Work Package 2
All WP Meeting
28th November 2013, Athens
Kai Eckert

co-funded by the European Union
Timetable
Q1

• 2.1, 2.3: Test of external components (MINT, Silk, jMet2Ont, D2R), Data Survey
• 2.2: Basic entity structure of the model

Q2

• 2.1, 2.4, 2.5: Prototype of Infrastructure, Workflows, UI
• 2.2: Mapping workshops, further work on the DM2E data model

Q3

• Softwaredesign of Intermediate Architecture, Foundational work
• Combination of the DM2E model with Linked Data principles, Integration with Architecture

Q4

• Provenance Model for DM2E Infrastructure, Web service development
• DM2E Data Model 1.0, first stable and operational version.

Q5

• Implementation of the DM2E Data Model in Mint.
• Development of the Intermediate Version of the Infrastructure: due on July 15th.

Q6

• Implementation of OmNom as final version.
• Further refinements of the DM2E model.

Q7

• Setup of the Silk instance, finalization of OmNom (final version due January 2014).
• Data ingestions (Wittgenstein, ÖNB Codices), data evaluation for contextualisation.

16.04.2013

DM2E Review: Work Package 2

2
Data Model (Evelyn, Julia, Steffen)

11.06.2013

DM2E All WP Meeting: Work Package 2

3
DM2E Model Refinements
• Currently working on Version 1.1:
– dm2e:
<http://onto.dm2e.eu/schemas/dm2e/1.1/> .
– dm2edata: <http://data.dm2e.eu/data/> .

• Model specification
– Revision 0.8
– Updated EDM definitions

• New property
– dm2e:hasAnnotatableVersionAt

• Hierarchies now analogue to EDM
– dcterms:hasPart and dcterms:isPartOf
DM2E Model Documentation
Current stable Version: DM2E Model v1.0
• Version 1.0 documentation on dm2e.eu: http://
dm2e.eu/document#DM2EModelSpecification
• Version 1.1 revision history and latest releases on
Redmine
– Revision history
– Model specification
– OWL file
Modelling Issue: Uncertainty
• Thoughts on modelling uncertainty:
– https://dm2e.hu-berlin.de/redmine/projects/wp2/ wiki
/DM2E#Uncertainty-in-ontology-modelling
Provider
Representation in DM2E

• Example of the
simple solution:

data
1909-[19--?] edm:TimeSpan
begin:
1909
end:
uncertainBegin:
uncertainEnd: 1999
[18--?]

edm:TimeSpan
begin:
end:
uncertainBegin: 1800
uncertainEnd:
1899

17XX

edm:TimeSpan
begin:
end:
uncertainBegin: 1700
uncertainEnd:
1799
Mappings
Mappings of the Polytechnische Journal
Consistency Check of the Ontology
• First test using Protege (4.3.0)
– Reasoner (HermiT 1.3.8, FaCT ++)
– discovered some inconsistent inferred axioms
– changed the illogical relations

• Second test using Protege (4.3.0)
– Reasoner (HermiT 1.3.8, FaCT ++)
– included individuals from the UBER mappings
– work in progress
First Test - First Example
Inconsistent Inferred Axiom

Explanation:
First Test - Second Example
Inconsistent Inferred Axiom

Explanation:
Second Test with Individuals
Next Steps of UBER in WP2
DM2E model
• Analyse your mappings
– Which resources are not used?
– Are any resources missing?

• Revise the model
– A last time after the mapping analysis

DM2E to EDM
• XSLT from the specialised to the generalised model
OmNom (Konstantin, Kai)

11.06.2013

DM2E All WP Meeting: Work Package 2

13
The DM2E Data Bridge

This is YOUR data.

This is the void:Dataset
in DM2E.
16.04.2013

DM2E Review: Work Package 2

14
Some more links are actually available...

11.06.2013

DM2E All WP Meeting: Work Package 2

15
Implementation
• Implementation (mostly) completed.
• Currently mainly testing, bug fixes and performance
tuning.
• Final version due: January 2014!
• Optimized visualization as eye candy (compare to
previous slide):

11.06.2013

DM2E All WP Meeting: Work Package 2

16
RESTful Linked Data API
• Mostly unchanged since intermediate version.
• Feedback needed from WP3, but also from other
potential users of the data.
• Feature added on request: Data can be marked as
temporary (the default for now). So ingest and play
with it, it won‘t be carved in stone (yet).

11.06.2013

DM2E All WP Meeting: Work Package 2

17
Alternative ingestion tool
• Fallback solution: Direct XSLT transformation and ingestion.
• Command line based, no GUI, only basic provenance information
(but complete).
• Why?
– Avoid bottleneck due to ongoing development of OmNom
– performance evaluations.

Available on request. Or provide links to your data files and an XSLT
mapping and we (Dominique, Kai) are happy to ingest for you.

11.06.2013

DM2E All WP Meeting: Work Package 2

18
Single Sign On, Security (Jorge)

11.06.2013

DM2E All WP Meeting: Work Package 2

19
Authentication Service
●
●
●
●
●
●
●

Centralized Authentication and Authorization Service
Centralized Storage of User Accounts
User Account Schema based on MINT Model.
Single Sign-On
Standard Based: JAAS, Web Services/SOAP
"Remember Me" support
Password reset support
Status
• Implementation of WS for Single Sign-on (WS-SSO) (
http://dm2e-security.rz-berlin.mpg.de/josso/).
• Implementation of a User Manager Application
integrated with WS-SSO (
http://dm2e-security.rz-berlin.mpg.de/userManager/).
• OmNom integration.

11.06.2013

DM2E All WP Meeting: Work Package 2

21
Contextualisation (Dominique, Klaus)

11.06.2013

DM2E All WP Meeting: Work Package 2

22
Status
• Evaluation of provided datasets regarding the
contextualisation.
• Setup of a Silk instance with preloaded
contextualisation targets (GND, LCSH).
• Planned: Geonames, VIAF, Yago, Freebase, DDC, InPho,
JudaicaLink, …
• Directly usable: Dbpedia

16.04.2013

DM2E Review: Work Package 2

23
JudaicaLink (Dov, Kai)
• Permission to create and publish Linked Data from:
– Russian Jewish Encyclopedia (rujen.ru)
– YIVO Encyclopedia

• UMA hired a student assistant to work with both
encyclopedias and create Linked Data representations.
• More encyclopediae would be great (Contact Dov)
• Further information: JudaicaLink.org

11.06.2013

DM2E All WP Meeting: Work Package 2

24
Deployment (Julia)

11.06.2013

DM2E All WP Meeting: Work Package 2

25
Berlin Server
• What do we have?
– 2 Processors (Multicore)
– 64 GB RAM
– 4 TB Hard Drive (Usable space might be less, depends on
RAID configuration)
– That‘s all, we won‘t get more!

• UBER and UMA will reinstall the server from scratch.
• Virtual Machines:
– Ontologies, OWLIM, OmNom, Silk, Pundit, Redmine, …

• Downtimes in December, will be announced.
16.04.2013

DM2E Review: Work Package 2

26
Next steps
• Final Version (January 2014)
–
–
–
–

Complete transformation and ingestion infrastructure
Integrated contextualization
Connection with scholarly environment (WP3)
Documentation Drafts

• Data ingestions(!)
• Maintenance phase
–
–
–
–

Bug fixing and performance tuning in OmNom
Documentation and introductory materials with WP4
Data provision for Europeana
Search and browse interface

16.04.2013

DM2E Review: Work Package 2

27
Thank you.

16.04.2013

DM2E Review: Work Package 2

28

All-WP Meeting 4 Athens, Workpackage 2, Kai Eckert

  • 1.
    Work Package 2 AllWP Meeting 28th November 2013, Athens Kai Eckert co-funded by the European Union
  • 2.
    Timetable Q1 • 2.1, 2.3:Test of external components (MINT, Silk, jMet2Ont, D2R), Data Survey • 2.2: Basic entity structure of the model Q2 • 2.1, 2.4, 2.5: Prototype of Infrastructure, Workflows, UI • 2.2: Mapping workshops, further work on the DM2E data model Q3 • Softwaredesign of Intermediate Architecture, Foundational work • Combination of the DM2E model with Linked Data principles, Integration with Architecture Q4 • Provenance Model for DM2E Infrastructure, Web service development • DM2E Data Model 1.0, first stable and operational version. Q5 • Implementation of the DM2E Data Model in Mint. • Development of the Intermediate Version of the Infrastructure: due on July 15th. Q6 • Implementation of OmNom as final version. • Further refinements of the DM2E model. Q7 • Setup of the Silk instance, finalization of OmNom (final version due January 2014). • Data ingestions (Wittgenstein, ÖNB Codices), data evaluation for contextualisation. 16.04.2013 DM2E Review: Work Package 2 2
  • 3.
    Data Model (Evelyn,Julia, Steffen) 11.06.2013 DM2E All WP Meeting: Work Package 2 3
  • 4.
    DM2E Model Refinements •Currently working on Version 1.1: – dm2e: <http://onto.dm2e.eu/schemas/dm2e/1.1/> . – dm2edata: <http://data.dm2e.eu/data/> . • Model specification – Revision 0.8 – Updated EDM definitions • New property – dm2e:hasAnnotatableVersionAt • Hierarchies now analogue to EDM – dcterms:hasPart and dcterms:isPartOf
  • 5.
    DM2E Model Documentation Currentstable Version: DM2E Model v1.0 • Version 1.0 documentation on dm2e.eu: http:// dm2e.eu/document#DM2EModelSpecification • Version 1.1 revision history and latest releases on Redmine – Revision history – Model specification – OWL file
  • 6.
    Modelling Issue: Uncertainty •Thoughts on modelling uncertainty: – https://dm2e.hu-berlin.de/redmine/projects/wp2/ wiki /DM2E#Uncertainty-in-ontology-modelling Provider Representation in DM2E • Example of the simple solution: data 1909-[19--?] edm:TimeSpan begin: 1909 end: uncertainBegin: uncertainEnd: 1999 [18--?] edm:TimeSpan begin: end: uncertainBegin: 1800 uncertainEnd: 1899 17XX edm:TimeSpan begin: end: uncertainBegin: 1700 uncertainEnd: 1799
  • 7.
    Mappings Mappings of thePolytechnische Journal
  • 8.
    Consistency Check ofthe Ontology • First test using Protege (4.3.0) – Reasoner (HermiT 1.3.8, FaCT ++) – discovered some inconsistent inferred axioms – changed the illogical relations • Second test using Protege (4.3.0) – Reasoner (HermiT 1.3.8, FaCT ++) – included individuals from the UBER mappings – work in progress
  • 9.
    First Test -First Example Inconsistent Inferred Axiom Explanation:
  • 10.
    First Test -Second Example Inconsistent Inferred Axiom Explanation:
  • 11.
    Second Test withIndividuals
  • 12.
    Next Steps ofUBER in WP2 DM2E model • Analyse your mappings – Which resources are not used? – Are any resources missing? • Revise the model – A last time after the mapping analysis DM2E to EDM • XSLT from the specialised to the generalised model
  • 13.
    OmNom (Konstantin, Kai) 11.06.2013 DM2EAll WP Meeting: Work Package 2 13
  • 14.
    The DM2E DataBridge This is YOUR data. This is the void:Dataset in DM2E. 16.04.2013 DM2E Review: Work Package 2 14
  • 15.
    Some more linksare actually available... 11.06.2013 DM2E All WP Meeting: Work Package 2 15
  • 16.
    Implementation • Implementation (mostly)completed. • Currently mainly testing, bug fixes and performance tuning. • Final version due: January 2014! • Optimized visualization as eye candy (compare to previous slide): 11.06.2013 DM2E All WP Meeting: Work Package 2 16
  • 17.
    RESTful Linked DataAPI • Mostly unchanged since intermediate version. • Feedback needed from WP3, but also from other potential users of the data. • Feature added on request: Data can be marked as temporary (the default for now). So ingest and play with it, it won‘t be carved in stone (yet). 11.06.2013 DM2E All WP Meeting: Work Package 2 17
  • 18.
    Alternative ingestion tool •Fallback solution: Direct XSLT transformation and ingestion. • Command line based, no GUI, only basic provenance information (but complete). • Why? – Avoid bottleneck due to ongoing development of OmNom – performance evaluations. Available on request. Or provide links to your data files and an XSLT mapping and we (Dominique, Kai) are happy to ingest for you. 11.06.2013 DM2E All WP Meeting: Work Package 2 18
  • 19.
    Single Sign On,Security (Jorge) 11.06.2013 DM2E All WP Meeting: Work Package 2 19
  • 20.
    Authentication Service ● ● ● ● ● ● ● Centralized Authenticationand Authorization Service Centralized Storage of User Accounts User Account Schema based on MINT Model. Single Sign-On Standard Based: JAAS, Web Services/SOAP "Remember Me" support Password reset support
  • 21.
    Status • Implementation ofWS for Single Sign-on (WS-SSO) ( http://dm2e-security.rz-berlin.mpg.de/josso/). • Implementation of a User Manager Application integrated with WS-SSO ( http://dm2e-security.rz-berlin.mpg.de/userManager/). • OmNom integration. 11.06.2013 DM2E All WP Meeting: Work Package 2 21
  • 22.
  • 23.
    Status • Evaluation ofprovided datasets regarding the contextualisation. • Setup of a Silk instance with preloaded contextualisation targets (GND, LCSH). • Planned: Geonames, VIAF, Yago, Freebase, DDC, InPho, JudaicaLink, … • Directly usable: Dbpedia 16.04.2013 DM2E Review: Work Package 2 23
  • 24.
    JudaicaLink (Dov, Kai) •Permission to create and publish Linked Data from: – Russian Jewish Encyclopedia (rujen.ru) – YIVO Encyclopedia • UMA hired a student assistant to work with both encyclopedias and create Linked Data representations. • More encyclopediae would be great (Contact Dov) • Further information: JudaicaLink.org 11.06.2013 DM2E All WP Meeting: Work Package 2 24
  • 25.
    Deployment (Julia) 11.06.2013 DM2E AllWP Meeting: Work Package 2 25
  • 26.
    Berlin Server • Whatdo we have? – 2 Processors (Multicore) – 64 GB RAM – 4 TB Hard Drive (Usable space might be less, depends on RAID configuration) – That‘s all, we won‘t get more! • UBER and UMA will reinstall the server from scratch. • Virtual Machines: – Ontologies, OWLIM, OmNom, Silk, Pundit, Redmine, … • Downtimes in December, will be announced. 16.04.2013 DM2E Review: Work Package 2 26
  • 27.
    Next steps • FinalVersion (January 2014) – – – – Complete transformation and ingestion infrastructure Integrated contextualization Connection with scholarly environment (WP3) Documentation Drafts • Data ingestions(!) • Maintenance phase – – – – Bug fixing and performance tuning in OmNom Documentation and introductory materials with WP4 Data provision for Europeana Search and browse interface 16.04.2013 DM2E Review: Work Package 2 27
  • 28.

Editor's Notes

  • #10 Ein Beispiel für eine vom Reasoning erzeugte Inferenzbeziehung, die dem EDM-Modell konträr ist. Edm:ProvidedCHO darf nicht eine Subklasse von ore:Aggregation sein. Diese Beziehung kommt von der Defintion des dc:subject als Subproperties von edm:aggregatedCHO und die Domain restriction edm:ProvidedCHO. Lösung: in der aktuellen Version des Modells ist dc:subject, auch auf Wunsch der Provider, keine ObjectPropertie, sondern ein DataPropertie und hat somit als Domain Literalvalue. &lt;number&gt;
  • #11 Durch die logische Gleichung von dm2e:Manuscript == not (bibo:Manuscript) wird die Konsequenz gezogen, dass edm:InformationResource eine SubKlasse von bibo:Manuscript sei. Dies wäre falsch. Lösung: die unterschiedliche Definitionen von bibo:Manuscript (steht für Dokumententwurf und nicht für handschriftliche Manuscripte) und dm2e:Manuscript wird nicht mehr über owl:EquivalentTo ausgedrückt. Wir haben es nun rdfs:comment festgehalten und in owl:disjointWith-Propertie in der Ontologie verändert. &lt;number&gt;
  • #12 A snippet of the inferred classifications of the individuals. Bin noch dabei alle Inferenzen zu checken, und ggf. Die Ontology zu korrigieren &lt;number&gt;
  • #21 &lt;number&gt;