lucene/sandbox/contributions/indyo/xdocs/tutorial.xml

222 lines
5.9 KiB
XML

<?xml version="1.0"?>
<document>
<properties>
<title>Indyo Tutorial</title>
<author email="kelvint@apache.org">Kelvin Tan</author>
</properties>
<body>
<section name="About this Tutorial">
<p>
This tutorial is intended to give first-time users an
introduction to using Indyo, a datasource-independent
Lucene indexing framework.
</p>
<p>
This will include how to obtain Indyo, configuring Indyo
and indexing a directory on a filesystem.
</p>
</section>
<section name="Step 1: Obtaining Indyo">
<p>
First, you need to obtain Indyo. As
of this writing, Indyo is only available via CVS, from the
"jakarta-lucene-sandbox" repository. See
<a href="http://jakarta.apache.org/site/cvsindex.html">Jakarta CVS</a>
on accessing files via CVS.</p>
</section>
<section name="Step 2: Building Indyo">
<p>
Get a copy of <a href="http://jakarta.apache.org/ant">Ant</a> if
you don't already have it installed. Then simply type "ant" in the
directory where the local copy of the Indyo sources reside.
</p>
<p>
Voila! You should now have a jar file "indyo-&lt;version number&gt;.jar".
</p>
</section>
<section name="Step 3: Configuring Indyo">
<p>
The "src/conf" folder contains a default configuration file which is
sufficient for normal use.
</p>
</section>
<section name="Step 4: Using Indyo">
<p>
Congratulations, you have finally reached the fun the
part of this tutorial. This is where you'll discover
the power of Indyo.
</p>
<p>
To index a datasource, first instantiate the respective
datasource, then hand it to IndyoIndexer for indexing.
For example:
</p>
<source><![CDATA[
IndexDataSource ds = new FSDataSource("/usr/local/lucene/docs");
IndyoIndexer indexer = new IndyoIndexer("/usr/local/index",
"/usr/local/indyo/default.config.xml");
indexer.index(ds);
]]></source>
<p>
FSDataSource is a simple datasource which indexes both files
and directories. The metadata FSDataSource adds to each document is:
filePath, fileName, fileSize, fileFormat, fileContents,
fileLastModifiedDate. Based on the file extension of the files indexed,
Indyo will use file content-handlers according to the mappings found in the
configuration file. If you're not happy with this list of file
metadata, feel free to subclass FSDataSource, or, as we're about
to cover next, write your own custom IndexDataSource.
</p>
<p>
Get familiar with FSDataSource. You'll find it very handy, both for indexing
files directly, as well as nesting it within another datasource. For example,
you might need to index a database table, in which one of the rows represent
the location of a file, and you may want to use FSDataSource to index this
file as well.
</p>
<subsection name="Writing your custom IndexDataSource">
<p>
To write a custom IndexDataSource, you need to write a class
which implements IndexDataSource, and provides an implementation
for the getData() method which returns a Map[]. The javadoc of the
getData() method reads:
</p>
<source><![CDATA[
/**
* Retrieve a array of Maps. Each map represents the
* a document to be indexed. The key:value pair of the map
* is the metadata of the document.
*/
]]></source>
<p>
So, the getData() method provides a way for Indyo to retrieve document
metadata from each IndexDataSource. A simple example of a custom
IndexDataSource, HashMapDataSource is provided below.
</p>
<source><![CDATA[
public class HashMapDataSource implements IndexDataSource
{
private Map data;
public HashMapDataSource(Map data)
{
this.data = data;
}
public Map[] getData() throws Exception
{
return new Map[1]{data};
}
}
]]></source>
<p>
As you can see, HashMapDataSource doesn't do anything very useful. It
always results in one Document being indexed, and the document's fields
depend on the contents of the map that HashMapDataSource was initialized
with.
</p>
<p>
A slightly more useful IndexDataSource, SingleDocumentFSDataSource
provides an example of how to nest datasources. Given a directory,
SingleDocumentFSDataSource recursively indexes all directories
and files within that directory <i>as the same Document</i>. In other
words, only one Document is created in the index. This is accomplished
by the use of a nested datasource. The code for
SingleDocumentFSDataSource is listed below:
</p>
<source><![CDATA[
public class SingleDocumentFSDataSource
implements IndexDataSource
{
private File file;
public SingleDocumentFSDataSource(File file)
{
this.file = file;
}
public Map[] getData() throws Exception
{
Map data = new HashMap(1);
data.put(NESTED_DATASOURCE, new FSDataSource(file));
return new Map[1]{data};
}
}
]]></source>
<p>
Nested datasources don't result in a separate Document being created.
Use them when working with complex datasources, i.e., datasources
which are an aggregation of multiple datasources. The current way to
add a nested datasource is using the key "NESTED_DATASOURCE". Indyo
accepts an IndexDataSource object, a List of IndexDataSources,
or an IndexDataSource[] for this key.
</p>
</subsection>
</section>
<section name="Where to Go From Here">
<p>
Congratulations! You have completed the Indyo
tutorial. Although this has only been an introduction
to Torque, it should be sufficient to get you started
with Indyo in your applications. For those of you
seeking additional information, there are several other
documents on this site that can provide details on
various subjects. Lastly, the source code is an
invaluable resource when all else fails to provide
answers!
</p>
</section>
<section name="Acknowledgements">
<p>
This document was shamelessly ripped from the extremely well-written
and well-organized
<a href="http://jakarta.apache.org/turbine/torque/tutorial.html">Torque
</a> tutorial. Thanks Pete!
</p>
</section>
</body>
</document>