Edwin

A framework for building and analysing Equivalence Set Graphs.

Installation

Edwin can be installed using maven or docker.

Maven

$ git clone https://github.com/luigi-asprino/edwin.git
$ cd edwin
$ mvn clean install

Once installed, you can add Edwin as a dependency of your maven project.

<dependency>
  <groupId>it.cnr.istc.stlab</groupId>
  <artifactId>edwin</artifactId>
  <version>0.0.1</version>
</dependency>

Docker

• todo

Usage

You can compute an Equivalence Set Graph from Java as follows.

Edwin.computeESG("/path/to/configuration/file");

Configuration file

# MANDATORY: Path to RDF-HDT file to analyse
hdtFilePath=/Volumes/L2TB/LOD_a_lot/LOD_a_lot_v1.hdt

# MANDATORY: URI of the equivalence property to observe (cf. [1])
equivalencePropertyToObserve=http://www.w3.org/2002/07/owl#equivalentProperty

# MANDATORY: URI of the equivalence property to use for retrieving equivalence relations among properties  (cf. [1])
equivalencePropertyForProperties=http://www.w3.org/2002/07/owl#equivalentProperty

# OPTIONAL: A list of properties to not be considered as equivalence properties (cf. [1])
notEquivalenceProperties=http://www.w3.org/1999/02/22-rdf-syntax-ns#type,http://www.w3.org/2000/01/rdf-schema#subClassOf,http://www.w3.org/2002/07/owl#equivalentClass

# MANDATORY: URI of the specialization property to observe (cf. [1])
specializationPropertyToObserve=http://www.w3.org/2000/01/rdf-schema#subPropertyOf

# MANDATORY: URI of the specialization property to use for retrieving specialization relations among properties (cf. [1])
specializationPropertyForProperties=http://www.w3.org/2000/01/rdf-schema#subPropertyOf

# OPTIONAL: A list of properties to not be considered as specialization properties (cf. [1])
notSpecializationProperties=http://www.w3.org/1999/02/22-rdf-syntax-ns#type,http://www.w3.org/2000/01/rdf-schema#subClassOf,http://www.w3.org/2002/07/owl#equivalentClass

# OPTIONAL: A class that implements a set of criteria for selecting entities to observe
observedEntitiesSelector=it.cnr.istc.stlab.edwin.PropertiesSelector

# OPTIONAL: A class that implements a set of criteria for estimating the extensional size of the observed entities
extensionalSizeEstimator=it.cnr.istc.stlab.edwin.PropertySizeEstimator

# OPTIONAL: A precomputed ESG for properties and for classes (these ESGs might be useful for methods for selecting observed entities and estimate their extensional size)
esgPropertiesFolder=/path/to/ESG/for/properties
esgClassesFolder=/path/to/ESG/for/classes

# MANDATORY: The path where the ESG will be stored
esgFolder=/path/to/esgFolder

Output

In esgFolder you can find:

1. The computed ESG is stored in a number of key-value maps and can be loaded from Java as follows:

EquivalenceSetGraph esg = EquivalenceSetGraphLoader.loadEquivalenceSetGraphFromFolder("/path/to/esgFolder");

2. A TSV file (stats.tsv) that summarizes a set of statistics computed on the ESG.
3. Two TSV files (nodelist.tsv and edgelist.tsv) that allow to analyse the structural features of the graph using SNAP. An analysis of the connected components of the ESG is performed by the script src/main/python/compute_graph_stats.py. This script takes as input the folder where the nodelist.tsv and edgelist.tsv are stored (which is /path/to/esgFolder by default) and the path where to store the connected components of the graph.
4. The triplification of the ESG following the Edwin Ontology (cf. [2]).

Computed Equivalence Set Graphs

A number of Computed Equivalence Set Graphs are available for download from this page.

Publications

For citing Edwin in academic papers please use:

[1] Luigi Asprino, Wouter Beek, Paolo Ciancarini, Frank van Harmelen and Valentina Presutti. Observing LOD using Equivalent Set Graphs: it is mostly flat and sparsely linked. In: Proceedings of the 18th International Semantic Web Conference (ISWC 2019). (to appear). Preprint

[2] Luigi Asprino, Wouter Beek, Paolo Ciancarini, Frank van Harmelen and Valentina Presutti. Triplifying Equivalence Set Graphs. In: Proceedings of the ISWC 2019 Posters & Demonstrations, Industry, and Outrageous Ideas Tracks co-located with 18th International Semantic Web Conference (ISWC 2019)

License

Apache 2.0

This framework is named in honor of the American astronomer Edwin Powell Hubble