Configurations2/floater/
Configurations2/popupmenu/
Configurations2/progressbar/
Configurations2/menubar/
Configurations2/toolbar/
Configurations2/images/Bitmaps/
content.xml
styles.xml
meta.xml
Thumbnails/thumbnail.png
settings.xml
META-INF/manifest.xml
 

and the code

import java.util.*;
import java.io.*;
import java.util.zip.*;
 

public class Unzipper
{
public static void main(String[] args)
{
try {
        // Open the ZIP file
        String inFilename = "filename.odt";
        ZipInputStream in = new ZipInputStream(new FileInputStream(inFilename));
    
        // Get the first entry
        ZipEntry entry = in.getNextEntry();
    
        // Open the output file
        String outFilename = "o";
        OutputStream out = new FileOutputStream(outFilename);
    
        // Transfer bytes from the ZIP file to the output file
        byte[] buf = new byte[1024];
        int len;
        while ((len = in.read(buf)) > 0) {
            out.write(buf, 0, len);
        }
    
        // Close the streams
        out.close();
        in.close();
    } catch (IOException e) {
    }
}
}

retrieves only the first entry whereas I would want the file called content.xml to be extracted, if not the entire odt file, onto a specified folder. How or what do I need to modify in the code?

[Joanne Neal]

You need a loop. ZipInputStream.getNextEntry() returns null when there are no more entries, so you can use that as the loop controller.
If you only want to output certain files, then you need to check the name of each ZipEntry object before you write it out. Check the API docs to see if there is a method that gets the name of the ZipEntry object.

[Arthur Buliva]

My redo of the code is

  
import java.util.*;
import java.io.*;
import java.util.zip.*;
 
 
public class Unzipper
{
    private String inFilename;
    private ZipInputStream zipInputStream;
    private OutputStream out;
    private ZipEntry entry;
    
    public Unzipper()
    {
        inFilename = "/home/arthur/filename.odt";
        
    }
    
    public void unzip()
    {
        try
        {
            // Open the ZIP file
            
            zipInputStream = new ZipInputStream(new FileInputStream(inFilename));
            
            entry = zipInputStream.getNextEntry();
            
            // Get the first entry
            while(entry != null)
            {
                System.out.println("Opening...");
                // Open the output file
                String outFilename = entry.getName();
                new File("/home/arthur/unzipped/" + outFilename).mkdirs();
                out = new FileOutputStream("/home/arthur/unzipped/" + outFilename);
                
                
                // Transfer bytes from the ZIP file to the output file
                byte[] buf = new byte[1024];
                int len;
                while ((len = zipInputStream.read(buf)) > 0)
                {
                    out.write(buf, 0, len);
                }
                out.flush();
                entry = zipInputStream.getNextEntry();
            }
            // Close the streams
            out.close();
            zipInputStream.close();
        }
        catch (IOException e)
        {
            e.printStackTrace();
        }
    }
 
static public void main(String[] args)
{
new Unzipper().unzip()
}
}
 
  

What could be the issue here as it returns lots of null and empty folders?

[Joanne Neal]

You create a new FileOutputStream object every time through the loop, but you only close the last one after you exit the loop. Put the out.close() call inside the loop.

[Arthur Buliva]

After doing that it renders the files as folders/directories instead of just files.

[Ulf Dittmer]

After doing that it renders the files as folders/directories instead of just files.

It looks as if the code is supposed to extract the files that are port of the ODF file, and write them to disk, each in the physical directory where it would logically be inside of the ODF file. Doesn't it do that? If not, what does it do, and where does it go wrong?

[Arthur Buliva]

Yes, this is what this code is supposed to do:

From the list of files inside the ODT file, which are

mimetype
Configurations2/statusbar/
Configurations2/accelerator/current.xml
Configurations2/floater/
Configurations2/popupmenu/
Configurations2/progressbar/
Configurations2/menubar/
Configurations2/toolbar/
Configurations2/images/Bitmaps/
content.xml
styles.xml
meta.xml
Thumbnails/thumbnail.png
settings.xml
META-INF/manifest.xml

I need to extract the files as they are in the odt file. For instance, thumbnail.png is to be extracted in a folder called Thumbnails.

So far, the code

import java.util.*;
import java.io.*;
import java.util.zip.*;
 
 
public class Unzipper
{
public static void main(String[] args)
{
try
{
String inFilename = "filename.odt";
        ZipInputStream in = new ZipInputStream(new FileInputStream(inFilename));
    
        ZipEntry entry = in.getNextEntry();
    OutputStream out = null;

while(entry != null)
{
        String outFilename = String.valueOf(entry.getName());
        out = new FileOutputStream(outFilename);
new File(outFilename).mkdirs();    
 
        byte[] buf = new byte[1024];
        int len;
        while ((len = in.read(buf)) > 0)
{
            out.write(buf, 0, len);
        }
entry = in.getNextEntry();
}
    
out.close();
        in.close();
    }
catch (IOException e)
{
System.out.print(e);
    }
}
}

returns

[root@localhost]# java Unzipper
java.io.FileNotFoundException: Configurations2\statusbar (The system cannot find
 the path specified)

Sorry if I am going back on my progress but I hope its for the best interest of clarity here.

Thanks.

[Ulf Dittmer]

There are a couple of problems with the code.

Firstly, you're creating the FileOutputStream (FOS) before creating the directories. That won't work, because the FOS constructor tries to access the file.

Secondly, if the name does not contain a "/", then it's a top-level file, and no directories should be created.

Thirdly, if the name ends with a "/", then it's an empty directory, and you should not open a FOS and try to copy bytes.

Lastly, if the name contains a "/" somewhere in the middle, then it's a file, and mkdirs should only be called with the part up to the last "/".

You can list the entry names your code should be expecting via "jar tf filename.odt"
[ November 17, 2007: Message edited by: Ulf Dittmer ]

[Arthur Buliva]

Eureka!

Thanks Dittmer

import java.util.*;
import java.io.*;
import java.util.zip.*;
 
 
public class Unzipper
{
public static void main(String[] args)
{
try
{
String inFilename = "filename.odt";
        ZipInputStream in = new ZipInputStream(new FileInputStream(inFilename));
    
        ZipEntry entry = in.getNextEntry();
    OutputStream out = null;

readfiles: while(entry != null)
{
        String outFilename = String.valueOf(entry.getName());
 

        if(!outFilename.contains("/"))
{
System.out.println(outFilename + " is a top level file. No folders will be created.\n");
out = new FileOutputStream(outFilename);
         byte[] buf = new byte[1024];
   int len;
        while ((len = in.read(buf)) > 0)
{
            out.write(buf, 0, len);
        }
entry = in.getNextEntry();
continue readfiles;
}
else
{
new File(outFilename.substring(0, outFilename.lastIndexOf("/"))).mkdirs();
if(outFilename.endsWith("/"))
{
System.out.println(outFilename + " is a folder\n");
entry = in.getNextEntry();
continue readfiles;
}
else
{
System.out.println(outFilename + " is a file within a folder\n");
new File(outFilename.substring(0, outFilename.lastIndexOf("/"))).mkdirs();
out = new FileOutputStream(outFilename);
byte[] buf = new byte[1024];
       int len;
    while ((len = in.read(buf)) > 0)
{
            out.write(buf, 0, len);
        }
entry = in.getNextEntry();
}
}
out.close();
}
        in.close();
    }
catch (IOException e)
{
System.out.print(e);
    }
finally
{
System.exit(0);
}
}
}

Has solved my problem

[Arthur Buliva]

   
 
import java.io.BufferedReader;
import java.io.File;
import java.io.FileReader;
import java.io.IOException;
 
public class XMLParser
{
    public String parseXMLFile(File xmlFile)
    {
       
        String fileContent = "";
        String contents = "";
        boolean startMark = true;
        boolean endMark = false;
        String possibleToEdit = "";
        
        try
        {
            File file = new File(xmlFile.getCanonicalPath());
            FileReader fis = new FileReader(file);
            BufferedReader buffer = new BufferedReader(fis);
            while(buffer.ready())
            {
                fileContent += buffer.readLine() + "\n";
            }
            // We now break down the contents of the file
            for (int i = 0; i < fileContent.length(); i++)
            {
                contents += String.valueOf(fileContent.charAt(i));
                if(String.valueOf(fileContent.charAt(i)).equalsIgnoreCase(">"))
                    contents += "\n";
            }
 
            for (int i = 0; i < contents.length(); i++)
            {
                if(String.valueOf(contents.charAt(i)).equalsIgnoreCase("<"))
                {
                    startMark = true;
                    endMark = false;
                }
                else if(String.valueOf(contents.charAt(i)).equalsIgnoreCase(">"))
                {
                    startMark = false;
                    endMark = true;
                }
                if(startMark)
                {
                    possibleToEdit = possibleToEdit;
                }
                if(endMark)
                {
                    possibleToEdit += String.valueOf(contents.charAt(i)).replaceAll(">", "\n");
                }
            }
            return possibleToEdit;
        }
        catch (Exception e)
        {
            e.printStackTrace();
            return String.valueOf(e);
        }
    }
}
 

Has crudely but successfully 'ripped' up the xml file. Now it is to 'reverse engineer' this class to save the changes.

[Ulf Dittmer]

I think you'd be better off with a real XML parser. Scanning an XML file like you showed above gets tricky in the presence of nested elements. Something like the following would do the trick (it doesn't extract the files, just reads them in-place, but it's easy to adapt to your situation).

import java.io.*;
import java.util.*;
import java.util.zip.*;
import javax.xml.parsers.*;
 
import org.xml.sax.Attributes;
import org.xml.sax.helpers.DefaultHandler;
 
public class SimpleODFReader {
 
    public static void main (String[] args) throws Exception {
        SimpleODFReader self = new SimpleODFReader();
        self.read(args);
    }
 
    private void read (String[] args) {
        for (int i=1; i<args.length; i++) {
            File f = new File(args[i]);
            if (! f.isFile()) {
                System.err.println(f.getName() + " does not seem to be a file");
                return;
            }
            if (! f.getName().endsWith(".odt")) {
                System.err.println(f.getName() + " does not seem to be an ODT file");
                return;
            }
 
            try {
                // Create a builder factory
                SAXParserFactory factory = SAXParserFactory.newInstance();
                // Create the builder 
                SAXParser parser = factory.newSAXParser();
 
                // now open the document file and parse its various constituent XML files
                ZipFile zf = new ZipFile(f);
 
                ZipEntry entry = zf.getEntry("content.xml");
                parser.parse(zf.getInputStream(entry), new ODFTextParser());
            } catch (Exception ex) {
                System.err.println("error reading file '"+f.getName()+"': "+ex.getMessage());
                ex.printStackTrace();
            }
        }
    }
 
    public class ODFTextParser extends DefaultHandler {
 
        public void startDocument() {
            System.out.println("startDocument");
        }
 
        public void endDocument() {
            System.out.println("endDocument");
        }
 
        public void startElement (String uri, String localName, String qName, Attributes atts) {
            System.out.println("start: "+qName);
        }
 
        public void characters (char[] ch, int start, int length) {
            System.out.println(new String(ch, start, length));
        }
 
        public void endElement (String uri, String localName, String qName) {
            System.out.println("end: "+qName);
        }
    }
 
}

[Arthur Buliva]

How is this class used?

This is what am getting:

E:\>javac  SimpleODFReader.java
 
E:\>java SimpleODFReader
 
E:\>java SimpleODFReader filename.odt
 
E:\>
    

Is there something am missing?

[Ulf Dittmer]

The for loop should run from 0, not 1. Silly typo.

[Arthur Buliva]

Great

 
E:\>java SimpleODFReader filename.odt
startDocument
start: office :mrgreen: ocument-content
start: office:scripts
end: office:scripts
start: office:font-face-decls
start: style:font-face
end: style:font-face
start: style:font-face
end: style:font-face
start: style:font-face
end: style:font-face
start: style:font-face
end: style:font-face
start: style:font-face
end: style:font-face
end: office:font-face-decls
start: office:automatic-styles
end: office:automatic-styles
start: office:body
start: office:text
start: office:forms
end: office:forms
start: text:sequence-decls
start: text:sequence-decl
end: text:sequence-decl
start: text:sequence-decl
end: text:sequence-decl
start: text:sequence-decl
end: text:sequence-decl
start: text:sequence-decl
end: text:sequence-decl
end: text:sequence-decls
start: text :-P	 
This is a test file
end: text :-P	 
end: office:text
end: office:body
end: office :mrgreen: ocument-content
endDocument
 
E:\>
 

is what I get.

Now, my main aim so far of going through the unzipping process is the editing part. Am building a simple online odt editor. So on saving, I was thinking of working with the content.xml file, editing whatever I want, then zipping up the entire package and renaming the resulting file to a dot odt. Thus I would have achieved the intended result.