BeanShell - Revision history

Tredmond: /* Duplicated Axioms */

2011-06-15T19:59:11Z

‎Duplicated Axioms

Show changes

Tredmond: /* Missing Annotation Properties */

2011-06-15T19:52:26Z

‎Missing Annotation Properties

Tredmond: /* Duplicated Axioms */

2011-06-15T19:52:12Z

‎Duplicated Axioms

Tredmond: /* Duplicated Axioms */

2011-06-15T19:51:15Z

‎Duplicated Axioms

Tredmond: /* Counting With a Reasoner */

2011-06-15T19:50:51Z

‎Counting With a Reasoner

Tredmond: /* Counting With a Reasoner */

2011-06-15T17:21:27Z

‎Counting With a Reasoner

Show changes

Tredmond: /* Counting (with and without a reasoner) */

2011-06-15T17:14:28Z

‎Counting (with and without a reasoner)

Tredmond: /* The BeanShell */

2011-06-15T14:22:12Z

‎The BeanShell

Tredmond: /* Property Puns */

2011-06-15T14:21:03Z

‎Property Puns

Tredmond: /* The BeanShell */

2011-06-15T14:19:21Z

‎The BeanShell

← Older revision		Revision as of 19:52, June 15, 2011
Line 349:		Line 349:

	== Missing Annotation Properties ==		== Missing Annotation Properties ==
		+
		+	Coming soon...

← Older revision		Revision as of 19:52, June 15, 2011
Line 347:		Line 347:
	};		};
	</pre>		</pre>
		+
		+	== Missing Annotation Properties ==

@@ Line 310: / Line 310: @@
 == Duplicated Axioms ==
-At one point there was some discussion about how Protege should handle axioms that are duplicated in more than one location in the imports closure of a ontology.  Obviously this lead to a desire to find all the duplicated axioms.  The [https://obi.svn.sourceforge.net/svnroot/obi/trunk/src/ontology/branches/obi.owl obi ontology] is an interesting ontology in this regard - it has several duplicated axioms.  I have added some code to ignore the "external" ontologies because these don't really count for this analysis in the obi case.  For some reason this bean shell code only works when it is cut and pasted.  It does not work when it is run from a script file.  I don't know why yet.
+At one point there was some discussion about how Protege should handle axioms that are duplicated in more than one location in the imports closure of a ontology.  Obviously this led to a desire to find all the duplicated axioms.  The [https://obi.svn.sourceforge.net/svnroot/obi/trunk/src/ontology/branches/obi.owl obi ontology] is an interesting ontology in this regard - it has several duplicated axioms.  I have added some code to ignore the "external" ontologies because these don't really count for this analysis in the obi case.  For some reason this bean shell code only works when it is cut and pasted.  It does not work when it is run from a script file.  I don't know why yet.
 <pre>
 import java.util.TreeSet;

@@ Line 306: / Line 306: @@
 bsh % manager.applyChanges(changes);
 bsh %
 </pre>

← Older revision		Revision as of 17:14, June 15, 2011
Line 96:		Line 96:
	It would be better if the bean shell offered some command completion like eclipse and I will see what can be done about this (it might be tricky?).		It would be better if the bean shell offered some command completion like eclipse and I will see what can be done about this (it might be tricky?).

−	== Counting ~~(with~~ and ~~without~~ a ~~reasoner~~) ==	+	== Counting With a Reasoner ==
		+
		+	This is a problem that came off the list and it is a bit complicated for the bean shell. But it is a good illustration because it uses several things that might be useful. The user wanted to count the individuals in two classes, calculate the percentage of them that are in the second class and then attach the result to a data property assertion. I am not positive exactly what the data property assertion was supposed to look like but I made an example here. The ontology is as follows:
		+	<pre>
		+	Prefix(owl:=<http://www.w3.org/2002/07/owl#>)
		+	Prefix(:=<http://www.semanticweb.org/ontologies/2011/5/furniture#>)
		+	Prefix(Marouns_problem:=<http://www.semanticweb.org/ontologies/2011/5/Marouns_problem#>)
		+
		+	Ontology(<http://www.semanticweb.org/ontologies/2011/5/furniture>
		+
		+	EquivalentClasses(Marouns_problem:having_final_value_0 ObjectComplementOf(Marouns_problem:having_final_value_1))
		+	EquivalentClasses(Marouns_problem:having_final_value_1
		+	ObjectUnionOf(
		+	ObjectSomeValuesFrom(Marouns_problem:has_mechanism owl:Thing)
		+	ObjectSomeValuesFrom(Marouns_problem:has_protocol owl:Thing)
		+	ObjectSomeValuesFrom(Marouns_problem:has_service owl:Thing)))
		+
		+	ObjectPropertyRange(Marouns_problem:has_mechanism Marouns_problem:mechanism)
		+	ObjectPropertyRange(Marouns_problem:has_protocol Marouns_problem:protocol)
		+	ObjectPropertyRange(Marouns_problem:has_service Marouns_problem:service)
		+
		+	ClassAssertion(Marouns_problem:moderate Marouns_problem:a)
		+	ClassAssertion(ObjectExactCardinality(1 Marouns_problem:has_mechanism Marouns_problem:mechanism) Marouns_problem:a)
		+	ClassAssertion(ObjectExactCardinality(1 Marouns_problem:has_protocol Marouns_problem:protocol) Marouns_problem:a)
		+	ClassAssertion(ObjectExactCardinality(1 Marouns_problem:has_service Marouns_problem:service) Marouns_problem:a)
		+
		+	ClassAssertion(Marouns_problem:moderate :b)
		+	ClassAssertion(ObjectExactCardinality(1 Marouns_problem:has_mechanism Marouns_problem:mechanism) :b)
		+	ClassAssertion(ObjectExactCardinality(1 Marouns_problem:has_protocol Marouns_problem:protocol) :b)
		+
		+	ClassAssertion(Marouns_problem:moderate :c)
		+
		+	ClassAssertion(Marouns_problem:moderate :d)
		+	ClassAssertion(ObjectExactCardinality(0 Marouns_problem:has_mechanism) :d)
		+	ClassAssertion(ObjectExactCardinality(0 Marouns_problem:has_protocol) :d)
		+	ClassAssertion(ObjectExactCardinality(0 Marouns_problem:has_service) :d)
		+	)
		+	</pre>

@@ Line 3: / Line 3: @@
 __TOC__
-As a Protege developer, I have a unique ability to answer complex questions about large ontologies.  All I have to do is to start Protege inside eclipse, open the ontology and set a breakpoint and I will have access to a full array of powerful Java tools for analyzing the ontology.  These tools range from the full set of OWL api features, direct access to reasoners and the ability to make a complex script out of simple primitives.  I have often thought that this power would be much more useful in the hands of the end user but it has been unclear how to get there.  What is really needed is a very powerful scripting language.
+As a Protege developer, I have a unique ability to answer complex questions about large ontologies.  All I have to do is to start Protege inside eclipse, open the ontology and set a breakpoint and I will have access to a full array of powerful Java tools for analyzing the ontology.  These tools range from the full set of [http://owlapi.sourceforge.net OWL API] features, direct access to reasoners and the ability to make a complex script out of simple primitives.  I have often thought that this power would be much more useful in the hands of the end user but it has been unclear how to get there.  What is really needed is a very powerful scripting language.
-The beanshell (a CO-ODE plugin developed at the University of Manchester) can fill some of these needs.  The primary disadvantage of the bean shell is that it requires some experience with Java and knowledge of the OWL API.  This will severely limit the audience but I thought it would still be useful to include some pages on how to use the beanshell.
+The beanshell (a CO-ODE plugin developed at the University of Manchester) can fill some of these needs.  The primary disadvantage of the bean shell is that it requires some experience with Java and knowledge of the [http://owlapi.sourceforge.net OWL API].  This will severely limit the audience but I thought it would still be useful to include some pages on how to use the beanshell.
 The beanshell also requires some understanding of the Protege API but to start with all that is really needed is to know one call:
 <pre>

← Older revision		Revision as of 14:21, June 15, 2011
Line 95:		Line 95:
	</pre>		</pre>
	It would be better if the bean shell offered some command completion like eclipse and I will see what can be done about this (it might be tricky?).		It would be better if the bean shell offered some command completion like eclipse and I will see what can be done about this (it might be tricky?).
		+
		+	== Counting (with and without a reasoner) ==

← Older revision		Revision as of 14:19, June 15, 2011
Line 1:		Line 1:
	= The BeanShell =		= The BeanShell =
		+
		+	__TOC__

	As a Protege developer, I have a unique ability to answer complex questions about large ontologies. All I have to do is to start Protege inside eclipse, open the ontology and set a breakpoint and I will have access to a full array of powerful Java tools for analyzing the ontology. These tools range from the full set of OWL api features, direct access to reasoners and the ability to make a complex script out of simple primitives. I have often thought that this power would be much more useful in the hands of the end user but it has been unclear how to get there. What is really needed is a very powerful scripting language.		As a Protege developer, I have a unique ability to answer complex questions about large ontologies. All I have to do is to start Protege inside eclipse, open the ontology and set a breakpoint and I will have access to a full array of powerful Java tools for analyzing the ontology. These tools range from the full set of OWL api features, direct access to reasoners and the ability to make a complex script out of simple primitives. I have often thought that this power would be much more useful in the hands of the end user but it has been unclear how to get there. What is really needed is a very powerful scripting language.