Publishing Data Using Semantic Web Technologies

Publishing Data Using Semantic
Web Technologies
An introduction for software engineers

Nikolaos Konstantinou, Ph.D.
National Documentation Centre / N.H.R.F.

Lecture Outline
• Introduction to the Semantic Web
• Semantic Web Languages
• Publishing RDF Using OpenLink’s Virtuoso
• Linked Open Data and Examples

November 9, 2011 National Documentation Centre / NHRF 2

The Problem (1)
• Keyword-based queries cannot be expressive
• E.g. search for:
– Cities in the U.S. with more than 100,000
inhabitants
– Italian painters of the 18th century
• Web resources
– Do not (usually) convey their meaning


The Problem (2)
• Seeking specific information in the Web or a
repository
• Integrating distributed data sources
• Need for data annotation
– Necessary for data non-readable by human
• E.g. binary information, multimedia
– Annotation may be redundant, incomplete, or
erroneous
– When it is present it does not necessarily follow a
standard pattern

The Semantic Web Paradigm (1)
• ‘Web of Data’ as in a ‘Web of Documents’
– Web resources uniquely identified by their URI
• Assign an unambiguously defined meaning to
information, its semantics
– Ontology, a well defined vocabulary
– Queries can be posed by any third parties
• Knowledge modeled in the form of a graph
– subject, predicate, object
• Interconnected data sets on the Web
– Provide context Documentation Centre / NHRF
November 9, 2011 National 5

The Semantic Web Paradigm (2)
• Enables semantic annotation, interoperability,
integration of information
• Enables reasoning
– Extract implicit information
– Assure concept consistency
• Variety of mature, open source tools available
– Protégé, Jena, Virtuoso, D2RQ, …
• Allows information to be exposed as Linked Open
Data (to be discussed later on)
• Data  Information  Knowledge

What is an Ontology?
• In philosophy, Ontology is the study of beings
– Onto (ὤν/ὄντος) + logy (λογία)
– Along with their properties and relations
• In computer science, an ontology is the formal
representation of knowledge
– A formal, explicit specification of a shared
conceptualisation
– Concepts of a domain, objects and their relations
– Allows complexity in schemas
• The RDF and OWL approaches

Lecture Outline


The Resource Description Framework
• The Resource Description Framework is about
describing resources
– Was initially proposed for describing Web resources
• RDF can be viewed as a graph where
– Objects are graph nodes
– Properties are graph edges
foaf:name Graph triples
ex:Author “J. Smith” ex:Author foaf:name “J. Smith”
ex:Author ex:participatesIn ex:Publication
ex:participatesIn
ex:Author foaf:knows ex:anotherAuthor
foaf:knows
ex:Publication
ex:anotherAuthor
National Documentation Centre / NHRF

The RDF Schema (1)
• Describing Web Resources using RDF
– rdfs:Resource
• All things described by RDF are resources
– rdfs:Class
• The class of resources that are classes, i.e. the class of
classes
– rdf:type
• States resource membership
• E.g.: ex:Person rdf:type rdfs:Class
– rdf:Property
• The relations between subjects and objects
November 9, 2011 National Documentation Centre / NHRF

The RDF Schema (2)
– rdfs:SubClassOf
• foaf:Agent rdfs:subClassOf foaf:Person
– rdfs:SubPropertyOf
• Allow class and property hierarchies
• E.g.: ex:hasFirstName rdfs:subpropertyOf ex:hasName
– rdfs:domain
• ex:employer rdfs:domain foaf:Person
– rdfs:range
• ex:employer rdfs:range foaf:Organization


The RDF Schema (3)
– rdfs:Container – rdf:List – rdf:statement
• rdf:Bag • rdf:first – rdf:subject
• rdf:Seq • rdf:rest – rdf:predicate
• rdf:Alt • rdf:nil
– rdf:object
• rdfs:ContainerMembershipProperty
– rdf:value
• rdfs:member
– rdfs:seeAlso
– rdfs:label
– rdfs:isDefinedBy
– rdfs:comment


The RDF Schema (4)
• Example 1
<rdfs:Class rdf:ID="animal" />
<rdfs:Class rdf:ID="horse">
<rdfs:subClassOf rdf:resource="#animal"/>
</rdfs:Class>

• Example 2
<rdf:Description rdf:about="http://www.ekt.gr">
<dc:description>National Documentation Centre</dc:description>
<dc:publisher>NHRF</dc:publisher>
<dc:date>2001-02-16</dc:date>
<dc:format>text/html</dc:format>
<dc:language>el</dc:language>
</rdf:Description>

Web Ontology Language (1)
• Based on Description Logics
– Decidable fragment of First Order Logic
• Allows more complex schema definitions
• OWL builds on top of RDF Woman ≡ Person ∩ Female

Father ≡ Man ∩ ∃hasChild.Person
• Current version is OWL 2 Wife ≡ Woman ∩ ∃hasHusband.Man

rdfs:Resource MotherWithoutDaughter ≡ Mother ∩
∀hasChild. ¬Woman

rdfs:Class rdf:Property

owl:datatypeProperty owl:functionalProperty

owl:Class owl:objectTypeProperty

• Class description • Property description
– owl:intersectionOf – owl:datatypeProperty
– owl:unionOf – owl:objectProperty
– owl:complementOf – owl:equivalentProperty
– owl:equivalentClass – owl:inverseOf
– owl:disjointWith • isTaughtBy ↔ teaches
– Cardinality – owl:functionalProperty
• owl:maxCardinality – owl:inverseFunctionalProperty
• owl:minCardinality – owl:transitiveProperty
• owl:cardinality – owl:symmetricProperty

• owl:Thing • Individuals
• owl:Nothing – owl:sameAs
• Version information – owl:differentFrom
– owl:versionInfo – owl:allDifferent
– owl:priorVersion • Value constraints
– owl:backwardCompatibleWith– owl:allValuesFrom
– owl:incompatibleWith – owl:someValuesFrom
– owl:deprecatedClass – owl:hasValue
– owl:deprecatedProperty

• Example 1
:RedBordeaux rdf:type owl:Class ;
owl:equivalentClass [ rdf:type owl:Class ;
owl:intersectionOf ( :Bordeaux :RedWine ) ] .

• Example 2
:locatedIn rdf:type owl:ObjectProperty ,
owl:TransitiveProperty ;
rdfs:domain owl:Thing ;
rdfs:range :Region .

• Example 3
:BordeauxRegion rdf:type owl:NamedIndividual ,
:Region ;
November 9, 2011
:locatedIn :FrenchRegion .

• Example 4
:hasColor rdf:type owl:FunctionalProperty ,
owl:ObjectProperty ;
rdfs:domain :Wine ;
rdfs:range :WineColor ;
rdfs:subPropertyOf :hasWineDescriptor .

• Example 5
:CabernetSauvignon rdf:type owl:Class
owl:equivalentClass [ rdf:type owl:Class ;
owl:intersectionOf ( :Wine
[ rdf:type owl:Restriction ;
owl:onProperty :madeFromGrape ;
owl:hasValue :CabernetSauvignonGrape
] [ rdf:type owl:Restriction ;
owl:onProperty :madeFromGrape ;
November 9, 2011
owl:maxCardinality "1"^^xsd:nonNegativeInteger] ) ] ;

• OWL 1 flavors
– OWL Full, full language expressivity
– OWL DL, maximal subset allowing reasoner support
– OWL Lite, minimal useful subset of language features
• OWL 2 profiles
– OWL 2 EL, for large numbers of classes/properties
– OWL 2 QL, large volume of instance data support,
relational database-friendly
– OWL 2 RL, RDFS with extra expressivity, scalable
reasoning

Reasoning
• Check ontology consistency
• Class expression subsumption
• Concept satisfiability
• Infer implicit information
– Produces extra (inferred) triples
• Numerous reasoners available
– Free
• Pellet, FaCT++, Jena, Hermit
– Non-free
• OWLIM, OntoBroker

Ontology Authoring (1)
• Protégé is a prominent
GUI solution
– Java-based, open-source
– OWL/RDF capabilities
– Includes FaCT++ reasoner
– WebProtégé in beta
– Extensible through
plugins
• E.g. Ontograf

Available online at http://protege.stanford.edu/

November 9, 2011

Ontology Authoring (2)
• Using HP’s Jena
– Large, active community
– Apache Maven group id com.hp.hpl.jena
– API Example
String ns = "http://example.com/sample#";
Model model = ModelFactory.createDefaultModel();
Resource resource = model.createResource(ns +
"Individual1");
resource.addProperty(DC.title, title);
model.write(file, “RDF/XML”);


Adding Reasoning Capabilities
• Using HP’s Jena
– Create an RDFS model using the Jena API
String ns = "http://www.example.com/ex#";
Model rdfsEx = ModelFactory.createDefaultModel();
Property p = rdfsEx.createProperty(ns, "p");
Property q = rdfsEx.createProperty(ns, "q");
rdfsEx.add(p, RDFS.subPropertyOf, q);
rdfsEx.createResource(NS+"a").addProperty(p, "foo");
– Adding the internal RDFS reasoner
Reasoner reasoner = ReasonerRegistry.getRDFSReasoner();
InfModel inf = ModelFactory.createInfModel(reasoner, rdfsEx);


Querying Ontologies
• SPARQL is to ontologies what SQL is to
relational databases
– W3C recommendation since 2008
• Designed using an SQL-like syntax
– SELECT … FROM … WHERE
• The WHERE conditions are a triple pattern
• Returns graphs instead of a tables
• Example
SELECT ?x ?y ?z WHERE { ?x ?y ?z }
returns all theNational Documentation Centregraph
November 9, 2011
triples in the / NHRF

Introduction to SPARQL (1)
• Selecting a single value
SELECT ?x
WHERE { ?x <ex:hasName> "John Smith" }
• Matching values from a graph
SELECT ?x ?fname
WHERE {?x <ex:hasName> ?fname}
• Also
SELECT ?name ?value
WHERE { ?x <ex:hasAttribute> ?attr .
November 9, 2011
?attr <ex:hasValue> ?value . }
National Documentation Centre / NHRF

• String matching using regular expressions
SELECT ?y
WHERE
{ ?x vcard:Given ?y .
FILTER regex(?y, "r", "i") }
• Filtering values
SELECT ?resource
WHERE {
?resource info:age ?age .
FILTER (?age >= 24) }

• The OPTIONAL construct to return information
where available
SELECT ?name ?age
WHERE {
?person vcard:FN ?name .
OPTIONAL { ?person info:age ?age } }
• The UNION construct
SELECT ?name
WHERE {
{ [] foaf:name ?name } UNION { [] vCard:FN ?name }

• Result handling
– ORDER BY, DISTINCT, OFFSET and LIMIT
• Same as in SQL
– CONSTRUCT
• Ability to construct a new graph based on the results
PREFIX foaf: <http://xmlns.com/foaf/0.1/>
CONSTRUCT { <http://example.com/person#Alice>
foaf:knows ?x }
FROM <http://example.org/foaf/people>
WHERE { ?x foaf:name ?name }
ORDER BY desc(?name)
November 9, 2011 LIMIT 10 National Documentation Centre / NHRF

Common Vocabularies (1)
• DC
– Describe library asset information
• SKOS
– Simple Knowledge Organization Scheme
• FOAF
– Friend of a friend
• SIOC
– Semantically Interlinked Online Communities
• DBPedia
– Extract structured information from Wikipedia 29

• Music ontology
– Describe music concepts
• Good relations
– Used in the e-commerce context
– Supported by Google and Yahoo
• Basic Geo Vocabulary
– Expresses spatial information using WGS84
• Creative commons
– Express copyright information

• Microformats are open data standards for
publishing structured information on the Web
• Simple, solve specific problems
• No change in
display
• Examples
– hCard
– hCalendar
– RDFa
• For SEO, see also schema.org

Embedded RDF
• RDFa
– Embed RDF in XHTML documents
– Uses <span>, <div>
– Allows nested
descriptions
• GRDDL
– Obtain RDF from
HTML pages
– Uses XSLT for XML

32
Both RDFa and GRDDL are W3C recommendations

Ontology Serialisation Formats
• RDF+XML
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:dc="http://purl.org/dc/elements/1.1/">
<rdf:Description rdf:about="http://en.wikipedia.org/wiki/Tony_Benn">
<dc:title>Tony Benn</dc:title>
<dc:publisher>Wikipedia</dc:publisher>
</rdf:Description>
</rdf:RDF>

• N3 and Turtle (Turtle ⊆ N3)
@prefix dc: <http://purl.org/dc/elements/1.1/>.
<http://en.wikipedia.org/wiki/Tony_Benn>
dc:title "Tony Benn";
dc:publisher "Wikipedia".

• ... and, of course, in the database!


Lecture Outline


Triplestores
• A triplestore contains records in the form
– (subject, predicate, object)
• Uses a relational database backend
• Saving an Ontology in a triplestore
– Jena
• http://jena.sourceforge.net
– Sesame
• http://www.openrdf.org
– Virtuoso
• http://virtuoso.openlinksw.com
– Oracle
• http://www.oracle.com/technetwork/database

RDB to RDF Mapping Language
• Several tools proposed in the early years
– Triplify, D2OMapper, DB2OWL, VisaVis, R2O,
MapOnto, …
• R2RML: a W3C working draft
– Implementations
• D2RQ
• Virtuoso


Virtuoso Overview (1)
• Open source and commercial version
• Can be used as
– A web application server
– A relational database repository
• Offers a JDBC Driver
• Collaborates with Jena
• Offers Conductor, a GUI for server administration
– A web service server
– A triplestore
• Export RDF data from same DB or others

• RDF Views
– Export relational data as triples
• SPARQL 1.1 support, plus
– Full Text Queries
– Geo Spatial Queries
– Business Analytics and Intelligence
– SQL Stored Procedure and Built-In Function
exploitation from SPARQL
– Create, Update, and Delete (SPARUL)
• Cluster Configuration
– Parallel and Horizontal scaling

• Extendable through VAD* Packages
– Interactive SPARQL Query Builder
• A GUI to create SPARQL queries
– Sponger Middleware
• Offers RDF Mappers to import data into Virtuoso
– PubSubHub Protocol (for RSS)
• Can be used to allow push behavior and subscriptions
by clients
– OAT (OpenLink AJAX Toolkit) Framework
• Rich web application development

* Virtuoso Application Distribution

Virtuoso Reasoning Engine
• Backward-chaining OWL reasoner coverage
– rdfs:subClassOf
– rdfs:subPropertyOf
– owl:sameAs
– owl:equivalentClass
– owl:equivalentProperty
– owl:InverseFunctionalProperty
– owl:inverseOf
– owl:SymmetricalProperty
– owl:TransitiveProperty


Virtuoso Sponger
• An RDF-iser to bring data into the Semantic
Web
• Sponger extracts RDF data from non-RDF
sources
• A Cartridge per data source
• XSLT templates do the work
• Customisable and Programmable
– Virtuoso PL, C++, Java


Virtuoso as a DB Server (1)
• Conductor: a GUI for server administration

42

Virtuoso as a DB Server (2)
• Can export data as RDF using RDF Views

43

Virtuoso as an RDF Server (1)
• A URI for every resource, browseable repository

44

• Example: Measurement URI


• RDF data also accessible via
– ODBC, JDBC, OLE DB, XMLA, ADO.NET
• Difficulties in extracting RDF Data
– Tables must have a primary key
– Mappings are defined using regular expressions
and tend to be complicated
DB.DBA.URLREWRITE_CREATE_REGEX_RULE ( 'demo_rule6‘, 1, '/Demo/objects/(*^#+*)‘, vector('path'), 1,
'/sparql?query=DESCRIBE+%%3Chttp%%3A//^{URIQADefaultHost}^/Demo/objects/%U%%3E+FROM+%
%3Chttp%%3A//^{URIQADefaultHost}^/Demo%%23%%3E&format=%U‘, vector('path', '*accept*'), null,
'(text/rdf.n3)|(application/rdf.xml)‘, 2, null );


Virtuoso Jena Provider
• Offered by
OpenLink
• Native Graph
Model Storage
Provider
• Enables access to
the Virtuoso RDF
Quad store
through Jena

Querying Remote Repositories
• XML over HTTP (RESTful approach)
– http://demo.openlinksw.com/sparql?default-
graph-
uri=urn:lsid:ubio.org:namebank:11815&should-
sponge=soft&query=SELECT+*+WHERE+{?s+?p+?
o}&format=text/html
• No create/update/delete capabilities


SPARQL Query Interface (1)
• SPARQL queries can be named and stored
– A query named sparql-demo listens to:
http://localhost:8890/DAV/sparql-demo
• Can return results over HTTP (XML by default)
• MIME type of the RDF data
– 'rdf+xml' (default) | 'n3' | 'turtle' | 'ttl‘


SPARQL Query Interface (2)
• SPARQL results example in RDF/XML
<ROOT>
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#” xmlns:rs="http://www.w3.org/2005/sparql-results#"
xmlns:xsd="http://www.w3.org/2001/XMLSchema#">
<rs:results rdf:nodeID="rset">
<rs:result rdf:nodeID="sol193">
<rs:binding rdf:nodeID="sol193-0" rs:name="x“>
<rs:value rdf:resource="http://localhost:8890/Demo/temperature/PK/4#this"/>
</rs:binding>
<rs:binding rdf:nodeID="sol193-1" rs:name="y“>
<rs:value rdf:resource="http://www.w3.org/1999/02/22-rdf-syntax-ns#type"/></rs:binding>
<rs:binding rdf:nodeID="sol193-2" rs:name="z“>
<rs:value rdf:resource="http://localhost:8890/schemas/Demo/temperature"/>
</rs:binding>
</rs:result>
…
</rs:results>
</rdf:RDF>
</ROOT>


Lecture Outline


The Linked Open Data Cloud (1)
• Data available on the Web
– Under an open license
• Available as structured data
– Excel sheet instead of a scanned image
• Use non-proprietary format
– CSV, RDF instead of DOC, XLS
• Use linked data format
– URIs to identify things
• Linked to other people’s data
– Provision of context 52
Also see: http://lab.linkeddata.deri.ie/2010/star-scheme-by-example/

• Interconnected datasets using URI’s and RDF

2011

2007

Source: http://linkeddata.org
Also see the datahub: http://thedatahub.org/group/lodcloud

• Consumer capabilities
– Access it, print it, store it locally, enter the data in
another system
– Process, aggregate, visualise, manipulate, export
in another format, reuse
– Avoid vendor lock-ins
• Publisher capabilities
– Make data discoverable
– Increase the value of the data
• Allow added-value services
– Fine-granular control over the data
Also see: http://lab.linkeddata.deri.ie/2010/star-scheme-by-example/
54

Open Government Data
• Prominent examples include
– data.gov (US)
– data.gov.uk, (UK)
– data.london.gov.uk (UK)
– digitaliser.dk (DK)
– data.govt.nz (NZ)
– linkedopendata.it (IT)
– geodata.gov.gr (GR)

Promotional video: http://opengovernmentdata.org/film/

The DBPedia Project
• Structured information based on Wikipedia
• SPARQL endpoint: dbpedia.org/sparql
• Example: Greece on DBPedia

New York Times Public Data
• News
data


The Claros Project
• The world of art on the Semantic Web
• Uses the CIDOC-CRM vocabulary to describe
– Objects
– Places
– Periods
– People
• OWL DL
• RESTful

November 9, 2011

The OpenCalais Project
• Creates semantic metadata for submitted
content
• By Thomson
Reuters
• Free to all
• Extracts RDF
using NLP
• Uses WS

November 9, 2011

The TrueKnowledge Engine
• Based on facts, not keywords

Thank you for your attention!
Questions?

Appendix
• Installing Virtuoso as a system service
– Windows 7
• Download and extract from
http://virtuoso.openlinksw.com/dataspace/dav/wiki/Mai
n/VOSDownload
• Open a command line prompt as administrator
• Register required DLL
– regsvr32 virtodbc.dll
• Install service
– virtuoso-t +service screate +instance "DB" +configfile virtuoso.ini
– Ubuntu 10.04 LTS
• sudo apt-get install virtuoso-server

Publishing Data Using Semantic Web Technologies

More Related Content

What's hot

Viewers also liked

Similar to Publishing Data Using Semantic Web Technologies

More from Nikolaos Konstantinou

Publishing Data Using Semantic Web Technologies