-
Jérôme Euzenat authored
- added reference to the tutorial
Jérôme Euzenat authored- added reference to the tutorial
index.html 27.76 KiB
<html><head>
<title>A small tutorial on the Alignment API</title>
<script language="JavaScript">
function show(id) {
var element = document.getElementById(id);
element.style.display = "block";
}
function hide(id) {
var element = document.getElementById(id);
element.style.display = "none";
}
</script>
<link rel="stylesheet" type="text/css" href="../base.css" />
<link rel="stylesheet" type="text/css" href="../style.css" />
</head><body bgcolor="white">
<h1>A small tutorial on the Alignment API</h1>
<dl>
<dt>This version:</dt>
<dd>
http://alignapi.gforge.inria.fr/tutorial/
</dd>
<dt>Author:</dt>
<dd>
<a href="http://exmo.inrialpes.fr/people/euzenat">Jérôme Euzenat</a>,
INRIA Rhône-Alpes<br />
</dd>
</dl>
<hr />
<p>Here is a small tutorial for the alignment API. Since the API has no
dedicated GUI, most of the tutorial is based on command-lines
invocations. Of course, it is not the natural way to use this API:
it is made for being embedded in some application programme and we
are working towards implementing an alignment server that can help
programmes to use the API remotely. The complete tutorial is also
available as a self-contained <a href="script.sh">script.sh</a> or <a href="script.bat">script.bat</a>.</p>
<hr />
<h2>Preparation</h2>
<p>For running the alignment API, you must have a Java interpreter
available. We wil call it <tt>java</tt>.</p>
<p>Download the last version of the Alignement API from
<a href="http://gforge.inria.fr/frs/?group_id=117">http://gforge.inria.fr/frs/?group_id=117</a>.
Unzip it and go to the created directory:
<div class="fragment"><pre>
$ mkdir alignapi
$ cd alignapi
$ unzip align*.zip
</pre></div></p>
<p>You can check that everything works by only typing
<div class="fragment"><pre>
$ java -jar lib/procalign.jar --help
</pre></div>
<div class="button"><form><input type="button" value="Show
output" onclick="show('qu3')"/><input type="button" value="Hide
output" onclick="hide('qu3')"/></form></div>
<div class="explain" id="qu3">
<pre>
usage: Procalign [options] URI1 URI2
options are:
--impl=className -i classname Use the given alignment implementation.
--renderer=className -r className Specifies the alignment renderer
--output=filename -o filename Output the alignment in filename --params=filename -p filename Reads parameters from filename
--alignment=filename -a filename Start from an XML alignment file
--threshold=double -t double Filters the similarities under threshold
--cutmethod=hard|perc|prop|best|span -T hard|perc|prop|best|span method for computing the threshold
--debug[=n] -d [n] Report debug info at level n
-Dparam=value Set parameter
--help -h Print this message
</pre></div></p>
The above command outputs the command line usage of the Procalign class.
We do not detail it here, this tutorial will present
it entirelly.</p>
<p>If you want to modify the Alignment API and its implementation,
please refer to <a href="../align.html">this page</a>. In this
tutorial, we will simply learn how to use it.</p>
<p>You will then go to the tutorial directory by doing
<div class="fragment"><pre>
$ cd html/tutorial
</pre></div>
You can clean up previous trials by:
<div class="fragment"><pre>
$ rm results/
</pre></div>
</p>
<p>The goal of this tutorial is only to help you realize the
possibilities of the Alignment API and implementation. It can be
played by invoking each command line from the command-line
interpreter. In this example we use the <tt>tcsh</tt> syntax but the main
specific syntax is the first one:
<div class="fragment"><pre>
$ setenv CWD `pwd`
</pre></div>
which puts in variable <tt>$CWD</tt> the name of the current directory.</p>
<p>Beside a Java interpreter, if one wants to generate the HTML
translations of the alignements, this can be done with the help of
an XSLT 1.0 processor like <tt>xsltproc</tt>. Hence:
<div class="fragment"><pre>
$ xsltproc ../form-align.xsl results/file.rdf > results/file.html
</pre></div>
generates <tt>results/file.html</tt> from the alignment file <tt>results/file.rdf</tt>.</p>
<h2>The data</h2>
<p>Your mission, if you accept it, will be to find the best alignment
between two bibliographic ontologies. They can be seen here:
<dl>
<dt>edu.mit.visus.bibtex.owl</dt><dd>is a relatively faithfull
transcription of BibTeX as an ontology. It can be seen here
in <a href="edu.mit.visus.bibtex.owl">RDF/XML</a> or
<a href="edu.mit.visus.bibtex.html">HTML</a>.</dd>
<dt>myOnto.owl</dt><dd>is an extension of the previous one that
contains a number of supplementary concepts. It can be seen here
in <a href="myOnto.owl">RDF/XML</a> or
<a href="myOnto.html">HTML</a>.</dd>
</dl>
These two ontologies have been used for a few year in the <a href="oaei.ontologymatching.org">Ontology
Alignment Evaluation Initiative</a>
</p>
<h2>Matching</h2>
<p>
For demonstrating the use of our implementation of the Alignment API,
we implemented a particular processor
(<tt>fr.inrialpes.exmo.align.util.Procalign</tt>) which:
<ul compact="1">
<li>Reads two OWL/RDF ontologies;</li>
<li>Creates an alignment object;</li><li>Computes the alignment between these ontologies;</li>
<li>Displays the result.</li>
</ul>
Let's try to match these two ontologies:
<div class="fragment"><pre>
$ java -jar ../../lib/procalign.jar file://localhost$CWD/myOnto.owl file://localhost$CWD/edu.mit.visus.bibtex.owl
</pre></div>
Additionaly a number of options are available:
<ul compact="1">
<li>displaying debug information (-d);</li>
<li>controling the way of rendering the output (-r);</li>
<li>deciding the implementation of the alignment method (-i);</li>
<li>providing an input alignent (-a) [implemented but not used by most
methods].</li>
</ul>
</p>
<p>The result is displayed on the standard output. Since the output is
too long we send it to a file by using the <tt>-o</tt> flag:
<div class="fragment"><pre>
$ java -jar ../../lib/procalign.jar file://localhost$CWD/myOnto.owl file://localhost$CWD/edu.mit.visus.bibtex.owl -o results/equal.rdf
</pre></div>
See the output in <a href="results/equal.rdf">RDF/XML</a> or <a href="results/equal.html">HTML</a>.</p>
<p>The result is expressed in the Alignment format.
This format, in RDF/XML, is made of a
header containing "metadata" about the alignment:
<div class="owl"><pre>
<?xml version='1.0' encoding='utf-8' standalone='no'?>
<rdf:RDF xmlns='http://knowledgeweb.semanticweb.org/heterogeneity/alignment'
xml:base='http://knowledgeweb.semanticweb.org/heterogeneity/alignment'
xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'
xmlns:xsd='http://www.w3.org/2001/XMLSchema#'>
<Alignment>
<xml>yes</xml>
<level>0</level>
<type>**</type>
<time>66</time>
<owl:Class rdf:about="Techreport">
<owl:equivalentClass rdf:resource="Techreport"/>
</owl:Class>
<method>fr.inrialpes.exmo.align.impl.method.StringDistAlignment</method>
<onto1>file://localhost/JAVA/alignapi/html/tutorial/myOnto.owl</onto1>
<onto2>file://localhost/JAVA/alignapi/html/tutorial/edu.mit.visus.bibtex.owl</onto2>
<uri1>http://alignapi.gforge.inria.fr/tutorial/myOnto.owl</uri1>
<uri2>http://alignapi.gforge.inria.fr/tutorial/edu.mit.visus.bibtex.owl</uri2>
</pre></div>
and the corresponding set of correspondences:
<div class="owl"><pre>
<map>
<Cell>
<entity1 rdf:resource='http://alignapi.gforge.inria.fr/tutorial/myOnto.owl#Article'/>
<entity2 rdf:resource='http://alignapi.gforge.inria.fr/tutorial/edu.mit.visus.bibtex.owl#Article'/>
<measure rdf:datatype='http://www.w3.org/2001/XMLSchema#float'>1.0</measure>
<relation>=</relation>
</Cell>
</map>
</pre></div>
each correspondence is made of two references to the aligned entities,
the relation holding between the entities (<tt>=</tt>) and a confidence measure
(<tt>1.0</tt>) in this correspondence. Here, because the default method that
has been used for aligning the ontologies is so simple (it only
compares the labels of the entities and find that there is a
correspondence if their labels are equal), the correspondences are
always that simple. But it is too simple so we will use a more
sophisticated method based on an edit distance:</p>
<div class="fragment"><pre>
$ java -jar ../../lib/procalign.jar -i fr.inrialpes.exmo.align.impl.method.StringDistAlignment -DstringFunction=levenshteinDistance file://localhost$CWD/myOnto.owl file://localhost$CWD/edu.mit.visus.bibtex.owl -o results/levenshtein.rdf</pre></div>
See the output in <a href="results/levenshtein.rdf">RDF/XML</a> or <a href="results/levenshtein.html">HTML</a>.</p>
<p>This is achieved by specifying the class of Alignment to be used
(through the <tt>-i</tt> switch) and the distance function to be used
(<tt>-DstringFunction=levenshteinDistance</tt>).</p>
<p>Look at the results: how are they different from before?
<div class="button"><form><input type="button" value="Show
Discussion" onclick="show('qu1')"/><input type="button" value="Hide
Discussion" onclick="hide('qu1')"/></form></div>
<div class="explain" id="qu1"><p>We can see that the
correspondences now contain confidence factors different than <tt>1.0</tt>,
they also match strings which are not the same and indeed far more
correspondences are available.
</p></div></p>
<p>We do the same with another measure (smoaDistance):
<div class="fragment"><pre>
$ java -jar ../../lib/procalign.jar -i fr.inrialpes.exmo.align.impl.method.StringDistAlignment -DstringFunction=smoaDistance file://localhost$CWD/myOnto.owl file://localhost$CWD/edu.mit.visus.bibtex.owl -o results/SMOA.rdf
</pre></div>
</p>
<p><div class="logic"><p><b>More work:</b> you can apply other available alignments
classes. Look in the
<a
href="../../src/fr/inrialpes/exmo/align/impl/method>"../../src/fr/inrialpes/exmo/align/impl/method</a>
directory for more simple alignment methods. Also look in the
<tt>StringDistances</tt> class the possible values
for <tt>stringFunction</tt> (they are the names of methods).</p></div></p>
<p><div class="logic"><p><b>Advanced:</b> You can also look at the instructions for
installing WordNet and its Java interface and use a WordNet based
distance provided with the API implementation by:
<div class="fragment"><pre>
$ java -jar ../../lib/alignwn.jar -i fr.inrialpes.exmo.align.ling.JWNLAlignment file://localhost$CWD/myOnto.owl file://localhost$CWD/edu.mit.visus.bibtex.owl -o results/jwnl.rdf
</pre></div>
See the output in <a href="jwnl.rdf">RDF/XML</a> or <a href="jwnl.html">HTML</a>.</p></div></p>
<h2>Manipulating</h2>
<p>As can be seen there are some correspondences that do not really make
sense. Fortunatelly, they also have very low confidence level. It is
thus interesting to use a threshold for eliminating these
values. Let's try a threshold of <tt>.33</tt> over the alignment
(with the <tt>-t</tt> switch):
<div class="fragment"><pre>
$ java -jar ../../lib/procalign.jar file://localhost$CWD/myOnto.owl file://localhost$CWD/edu.mit.visus.bibtex.owl -i fr.inrialpes.exmo.align.impl.method.StringDistAlignment -DstringFunction=levenshteinDistance -t 0.33 -o results/levenshtein33.rdf
</pre></div>
See the output in <a href="results/levenshtein33.rdf">RDF/XML</a>
or <a href="results/levenshtein33.html">HTML</a>.</p>
<p>As expected we have suppressed some of these inaccurate
correspondences. But did we also suppressed accurate ones?
<div class="button"><form><input type="button" value="Show
Discussion" onclick="show('qu4')"/><input type="button" value="Hide
Discussion" onclick="hide('qu4')"/></form></div>
<div class="explain" id="qu4"><p>
This operation has contributed eliminating a number of innacurate
correspondences like Journal-Conference or
Composite-Conference. However, there remains some unaccurate
correspondences like Institution-InCollection and Published-UnPublished!
</p></div></p>
<p>We can also apply this treatment to other methods available:
<div class="fragment"><pre>
$ java -jar ../../lib/procalign.jar -i fr.inrialpes.exmo.align.impl.method.StringDistAlignment -DstringFunction=smoaDistance file://localhost$CWD/myOnto.owl file://localhost$CWD/edu.mit.visus.bibtex.owl -t 0.5 -o results/SMOA5.rdf
</pre></div>
See the output in <a href="results/SMOA5.rdf">RDF/XML</a>
or <a href="results/SMOA5.html">HTML</a>.</p>
<p><b>Other manipulations:</b> It is possible to invert an alignement
with the following command:
<div class="fragment"><pre>
$ java -cp ../../lib/procalign.jar fr.inrialpes.exmo.align.util.ParserPrinter -i results/SMOA5.rdf -o results/AOMS5.rdf
</pre></div>
See the output in <a href="results/AOMS5.rdf">RDF/XML</a>
or <a href="results/AOMS5.html">HTML</a>. The results is an alignement from
the source to the target. Inverting alignment is only the exchange of
the order of the elements in the alignment file. This can be useful
when you have an alignment of <i>A</i> to <i>B</i>, an alignment
from <i>C</i> to <i>B</i> and you want to go from <i>A</i>
to <i>C</i>. The solution is then to invert the second alignment and
to compose them.</p>
<p><div class="logic"><p><b>More work:</b>There is another switch (<tt>-T</tt>) in Procalign that
specifies the way a threshold is applied (hard|perc|prop|best|span)
the default being "hard". The curious reader can apply these and see
the difference in results. How they work is explained in the
Alignment API documentation.
</p></div></p>
<p><div class="logic"><p><b>More work:</b> Try to play with the
thresholds in order to find the best values for
levenshteinDistance and smoaDistance.
</p></div></p>
<h2>Output</h2>
<p>Once a good alignment has been found, only half of the work has
been done. In order to actually use our result it is necessary to
transform it into some processable format. For instance, if one
wants to merge two OWL ontologies, the alignment can be changed into
as set of OWL "bridging" axioms. This is achieved by "rendering" the
alignment in OWL (through the <tt>-r</tt> switch):
<div class="fragment"><pre>
$ java -cp ../../lib/procalign.jar fr.inrialpes.exmo.align.util.ParserPrinter results/SMOA5.rdf -r fr.inrialpes.exmo.align.impl.renderer.OWLAxiomsRendererVisitor
</pre></div>
The result is a set of OWL assertions of the form:
<div class="owl"><pre>
<owl:Class rdf:about="http://alignapi.gforge.inria.fr/tutorial/myOnto.owl#Techreport">
<owl:equivalentClass rdf:resource="http://alignapi.gforge.inria.fr/tutorial/edu.mit.visus.bibtex.owl#Techreport"/>
</owl:Class>
<owl:ObjectProperty rdf:about="http://alignapi.gforge.inria.fr/tutorial/myOnto.owl#copyright">
<owl:equivalentProperty rdf:resource="http://alignapi.gforge.inria.fr/tutorial/edu.mit.visus.bibtex.owl#hasCopyright"/>
</owl:ObjectProperty>
</pre></div>
</p>
<p>If one wants to use the alignments only for infering on instances
without actually merging the classes, she can generate SWRL rules:
<div class="fragment"><pre>
$ java -cp ../../lib/procalign.jar fr.inrialpes.exmo.align.util.ParserPrinter results/SMOA5.rdf -r fr.inrialpes.exmo.align.impl.renderer.SWRLRendererVisitor
</pre></div>
which brings for the same assertions:
<div class="owl"><pre>
<ruleml:imp>
<ruleml:_body>
<swrl:classAtom>
<owllx:Class owllx:name="http://alignapi.gforge.inria.fr/tutorial/myOnto.owl#Techreport"/>
<ruleml:var>x</ruleml:var>
</swrl:classAtom>
</ruleml:_body>
<ruleml:_head>
<swrlx:classAtom>
<owllx:Class owllx:name="http://alignapi.gforge.inria.fr/tutorial/edu.mit.visus.bibtex.owl#Techreport"/>
<ruleml:var>x</ruleml:var>
</swrl:classAtom>
</ruleml:_head> </ruleml:imp>
<ruleml:imp>
<ruleml:_body>
<swrl:individualPropertyAtom swrlx:property="http://alignapi.gforge.inria.fr/tutorial/myOnto.owl#copyright"/>
<ruleml:var>x</ruleml:var>
<ruleml:var>y</ruleml:var>
</swrl:individualPropertyAtom>
</ruleml:_body>
<ruleml:_head>
<swrl:datavaluedPropertyAtom swrlx:property="http://alignapi.gforge.inria.fr/tutorial/edu.mit.visus.bibtex.owl#hasCopyright"/>
<ruleml:var>x</ruleml:var>
<ruleml:var>y</ruleml:var>
</swrl:datavaluedPropertyAtom>
</ruleml:_head>
</ruleml:imp>
</pre></div>
</p>
<p>Exchanging data can also be achieved more simply
through XLST transformations which will transform the OWL instance
files from one ontology to another:
<div class="fragment"><pre>
$ java -cp ../../lib/procalign.jar fr.inrialpes.exmo.align.util.ParserPrinter results/SMOA5.rdf -r fr.inrialpes.exmo.align.impl.renderer.XSLTRendererVisitor -o results/SMOA5.xsl
</pre></div>
this transformation can be applied to the data of <a href="data.xml">data.xml</a>:
<div class="fragment"><pre>
$ xsltproc results/SMOA5.xsl data.xml > results/data.xml
</pre></div>
for giving the <a href="results/data.xml">results/data.xml</a> file.
</p>
<h2>Evaluating</h2>
<p>We will evaluate alignments by comparing them to some reference
alignment which is supposed to express what is expected from an
alignment of these two ontologies. The reference alignment is
<a href="refalign.rdf">refalign.rdf</a>
(or <a href="results/refalign.html">HTML</a>).</p>
<p>
<p>For evaluating we use another class than <tt>Procalign</tt>. It is called
<tt>EvalAlign</tt> we should specify this to <tt>java</tt>. By default, it
computes precision, recall and associated measures. It can be
invoked this way:
<div class="fragment"><pre>
$ java -cp ../../lib/procalign.jar fr.inrialpes.exmo.align.util.EvalAlign -i fr.inrialpes.exmo.align.impl.eval.PRecEvaluator file://localhost$CWD/refalign.rdf file://localhost$CWD/results/equal.rdf
</pre></div>
The first argument is always the reference alignment, the second one
is the alignment to be evaluated. The result is given here:
<div class="owl"><pre>
<?xml version='1.0' encoding='utf-8' standalone='yes'?>
<rdf:RDF xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'
xmlns:map='http://www.atl.external.lmco.com/projects/ontology/ResultsOntology.n3#'>
<map:output rdf:about=''>
<map:reference rdf:resource="file://localhost/JAVA/alignapi/html/tutorial/results/refalign.rdf">
<map:input rdf:resource="file://localhost/JAVA/alignapi/html/tutorial/results/equal.rdf">
<map:precision>1.0</map:precision>
<map:recall>0.22916666666666666</map:recall>
<fallout>0.0</fallout>
<map:fMeasure>0.37288135593220334</map:fMeasure>
<map:oMeasure>0.22916666666666666</map:oMeasure>
<result>0.22916666666666666</result>
</map:output>
</rdf:RDF>
</pre></div>
Of course, since that method only match objects with the same name, it
is accurate, yielding a high precision. However, it has poor recall.
</p>
<p>We can now evaluate the edit distance. What to expect from the evaluation of this alignment?
<div class="button"><form><input type="button" value="Show
Discussion" onclick="show('qu5')"/><input type="button" value="Hide
Discussion" onclick="hide('qu5')"/></form></div>
<div class="explain" id="qu5"><p>Since it returns more correspondences
by loosening the constraints for being a correspondence, it is
expected that the recall will increase at the expense of precision.
</p></div></p>
<p>We can see the results of:
<div class="fragment"><pre>
$ java -cp ../../lib/procalign.jar fr.inrialpes.exmo.align.util.EvalAlign -i fr.inrialpes.exmo.align.impl.eval.PRecEvaluator file://localhost$CWD/refalign.rdf file://localhost$CWD/results/levenshtein33.rdf
</pre></div>
<div class="button"><form><input type="button" value="Show
result" onclick="show('qu6')"/><input type="button" value="Hide
result" onclick="hide('qu6')"/></form></div>
<div class="explain" id="qu6"><pre>
<?xml version='1.0' encoding='utf-8' standalone='yes'?>
<rdf:RDF xmlns:rdf='http://www.w3.org/1999/02/22-rdf-syntax-ns#'
xmlns:map='http://www.atl.external.lmco.com/projects/ontology/ResultsOntology.n3#'>
<map:output rdf:about=''>
<map:reference rdf:resource="file://localhost/JAVA/alignapi/html/tutorial/results/refalign.rdf">
<map:input rdf:resource="file://localhost/JAVA/alignapi/html/tutorial/results/levenshtein33.rdf">
<map:precision>0.6811594202898551</map:precision>
<map:recall>0.9791666666666666</map:recall>
<fallout>0.3188405797101449</fallout>
<map:fMeasure>0.8034188034188035</map:fMeasure>
<map:oMeasure>0.5208333333333333</map:oMeasure>
<result>1.4374999999999998</result>
</map:output>
</rdf:RDF>
</pre></div></p>
<p>It is possible to summarize these results by comparing them to
each others. This can be achieved by the <tt>GroupEval</tt>
class. This class can output several formats (by default html) and
takes all the alignments in the subdirectories of the current
directory. Here we only have the <tt>results</tt> directory:
<div class="fragment"><pre>
$ cp refalign.rdf results
$ java -cp ../../lib/procalign.jar fr.inrialpes.exmo.align.util.GroupEval -r refalign.rdf -l "refalign,equal,SMOA,SMOA5,levenshtein,levenshtein33" -c prf -o results/eval.html
</pre></div>
The results are displayed in
the <a href="results/eval.html">results/eval.html</a> file whose main
content is the table:<br />
<center>
<table border='2' frame='sides' rules='groups'>
<colgroup align='center' />
<colgroup align='center' span='2' />
<colgroup align='center' span='2' />
<colgroup align='center' span='2' />
<colgroup align='center' span='2' />
<colgroup align='center' span='2' />
<colgroup align='center' span='2' />
<thead valign='top'><tr><th>algo</th>
<th colspan='2'>refalign</th>
<th colspan='2'>equal</th>
<th colspan='2'>SMOA</th>
<th colspan='2'>SMOA5</th>
<th colspan='2'>levenshtein</th>
<th colspan='2'>levenshtein33</th>
</tr></thead><tbody><tr><td>test</td>
<td>Prec.</td>
<td>Rec.</td>
<td>Prec.</td>
<td>Rec.</td>
<td>Prec.</td>
<td>Rec.</td>
<td>Prec.</td>
<td>Rec.</td><td>Prec.</td>
<td>Rec.</td>
<td>Prec.</td>
<td>Rec.</td>
</tr></tbody><tbody>
<tr>
<td>results</td>
<td>1.00</td>
<td>1.00</td>
<td>1.00</td>
<td>0.23</td>
<td>0.56</td>
<td>0.98</td>
<td>0.69</td>
<td>0.96</td>
<td>0.53</td>
<td>1.00</td>
<td>0.68</td>
<td>0.98</td>
</tr>
<tr bgcolor="yellow"><td>H-mean</td><td>1.00</td>
<td>1.00</td>
<td>1.00</td>
<td>0.23</td>
<td>0.56</td>
<td>0.98</td>
<td>0.69</td>
<td>0.96</td>
<td>0.53</td>
<td>1.00</td>
<td>0.68</td>
<td>0.98</td>
</tr>
</tbody></table>
</center>
</p>
<!--div class="fragment"><pre>
$ java -jar ../../lib/Procalign.jar file://localhost$CWD/rdf/edu.umbc.ebiquity.publication.owl file://localhost$CWD/rdf/edu.mit.visus.bibtex.owl -i fr.inrialpes.exmo.align.impl.method.StringDistAlignment -DstringFunction=levenshteinDistance -DprintMatrix=1 -o /dev/null > matrix.tex
</pre></div-->
<p><div class="logic"><p><b>More work:</b> As you can see,
the <tt>PRecEvaluator</tt> does not only provide precision and
recall but also provides F-measure. F-measure is usually used as
an "absolute" trade-off between precision and recall (i.e., the
optimum F-measure is considered the best precision and
recall). Can you establish this point for SMOA and levenshtein and
tell which algorithm is more adapted?
</p></div></p>
<!--div class="logic"><p><b>More work:</b>If you want to compare several algorithms, there
is another class, GroupAlign, that allows to run an
</p></div-->
<h2>Embedding</h2>
<p>Of course, the goal of this API is not to be used at the command
line level (even if it can be very useful). So if you are
ready for it, you can develop in Java your own application that takes
advantage of the API.</p>
<p>A skeleton of program using the Alignment API
is <a href="Skeleton.java">Skeleton.java</a>. It can be compiled by
invoking:
<div class="fragment"><pre>
$ javac -classpath ../../lib/api.jar:../../lib/rdfparser.jar:../../lib/align.jar:../../lib/procalign.jar -d results Skeleton.java
</div>
and run by:
<div class="fragment"><pre>
$ java -cp ../../lib/Procalign.jar:results Skeleton myOnto.owl edu.mit.visus.bibtex.owl
</pre></div></p>
<p>Now considering the API (that can be consulted through its thin
<a href="../../javadoc/org/semanticweb/owl/align/Alignment.html">Javadoc</a>
for instance), can you modify the Skeleton program in order for it
performs the following:
<ul compact="1">
<li>Run two different alignment methods (e.g., ngram distance and
smoa);</li>
<li>Merge the two results;</li>
<li>Trim at various thresholds;</li>
<li>Evaluate them against the reference alignment and choose the one with the
best F-Measure;</li>
<li>Displays it as SWRL Rules.</li>
</ul>
Of course, you can do it progressively.
<div class="fragment"><pre>
$ javac -classpath ../../lib/api.jar:../../lib/rdfparser.jar:../../lib/align.jar:../../lib/procalign.jar -d results MyApp.java
$ java -cp ../../lib/Procalign.jar:results MyApp myOnto.owl edu.mit.visus.bibtex.owl > results/MyApp.owl
</pre></div></p>
<p>Do you want to see a possible solution?
<div class="button"><form>
<input type="button" value="Cheat" onclick="show('qu7')"/>
<input type="button" value="Teacher is comming" onclick="hide('qu7')"/>
</form></div>
<div class="explain" id="qu7"><p>
The main piece of code in Skeleton.java is replaced by
<pre>
// Run two different alignment methods (e.g., ngram distance and smoa)
AlignmentProcess a1 = new StringDistAlignment( onto1, onto2 );
params.setParameter("stringFunction","smoaDistance");
a1.align( (Alignment)null, params );
AlignmentProcess a2 = new StringDistAlignment( onto1, onto2 );
params = new BasicParameters();
params.setParameter("stringFunction","ngramDistance");
a1.align( (Alignment)null, params );
// Merge the two results.
((BasicAlignment)a1).ingest(a2);
// Threshold at various thresholds
// Evaluate them against the references
// and choose the one with the best F-Measure
AlignmentParser aparser = new AlignmentParser(0);
Alignment reference = aparser.parse( "file://localhost"+(new File ( "refalign.rdf" ) . getAbsolutePath()), loaded );
Evaluator evaluator = new PRecEvaluator( reference, a1 );
double best = 0.;
Alignment result = null;
for ( int i = 0; i <= 10 ; i = i+2 ){
a1.cut( ((double)i)/10 );
evaluator.eval( new BasicParameters() );
System.err.println("Threshold "+(((double)i)/10)+" : "+((PRecEvaluator)evaluator).getFmeasure());
if ( ((PRecEvaluator)evaluator).getFmeasure() > best ) {
result = (BasicAlignment)((BasicAlignment)a1).clone();
best = ((PRecEvaluator)evaluator).getFmeasure();
}
}
// Displays it as SWRL Rules
PrintWriter writer = new PrintWriter (
new BufferedWriter(
new OutputStreamWriter( System.out, "UTF-8" )), true);
AlignmentVisitor renderer = new SWRLRendererVisitor(writer);
result.render(renderer);
writer.flush();
writer.close();
</pre>
</p></div></p>
</p>
<p>A full working solution is <a href="MyApp.java">MyApp.java</a>.</p><p><div class="logic"><p><b>More work:</b> Can you add a switch like the <tt>-i</tt> switch
of <tt>Procalign</tt> so that the main class of the application can
be passed at commant-line.
</p></div></p>
<p><div class="logic"><p><b>Advanced:</b>
You can develop a specialized matching algorithm
by subclassing the Java programs provided in the Alignment API
implementation (like BasicAlignment or DistanceAlignment).
</p></div></p>
<p><div class="logic"><p><b>Advanced:</b>
What about writing an editor for the alignment API?
</p></div></p>
<h2>Further exercises</h2>
<p>
More info: <a href="http://alignapi.gforge.inria.fr">http://alignapi.gforge.inria.fr</a>
</p>
<!--
Planning:
- Alignment server (incl. DB storage, agents, WSDL service)
- Extensive composition operators (with comp. tables)
- Expressive alignment language (with SEKT/François Sharffe)
-->
<h2>Acknowledgements</h2>
<p>The format of this tutorial has been shamelessly borrowed from Sean
Bechhofer's <a href="http://owl.man.ac.uk/2005/07/sssw/">OWL tutorial</a>.</p>
<hr />
<center><small>http://alignapi.gforge.inria.fr/tutorial</small></center>
<hr />
$Id$
</body>
</html>