Towards maintainable constraint validation and repair for taxonomies: The PoolParty approach

http://aligned-project.eu COLD@ISWC, 18th October 2016, Kobe
Towards maintainable constraint validation
and repair for taxonomies
- The PoolParty approach
Monika Solanki
https://w3id.org/people/msolanki
@nimonika
University of Oxford
Joint work with
Christian Mader
Fraunhofer IAIS, Germany

PoolParty (SWC) Use case
PoolParty(PPT): leading commercial taxonomy
management application, authoring tool for knowledge
graphs, provides taxonomy import functionality to
interact with third party datasets
Taxonomists using PPT integrate a variety of models,
schemata, ontologies and vocabularies into their
knowledge bases.
Challenge: combining varied data sources to ensure that
these data mashups at any time conform to a set of quality
heuristics, as expected by the data processing algorithms.
monika.solanki@cs.ox.ac.uk, @nimonika Constraint validation and repair for taxonomies

Motivation
Consuming and interlinking enterprise data and openly
available data within an industry setting.
Ensuring that the interlinked datasets conﬁrm to a set of
quality heuristics.
Interactively detecting and repairing datasets with
constraint violations.

Ensuring Data Consistency
Current - checks to ensure that the data persisted in the triple
store do not violate it’s data consistency are scattered in the
code and sometimes performed multiple times.
Requirements
Provide a mechanism to specify data constraints in a
formal way,
Identify and analyse datasets that are imported into PPT
and are a source of constraint violations.

Constraint resolution
Current - checks to ensure that the data persisted in the triple
store do not violate it’s data consistency are scattered in the
code and sometimes performed multiple times.
Requirements
Provide a validation mechanism to check for constraint
violation and evaluate this against the selected datasets.
Combine formal data constraint deﬁnitions with reusable
repair strategies that can be easily applied by end-users in
a (semi-) automatic way.

Dataset selection
SWC-generated: Datasets for which a conversion to a
PPT-compatible taxonomy has been performed by SWC
(containing 10 datasets),
Custom-generated: Datasets for which a conversion to a
PPT-compatible taxonomy has been performed by
third-party institutions (containing 9 datasets), and
Web: Datasets that are using SKOS, but for which is
currently unknown if they are compatible with PPT
(containing 7 datasets).

Constraint speciﬁcation
ConceptTypeAssertion (cta):
SELECT DISTINCT ?resource WHERE {
?resource skos:broader|skos:narrower ?otherRes.
FILTER NOT EXISTS {?resource a skos:Concept}}
HierarchicalConsistency (hc):
SELECT DISTINCT ?resource WHERE {
?resource a skos:Concept
FILTER NOT EXISTS {
?resource (skos:broader|^skos:narrower)*/skos:
topConceptOf ?parent.
?parent a skos:ConceptScheme.}}

Validation using SHACL
HierarchicalConsistency (hc):
ppts:ConceptShape
a sh:Shape;
sh:scopeClass skos:Concept;
sh:property [
a sh:PropertyConstraint;
sh:predicate skos:prefLabel;
sh:minCount 1;
sh:minLength 1;
sh:datatype rdf:langString;
sh:uniqueLang true];
sh:constraint [
a sh:Constraint;
a sh:OrConstraint;
sh:shapes (ppts:ConceptHasBroaderShape ppts:
ConceptIsTopConceptShape)].

Repair strategies
AddInverseStrategy
ppts:ConceptHavingBroader
a sh:Shape;
sh:scope [
a sh:Scope;
a sh:PropertyScope ;
sh:predicate skos:broader];
sh:inverseProperty [
a sh:InversePropertyConstraint;
sh:predicate skos:narrower;
sh:minCount 1;
rs:strategy [
a rs:AddInverseStrategy]].

Implementation
SHACL implementation (TopQuadrant), Sesame, SWC
libraries ⇒ Java application
SKOS data model, Dataset file, Constraint specification ⇒
Violation report
Violation report, SKOS data model, Dataset file, Constraint
specification ⇒ Triples changeset
Not yet Optimised for runtime performance

Validation results
cta was never violated in datasets converted to PPT
taxonomies.
upl is a SKOS-level constraint, better respected by
vocabulary providers.
Violations observed across all datasets.

Validation performance
Omitted 10 datasets that contained ≤ 50000 triples.
No correlation between the dataset size and time taken to
perform the validation.
Structure of the dataset makes a difference.

Repair strategy execution performance
Repair strategy applied to a special case of the constraint
br - BidirectionalRelationsHierarical.
Only considered skos:broaderThan and
skos:narrowerThan. Did not consider owl:inverse.
Repair scales well even with larger datasets.

Summary and Conclusions
Interwoven SHACL-based data consistency specification
and validation with repair strategies.
Validation of datasets generated by PPT can be done with
reasonable performance.
Integrating repair strategies and data constraint
specification helps in building a unified, maintainable
model.
The model also plays a pivotal role in harmonizing data
and software development processes.

Towards maintainable constraint validation and repair for taxonomies: The PoolParty approach

More Related Content

What's hot (20)

Viewers also liked (17)

More from Monika Solanki (16)

Recently uploaded (20)

Towards maintainable constraint validation and repair for taxonomies: The PoolParty approach