Transcript Publishing data on the Web (with SPARQL)

Publishing data on the
Web (with SPARQL)
Andy Seaborne,
Semantic and Adaptive Systems Dept
© 2006 Hewlett-Packard Development Company, L.P.
The information contained herein is subject to change without notice
Publishing on the Web
•
Making information available
− without knowing the eventual use
− reuse, collaboration
− reproduction of results
•
This talk
− About SPARQL and RDF
− Not covered (but important)
• Economic issues
• DRM, copyright and licensing issues
• ...
Publishing on the Semantic Web
•
Data Access
 SPARQL
•
Information organisation
 OWL, RDFS, SKOS
•
Information format
 RDF
•
Identification
 URIs
•
Serialization
 XML
Publishing via a query service on the web
3
RDF/OWL
•
Why using RDF/OWL?
− natural representation for graphs of relationships
− extensibility, facilities an open eco-system
− ease of model integration
•
•
•
•
composing data from multiple sources
ontology not schema
ease of transformation
variable levels of abstraction
− support for partial models
− inference
SPARQL
SPARQL =
Query Language + Protocol + XML Results
Format
•
Access and query RDF graphs
•
Product of the RDF Data Access Working Group
•
Status: W3C Candidate Recommendation
5
SPARQL Query
PREFIX dc: <http://purl.org/dc/elements/1.1/>
SELECT ?title2
WHERE
{
?doc
dc:title
?doc
dc:creator
?docOther dc:creator
?docOther dc:title
"SPARQL at speed" .
?c .
?c .
?title2
}
•
On an abstracts/papers database:
“Find other papers by the authors of a given paper.”
6
SPARQL Query
PREFIX dc: <http://purl.org/dc/elements/1.1/>
PREFIX foaf: <http://xmlns.com/foaf/0.1/>
PREFIX shop: <http://example/shop#>
SELECT ?title2
WHERE
{
?doc
dc:title
?title .
FILTER regex(?title, "SPARQL") .
?doc
dc:creator ?c .
?c
foaf:name
?name .
OPTIONAL
{ ?doc shop:price ?price }
}
•
“Find books with ‘SPARQL’ in the title. Get the authors’
name and the price (if available).”
•
Multiple vocabularies
7
Inference
•
An RDF graph may be backed by inference
− OWL, RDFS, application, rules
:x rdf:type :C .
:C rdfs:subClassOf :D .
PREFIX rdf:
<http://www.w3.org/1999/02/22-rdf-syntax-ns#>
SELECT ?type
WHERE
{
?x rdf:type ?type .
}
-------| type |
========
| :C
|
| :D
|
-------8
SPARQL : Data Virtualization
•
SPARQL as integrator
− Data remains where it is
− Existing applications untouched
− data appears as RDF, remap query to native form
•
SPARQL to SQL
• Direct mapping of tables
− Semi-automatic generation of mapping
− SquirrelRDF (also does LDAP)
• Modelled: D2RQ
− High-quality mapping, manually developed
Federated Query
Single Point of Access
SPARQL
Query
• Inputs:
RDF
Corp
LDAP
SPARQL
Query
– Unified results
SPARQL
Query
• Outputs:
Query Broker
SPARQL => SPARQL
SPARQL
Query
– Service
Description
– Information
Directory
– Request
Doc
DB
10
Access Multiple Sources
PREFIX dc:
PREFIX ldap:
PREFIX foaf:
<http://purl.org/dc/elements/1.1/>
<http://ldap.hp.com/people#>
SELECT ?name ?name2
{
?doc
dc:title
?title .
FILTER regex(?title, “SPARQL”) .
?doc
dc:creator
?reseacher .
?researcher
ldap:email
?email .
?researcher
ldap:name
?name
}
•
“Find the name and email addresses of authors of a
paper”
•
Query tech reports first, then go to LDAP for more
information.
•
SQL database; LDAP server
11
Access Multiple Sources
PREFIX dc:
PREFIX ldap:
PREFIX foaf:
<http://purl.org/dc/elements/1.1/>
<http://ldap.hp.com/people#>
SELECT ?name ?name2
{
?doc
dc:title
?doc
dc:creator
?researcher
ldap:email
}
?title .
?reseacher .
<mailto:[email protected]> .
•
“Find the reports written by the person with
email address [email protected]
•
Go to LDAP first, then the reports database
12
SPARQL : Protocol
SPARQL =
Query Language + Protocol + XML Results
Format
•
One request type expressed in WSDL
•
HTTP and SOAP
•
Results: fixed XML form for further transformation
13
Links
•
Jena: Java and .Net Semantic Web Framework
− http://jena.sourceforge.net/
− RDF, SPARQL, OWL, Rules
− Open source: BSD-license
• Commercial and non-commercial use
•
SPARQL Query
− http://jena.sourceforge.net/ARQ
•
SPARQL Protocol
− http://www.joseki.org
•
SquirrelRDF: Access legacy SQL:
− http://jena.sourceforge.net/SquirrelRDF
14
15
SPARQL Update
INSERT
{ book:newBook dc:title
“More about SPARQL” .
book:newBook dc:creator
[ foaf:mbox <mailto:alex@example> ;
foaf:name “Alex Author” ]
}
DELETE
{ ?book dc:title
?title ;
dc:creator ?c }
WHERE
{ ?book dc:title ?title ;
dc:creator ?c
FILTER regex(?title, “first edition”)
}
•
Future HP proposal to W3C
16