Rhizomer Semantic Content Management System

1
Rhizomer Semantic Content Management System

• Roberto García, Juan Manuel Gimeno, Ferran
  Perdrix, Rosa Gil, Marta Oliva

HCI and Data Integration Research
Group Universitat de Lleida, Spain
Segre Media, Spain
2
Contents

• Introduction
• The Rhizomer Platform
• Use Case S5T Project
• Conclusions Future Work

3
Contents

• Introduction
• The Rhizomer Platform
• Use Case S5T Project
• Conclusions Future Work

4
Introduction

• Difficult to manage Information Architecture with
  current CMSs/Wikis
• Use Semantic Web technologies
• Semantic CMSs/Wikis dont fully exploit them
• Most semantic wikis
• focus on page annotation, limited reasoning, no
  import...
• Drupal RDF Module a Drupal extension, though
  Drupal 7 based on SWT 1
• 1 Buytaert, D. DrupalCon Boston 2008 State of
  Drupal presentation.http//buytaert.net/files/sta
  te-of-drupal-2008-boston.pdf

5
Contents

• Introduction
• The Rhizomer Platform
• Use Case S5T Project
• Conclusions Future Work

6
Rhizomer

• Platform for Semantic Content Management and
  Human-Semantic Web Interaction
• Based on Web 2.0 technologies
• Client (AJAX) / Server (REST)
• Web 3.0 data model
• Semantic metadata store RDF(OWL) / SPARQL
• Generic metadata browsing
• HTML Javascript
• Object-Action Interaction Paradigm
• Semantic Web Services

7
Rhizomer
search
new
edit
delete
GET
PUT
POST
DEL

• Server part
• REST interface over the repository
• Jena, Sesame,
• GET SPARQL end-point
• PUT/POST/DELETE RDF metadata management
• Semantic Web Services
• OWL-S descriptions
• Service Profile Process, hasInput, hasOutput
• REST adapters for
• Javascript APIs (e.g. Google Maps), SOAP,

Rhizomer
MetadataStore
8
Rhizomer

• Client part
• AJAX (asynchronously send queries and metadata)
• RDF/XML -XSLT? HTML (browsing)
• Semantic forms (edition)
• RDF/XML -XSLT? FORM -Javascript? RDF

RDF
JavaScript
ltHTML/gt
ltHTML/gt SemFORMS
9
Contents

• Introduction
• The Rhizomer Platform
• Use Case S5T Project
• Conclusions Future Work

10
S5T Project

• S5T Spanish research project (TIN2005-06885)
• Scaleable Semantic personalised Search of Spoken
  and written contents on the Semantic Web
• Application in the Segre media group
• Newspaper, TV, radio and web site
• Internal use (media house) facilitate content
  production, history tracking, contextualisation,
• External use (web users) advanced services (e.g.
  news personalisation, semantic podcasts,)

11
S5T Project

• Semantic annotation of audiovisual content
• voice ? text (keywords) ? semantic annotations
• Rhizomer user interface for browsing/interacting
  with content and associated knowledge

12
S5T Project query
s5tAudiospecific properties P rdfsdomain
s5tAudio or s5tAudio owlRestriction on P (or
superclass)

• Queries
• Metadata
• Content (transcript)
• Words
• Concepts

GET
ltform methodGET onSumbitformToSPARQLgt ltinpu
t typetext name http//purl.org/dc/elements
/1.1/title/gt ltinput typetext
name http//purl.org/dc/elements/1.1/date/gt
lt/formgt
?query DESCRIBE ?s WHERE  ?s dctitle ?y0
s5tgenre ?y1. FILTER regex(?y0, "")
?y1lt./SubjectsOnto 11000000gt
13
S5T Project store

• GET ? SPARQL endpoint
• Metadata store
• Jena, Sesame,
• DESCRIBE queries by default. Facilitate
• Browse heterogeneous metadata
• Without a priori knowledge
• DESCRIBE implementation
• Concise Bounded Description rdfslabels for
  all selected resources
• Result RDF/XML

14
S5T Project store

• DESCRIBE http//rhizomik.net/rosa CBD(rosa)
  Fragment A U rdfslabels

Fragment A
15
S5T Project browse
lts5tAudio rdfabout"http//www.segre.com/audio/2
0070323-bm.mp3"gt      ltdctitlegtButlletí
Migdialt/dctitlegt      ltdcdategt2007-03-23lt/dcda
tegt      lts5tgenre rdfresource"srs11000000"/
gt      lts5ttranscriptResourcegt             
http//www.segre.com/audio/20070323-bm.xml      
lt/s5ttranscriptResourcegt lt/srtAudiogt
RDF2HTML
http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio
title Butlletí Migdia
date 2007-03-23
genre politics
transcript http//www.segre.com/audio/20070323-bm.xml
Referrers Referrers
http//rhizomik.net/.../NewsCodes-SubjectsOnto11000000 a Subject http//rhizomik.net/.../NewsCodes-SubjectsOnto11000000 a Subject
label politics
description Local, regional, national and international exercise of power, or struggle for power, and the relationships between governing bodies and states.
Referrers Referrers
GET ? RDF2HTML
16
S5T Project edition
edit del new edit del new
http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio
title Butlletí Migdia
date 2007-03-23
genre politics
transcript http//www.segre.com/audio/20070323-bm.xml
Referrers Referrers
17
S5T Project edition

• Semantic form
• ltform methodPUT/POST onSumbitformToNTriplesgt
  ltinput typetext namehttp//...w3.org/.../rdf
  -syntax-nsID valuehttp//www.segre.com/audi
  o/20070323-bm.mp3/gt
• ltinput typetext namehttp//...w3.org/.../rdf
  -syntax-nstype valuehttp//rhizomik.net/.../
  s5t.owlAudio/gt
• ltinput typetext namehttp//rhizomik.net/.../
  s5t.owlgenre/ valuehttp//.../NewsCodes-Subje
  ctsOnto11000000 /gt
• lt/formgt
• PUT description for new resource
• POST replace resource description

ltSubject,
Predicate,Objectgt
Predicate,Objectgt
18
S5T Project uploading

• Semantically annotate uploaded content(facilitate
  metadata creation)
• Content-type specific metadata extraction plugins
• Audio text-to-speech
• Text keyword detection, disambiguation
• OpenCalais, Freeling, named entities
• Photo EXIFtoRDF,

19
S5T Project interaction

• Given a resource (object)determine available
  services (actions)
• Resource description type, properties, values,

http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio
title Butlletí Migdia Butlletí Migdia
date 2007-03-23 2007-03-23
genre politics politics
transcript http//www.segre.com/audio/20070323-bm.xml http//www.segre.com/audio/20070323-bm.xml
Referrers
20
S5T Project interaction

• Services descriptions
• ID Service Adapter URL
• hasInput an OWL Class
• Service applicable to any resource classified in
  this class
• hasOutput output content type
• text/html, application/rdfxml,

http//rhizomik.net/services/play a Process http//rhizomik.net/services/play a Process http//rhizomik.net/services/play a Process
label Play Play
hasInput http//rhizomik.net/.../s5t.owlAudio http//rhizomik.net/.../s5t.owlAudio
hasOutput text/html text/html
Referrers
21
S5T Project interaction

• For each resource
• Processes s.t. resource rdftype hasInput Class
• Call REST PUT/POST RDF resource
• Adapter service invocation(Javascript, SOAP,)

http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio
title Butlletí Migdia Butlletí Migdia
date 2007-03-23 2007-03-23
genre politics politics
transcript http//www.segre.com/audio/20070323-bm.xml http//www.segre.com/audio/20070323-bm.xml
Play Play Referrers
22
S5T Project interaction

• Butlletí Migdia (23/03/2007)
• La mobilització en contra dels transgènics i en
  favor de Josep Pàmies també ha servit per
  introduir altres reclamacions. En aquest cas,
  alguns dels collectius de la lluita contra
  aquests cultius demanen que la Universitat de
  Lleida rebi una especialització en Agricultura
  Ecològica. Asseguren que serien uns estudis
  pioners que servirien al centre per recuperar
  prestigi.

http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio http//www.segre.com/ audio/ 20070323-bm.mp3 a Audio
title Butlletí Migdia Butlletí Migdia
date 2007-03-23 2007-03-23
genre politics politics
transcript http//www.segre.com/audio/20070323-bm.xml http//www.segre.com/audio/20070323-bm.xml
Play Play Referrers
Search Keyword describe resources whose
transcript contains keyword Browse Term
describe the concept annotating the keyword
http//www.udl.cat a University http//www.udl.cat a University http//www.udl.cat a University
label Universitat de Lleida Universitat de Lleida
lat 41.37 41.37
long 0.37 0.37
Map Map Referrers
Search Keyword Browse Term
23
S5T Project interaction

• Map any resource with lat/long
• ltprocessProcess rdfabout"http//rhizomik.net/s
  ervices/map"gt  ltrdfslabelgtmaplt/rdfslabelgt  ltpr
  ocesshasInputgt  ltowlClass rdfID"GeolocatedE
  ntity"gt    ltowlintersectionOf
  rdfparseType"Collection"gt    
  ltowlRestrictiongt      ltowlonProperty
  rdfresource"poslat"/gt     
  ltowlminCardinalitygt1lt/owlminCardinalitygt   
    lt/owlRestrictiongt     ltowlRestrictiongt
        ltowlonProperty rdfresource"poslong"
  /gt      ltowlminCardinalitygt1lt/owlminCardina
  litygt    lt/owlRestrictiongt  
  lt/owlintersectionOfgt  lt/owlClassgt  lt/proc
  esshasInputgt  ltprocesshasOutputgttext/htmllt/proc
  esshasOutputgtlt/processProcessgt
• Service Adapter includes GoogleMap Javascript
• The same for events (Timeline, Calendar,), etc.

24
Contents

• Introduction
• The Rhizomer Platform
• Use Case S5T Project
• Conclusions Future Work

25
Conclusions

• Content Management System based on semantic
  metadata
• Framework for Human-Semantic Web Interaction
• Flexible
• Projects S5T, Rhizomik,
• Extensible
• Plugins for visualisation, interaction and
  metadata extraction
• Usable
• Preliminary usability tests

26
Future Work

• Integrate S5T project interface
• Assisted metadata edition
• Properties and values recommended based on
  available ontologies
• Create basic services set
• Map
• Time line
• Calendar
• Plot
• Complete usability test

27
Thank you for your attentionMore
http//rhizomik.net

• Roberto García
• http//rhizomik.net/roberto