Open Semantic Framework 3.2 Released

Structured Dynamics is happy to announce the immediate availability of the Open Semantic Framework version 3.2. This is the second important OSF release in a month and a half. triple_120

This new major release of OSF changes the way the web services communicate with the triple store. Originally, OSF web services were using a ODBC channel to communicate with the triple store (Virtuoso). This new release uses the SPARQL HTTP endpoints of the triple store to send queries to it. This is the only changes that occurs in this new version, but as you will see bellow, this is a major one.

Why switching to HTTP?

The problem with using ODBC as the primary communication channel between the OSF web services and the triple store is that it was adding a lot of complexity into OSF. Because the UnixODBC drivers that are shipped with Ubuntu had issues with Virtuoso, we had to use the iODBC drivers to make sure that everything was working properly. This situation forced us to recompile PHP5 such that it uses iODBC instead of UnixODBC as the ODBC drivers for PHP5.

This was greatly complexifying the deployment of OSF since we couldn’t use the default PHP5 packages that shipped with Ubuntu, but had to maintain our own ones that were working with iODBC.

The side effect of this is that system administrators couldn’t upgrade their Ubuntu instances normally since PHP5 needed to be upgraded using particular packages created for that purpose.

Now that OSF doesn’t use ODBC to communicate with the triple store, all this complexity goes away since no special handling is now required. All of the default Ubuntu packages can be used like system administrators normally do.

With this new version, the installation and deployment of a OSF instance has been greatly simplified.

Supports New Triple Stores

Another problem with using ODBC is that it was limiting the number of different triple stores that could be used for operating OSF. In fact, people could only use Virtuoso with their OSF instance.

This new release opens new opportunities. OSF still ships with Virtuoso Open Source as its default triple store, however any triple store that has the following characteristics could replace Virtuoso in OSF:

  1. It has a SPARQL HTTP endpoint
  2. It supports SPARQL 1.1 and SPARQL Update 1.1
  3. It supports SPARQL Update queries that can be sent to the SPARQL HTTP endpoint
  4. It supports the SPARQL 1.1 Query Results JSON Format
  5. It supports the SPARQL 1.1 Graph Store HTTP Protocol via a HTTP endpoint (optional, only required by the Datasets Management Tool)

Deploying a new OSF 3.2 Server

Using the OSF Installer

OSF 3.2 can easily be deployed on a Ubuntu 14.04 LTS server using the osf-installer application. It can easily be done by executing the following commands in your terminal:

[cc lang=”bash”]
[raw]
mkdir -p /usr/share/osf-installer/

cd /usr/share/osf-installer/

wget https://raw.github.com/structureddynamics/Open-Semantic-Framework-Installer/3.2/install.sh

chmod 755 install.sh

./install.sh

./osf-installer –install-osf -v
[/raw]
[/cc]

Using a Amazon AMI

If you are an Amazon AWS user, you also have access to a free AMI that you can use to create your own OSF instance. The full documentation for using the OSF AMI is available here.

Upgrading Existing Installations

It is not possible to automatically upgrade previous versions of OSF to OSF 3.2. It is possible to upgrade a older instance of OSF to OSF version 3.2, but only manually. If you have this requirement, just let me know and I will write about the upgrade steps that are required to upgrade these instances to OSF version 3.2.

Security

Now that the triple store’s SPARQL HTTP endpoint requires it to be enabled with SPARQL Update rights, it is more important than ever to make sure that the SPARQL HTTP endpoint of the triple store is only available to the OSF web services.

This can be done by properly configuring your firewall or proxy such that only local traffic, or traffic coming from the OSF web service processes, can reach the endpoint.

The SPARQL endpoint that should be exposed to the outside World is OSF’s SPARQL endpoint, which adds an authentication layer above the triple store’s endpoint, and restricts potentially armful SPARQL queries.

Conclusion

This new version of the Open Semantic Framework greatly simplifies its deployment and its maintenance. It also enables other triple stores that exist on the market to be used for OSF instead of Virtuoso Open Source.

Open Semantic Framework 3.1 Released

Structured Dynamics is happy to announce the immediate availability of the Open Semantic Framework version 3.1. This new version includes a set of fixes to different components of the framework in the last few months. The biggest change is deployment of OSF using Virtuoso Open Source version 7.1.0. triple_120

We also created a new API for Clojure developers called: clj-osf. Finally we created a new Open Semantic Framework web portal that better describes the project and is hopefully easier to use and more modern.

Quick Introduction to the Open Semantic Framework

What is the Open Semantic Framework?

The Open Semantic Framework (OSF) is an integrated software stack using semantic technologies for knowledge management. It has a layered architecture that combines existing open source software with additional open source components. OSF is designed as an integrated content platform accessible via the Web, which provides needed knowledge management capabilities to enterprises. OSF is made available under the Apache 2 license.

OSF can integrate and manage all types of content – unstructured documents, semi-structured files, spreadsheets, and structured databases – using a variety of best-of-breed data indexing and management engines. All external content is converted to the canonical RDF data model, enabling common tools and methods for tagging and managing all content. Ontologies provide the schema and common vocabularies for integrating across diverse datasets. These capabilities can be layered over existing information assets for unprecedented levels of integration and connectivity. All information within OSF may be powerfully searched and faceted, with results datasets available for export in a variety of formats and as linked data.

A new Open Semantic Framework website

The OSF 3.1 release also triggered the creation of a new website for the project. We wanted something leaner and more modern and that is what I think we delivered. We also reworked the content, we wrote about a series of usecases 1 2 3 4 5 6 and we better aggregated and presented information for each web service endpoint.

A new OSF sandbox

We also created an OSF sandbox where people can test each web service endpoint and test how each functionality works. All of the web services are open to users. The sandbox is not meant to be stable considering that everybody have access to all endpoints. However, the sandbox server will be recreated on a periodic basis. If the sandbox is totally broken and users experiment issues, they can always request a re-creation of the server directly on the OSF mailing list.

Each of the web service pages on the new OSF portal has a Sandbox section where you see some code examples of how to use the endpoint and how to send requests to the sandbox. Here are the instructions to use the sandbox server.

A new OSF API for Clojure: clj-osf

The OSF release 3.1 also includes a new API for Clojure developers: clj-osf.

clj-osf is a Domain Specific Language (DSL) that should lower the threshold to use the Open Semantic Framework.

To use the DSL, you only have to configure your application to use a specific OSF endpoint. Here is an example of how to do this for the Sandbox server:

[cc lang=”lisp”]
[raw]
;; Define the OSF Sandbox credentials (or your own):
(require ‘[clj-osf.core :as osf])

(osf/defosf osf-test-endpoint {:protocol :http
:domain “sandbox.opensemanticframework.org”
:api-key “EDC33DA4D977CFDF7B90545565E07324”
:app-id “administer”})

(osf/defuser osf-test-user {:uri “http://sandbox.opensemanticframework.org/wsf/users/admin”})
[/raw]
[/cc]

Then you can send simple OSF web service queries. Here is an example that sends a search query to return records of type foaf:Person that also match the keyword “bob”:

[cc lang=”lisp”]
[raw]
(require ‘[clj-osf.search :as search])

(search/search
(search/query “bob”)
(search/type-filters [“http://xmlns.com/foaf/0.1/Person”]))
[/raw]
[/cc]

A complete set of clj-osf examples is available on the OSF wiki.

Finally the complete clj-osf DSL documentation is available here.

A community effort

This new release of the OSF Installer is another effort of the growing Open Semantic Framework community. The upgrade of the installer to deploy the OSF stack using Virtuoso Open Source version 7.1.0 has been created by William (Bill) Anderson.

Deploying a new OSF 3.1 Server

Using the OSF Installer

OSF 3.1 can easily be deployed on a Ubuntu 14.04 LTS server using the osf-installer application. It can easily be done by executing the following commands in your terminal:

[cc lang=”bash”]
[raw]
mkdir -p /usr/share/osf-installer/

cd /usr/share/osf-installer/

wget https://raw.github.com/structureddynamics/Open-Semantic-Framework-Installer/3.1/install.sh

chmod 755 install.sh

./install.sh

./osf-installer –install-osf -v
[/raw]
[/cc]

Using a Amazon AMI

If you are an Amazon AWS user, you also have access to a free AMI that you can use to create your own OSF instance. The full documentation for using the OSF AMI is available here.

Upgrading Existing Installations

Existing OSF installations can be upgraded using the OSF Installer. However, note that the upgrade won’t deploy Virtuoso Open Source 7.1.0 for you. All the code will be upgraded, but Virtuoso will remain the version you were last using on your instance. All the code of OSF 3.1 is compatible with previous versions of Virtuoso, but you won’t benefit the latest improvements to Virtuoso (in terms of performances) and its latest SPARQL 1.1 implementations. If you want to upgrade Virtuoso to version 7.1.0 on an existing OSF instance you will have to do this by hands.

To upgrade the OSF codebase, the first thing is to upgrade the installer itself:

[cc lang=”bash”]
[raw]
# Upgrade the OSF Installer
./usr/share/osf-installer/upgrade.sh
[/raw]
[/cc]

Then you can upgrade the components using the following commands:

[cc lang=”bash”]
[raw]
# Upgrade the OSF Web Services
./usr/share/osf-installer/osf –upgrade-osf-web-services=”3.1.0″

# Upgrade the OSF WS PHP API
./usr/share/osf-installer/osf –upgrade-osf-ws-php-api=”3.1.0″

# Upgrade the OSF Tests Suites
./usr/share/osf-installer/osf –upgrade-osf-tests-suites=”3.1.0″

# Upgrade the Datasets Management Tool
./usr/share/osf-installer/osf –upgrade-osf-datasets-management-tool=”3.1.0″

# Upgrade the Data Validator Tool
./usr/share/osf-installer/osf –upgrade-osf-data-validator-tool=”3.1.0″
[/raw]
[/cc]

clj-turtle: A Clojure Domain Specific Language (DSL) for RDF/Turtle

Some of my recent work leaded me to heavily use Clojure to develop all kind of new capabilities for Structured Dynamics. The ones that knows us, knows that every we do is related to RDF and OWL ontologies. All this work with Clojure is no exception.

Recently, while developing a Domain Specific Language (DSL) for using the Open Semantic Framework (OSF) web service endpoints, I did some research to try to find some kind of simple Clojure DSL that I could use to generate RDF data (in any well-known serialization). After some time, I figured out that no such a thing was currently existing in the Clojure ecosystem, so I choose to create my simple DSL for creating RDF data.

The primary goal of this new project was to have a DSL that users could use to created RDF data that could be feed to the OSF web services endpoints such as the CRUD: Create or CRUD: Update endpoints.

What I choose to do is to create a new project called clj-turtle that generates RDF/Turtle code from Clojure code. The Turtle code that is produced by this DSL is currently quite verbose. This means that all the URIs are extended, that the triple quotes are used and that the triples are fully described.

This new DSL is mean to be a really simple and easy way to create RDF data. It could even be used by non-Clojure coder to create RDF/Turtle compatible data using the DSL. New services could easily be created that takes the DSL code as input and output the RDF/Turtle code. That way, no Clojure environment would be required to use the DSL for generating RDF data.

Installation

For people used to Clojure and Leinengen, you can easily install clj-turtle using Linengen. The only thing you have to do is to add Add [clj-turtle "0.1.3"] as a dependency to your project.clj.

Then make sure that you downloaded this dependency by running the lein deps command.

API

The whole DSL is composed of simply six operators:

  • rdf/turtle
    • Used to generate RDF/Turtle serialized data from a set of triples defined by clj-turtle.
  • defns
    • Used to create/instantiate a new namespace that can be used to create the clj-turtle triples
  • rei
    • Used to reify a clj-turtle triple
  • iri
    • Used to refer to a URI where you provide the full URI as an input string
  • literal
    • Used to refer to a literal value
  • a
    • Used to specify the rdf:type of an entity being described

Usage

Working with namespaces

The core of this DSL is the defns operator. What this operator does is to give you the possibility to create the namespaces you want to use to describe your data. Every time you use a namespace, it will generate a URI reference in the triple(s) that will be serialized in Turtle.

However, it is not necessary to create a new namespace every time you want to serialize Turtle data. In some cases, you may not even know what the namespace is since you have the full URI in hands already. It is why there is the iri function that let you serialize a full URI without having to use a namespace.

Namespaces are just shorthand versions of full URIs that mean to make your code cleaner an easier to read and maintain.

Syntactic rules

Here are the general syntactic rules that you have to follow when creating triples in a (rdf) or (turtle) statement:

  1. Wrap all the code using the (rdf) or the (turtle) operator
  2. Every triple need to be explicit. This means that every time you want to create a new triple, you have to mention the subject, predicate and the object of the triple
  3. A fourth “reification” element can be added using the rei operator
  4. The first parameter of any function can be any kind of value: keywords, strings, integer, double, etc. They will be properly serialized as strings in Turtle.

Strings and keywords

As specified in the syntactic rules, at any time, you can use a string, a integer, a double a keyword or any other kind of value as input of the defined namespaces or the other API calls. You only have to use the way that is more convenient for you or that is the cleanest for your taste.

More about reification

Note: RDF reification is quite a different concept than Clojure’s reify macro. So carefully read this section to understand the meaning of the concept in this context.

In RDF, reifying a triple means that we want to add additional information about a specific triple. Let’s take this example:

[cc lang=”lisp”]
[raw]
(rdf
(foo :bar) (iron :prefLabel) (literal “Bar”))
[/raw]
[/cc]

In this example, we have a triple that specify the preferred label of the :bar entity. Now, let’s say that we want to add “meta” information about that specific triple, like the date when this triple got added to the system for example.

That additional information is considered the “fourth” element of a triple. It is defined like this:

[cc lang=”lisp”][raw]
(rdf
(foo :bar) (iron :prefLabel) (literal “Bar”) (rei
(foo :date) (literal “2014-10-25” :type :xsd:dateTime)))
[/raw][/cc]

What we do here is to specify additional information regarding the triple itself. In this case, it is the date when the triple got added into our system.

So, reification statements are really “meta information” about triples. Also not that reification statements doesn’t change the semantic of the description of the entities.

Examples

Here is a list of examples of how this DSL can be used to generate RDF/Turtle data:

Create a new namespace

The first thing we have to do is define the namespaces we will want to use in our code.

[cc lang=”lisp”][raw]
(defns iron “http://purl.org/ontology/iron#”)
(defns foo “http://purl.org/ontology/foo#”)
(defns owl “http://www.w3.org/2002/07/owl#”)
(defns rdf “http://www.w3.org/1999/02/22-rdf-syntax-ns#”)
(defns xsd “http://www.w3.org/2001/XMLSchema#”)
[/raw][/cc]

Create a simple triple

The simplest example is to create a simple triple. What this triple does is to define the preferred label of a :bar entity:

[cc lang=”lisp”][raw]
(rdf
(foo :bar) (iron :prefLabel) (literal “Bar”))
[/raw][/cc]

Output:

[cce]<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#prefLabel> “””Bar””” .[/cce]

Create a series of triples

This example shows how we can describe more than one attribute for our bar entity:

[cc lang=”lisp”][raw]
(rdf
(foo :bar) (a) (owl :Thing)
(foo :bar) (iron :prefLabel) (literal “Bar”)
(foo :bar) (iron :altLabel) (literal “Foo”))
[/raw][/cc]

Output:

[cce]<http://purl.org/ontology/foo#bar> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2002/07/owl#Thing> .
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#prefLabel> “””Bar””” .
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#altLabel> “””Foo””” .
[/cce]

Note: we prefer having one triple per line. However, it is possible to have all the triples in one line, but this will produce less readable code:

Just use keywords

It is possible to use keywords everywhere, even in (literals)

[cc lang=”lisp”][raw]
(rdf
(foo :bar) (a) (owl :Thing)
(foo :bar) (iron :prefLabel) (literal :Bar)
(foo :bar) (iron :altLabel) (literal :Foo))
[/raw][/cc]

Output:

[cce]
<http://purl.org/ontology/foo#bar> <http://www.w3.org/1999/02/22-rdf-syntax-ns#> <http://www.w3.org/2002/07/owl#Thing> .
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#prefLabel> “””Bar””” .
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#altLabel> “””Foo””” .
[/cce]

Just use strings

It is possible to use strings everywhere, even in namespaces:

[cc lang=”lisp”][raw]
(rdf
(foo “bar”) (a) (owl “Thing”)
(foo “bar”) (iron :prefLabel) (literal “Bar”)
(foo “bar”) (iron :altLabel) (literal “Foo”))
[/raw][/cc]

Output:

[cce]
<http://purl.org/ontology/foo#bar> <http://www.w3.org/1999/02/22-rdf-syntax-ns#> <http://www.w3.org/2002/07/owl#Thing> .
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#prefLabel> “””Bar””” .
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#altLabel> “””Foo””” .
[/cce]

Specifying a datatype in a literal

It is possible to specify a datatype for every (literal) you are defining. You only have to use the :type option and to specify a XSD datatype as value:

[cc lang=”lisp”][raw]
(rdf
(foo “bar”) (foo :probability) (literal 0.03 :type :xsd:double))
[/raw][/cc]

Equivalent codes are:

[cc lang=”lisp”][raw]
(rdf
(foo “bar”) (foo :probability) (literal 0.03 :type (xsd :double)))
(rdf
(foo “bar”) (foo :probability) (literal 0.03 :type (iri “http://www.w3.org/2001/XMLSchema#double”)))
[/raw][/cc]

Output:

[cce]
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/foo#probability> “””0.03″””^^xsd:double .
[/cce]

Specifying a language for a literal

It is possible to specify a language string using the :lang option. The language tag should be a compatible ISO 639-1 language tag.

[cc lang=”lisp”][raw]
(rdf
(foo “bar”) (iron :prefLabel) (literal “Robert” :lang :fr))
[/raw][/cc]

Output:

[cce]
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#prefLabel> “””Robert”””@fr .
[/cce]

Defining a type using the an operator

It is possible to use the (a) predicate as a shortcut to define the rdf:type of an entity:

[cc lang=”lisp”][raw]
(rdf
(foo “bar”) (a) (owl “Thing”))
[/raw][/cc]

Output:

[cce]
<http://purl.org/ontology/foo#bar> <http://www.w3.org/1999/02/22-rdf-syntax-ns#> <http://www.w3.org/2002/07/owl#Thing> .
[/cce]

This is a shortcut for:

[cc lang=”lisp”][raw]
(rdf
(foo “bar”) (rdf :type) (owl “Thing”))
[/raw][/cc]

Specifying a full URI using the iri operator

It is possible to define a subject, a predicate or an object using the (iri) operator if you want to defined them using the full URI of the entity:

[cc lang=”lisp”][raw]
(rdf
(iri “http://purl.org/ontology/foo#bar”) (iri “http://www.w3.org/1999/02/22-rdf-syntax-ns#type) (iri http://www.w3.org/2002/07/owl#Type))
[/raw][/cc]

Output:

[cce]
<http://purl.org/ontology/foo#bar> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/2002/07/owl#Type> .
[/cce]

Simple reification

It is possible to reify any triple suing the (rei) operator as the fourth element of a triple:

[cc lang=”lisp”][raw]
(rdf
(foo :bar) (iron :prefLabel) (literal “Bar”) (rei
(foo :date) (literal “2014-10-25” :type :xsd:dateTime)))
[/raw][/cc]

Output:

[cce]
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#prefLabel> “””Bar””” .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/1999/02/22-rdf-syntax-ns#Statement> .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#subject> <http://purl.org/ontology/foo#bar> .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#predicate> <http://purl.org/ontology/iron#prefLabel> .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#object> “””Bar””” .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://purl.org/ontology/foo#date> “””2014-10-25″””^^xsd:dateTime .
[/cce]

Reify multiple properties

It is possible to add multiple reification statements:

[cc lang=”lisp”][raw]
(rdf
(foo :bar) (iron :prefLabel) (literal “Bar”) (rei
(foo :date) (literal “2014-10-25” :type :xsd:dateTime)
(foo :bar) (literal 0.37)))
[/raw][/cc]

Output:

[cce]
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#prefLabel> “””Bar””” .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#type> <http://www.w3.org/1999/02/22-rdf-syntax-ns#Statement> .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#subject> <http://purl.org/ontology/foo#bar> .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#predicate> <http://purl.org/ontology/iron#prefLabel> .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://www.w3.org/1999/02/22-rdf-syntax-ns#object> “””Bar””” .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://purl.org/ontology/foo#date> “””2014-10-25″””^^xsd:dateTime .
<rei:6930a1f93513367e174886cb7f7f74b7> <http://purl.org/ontology/foo#bar> “””0.37″”” .
[/cce]

Using clj-turtle with clj-osf

clj-turtle is meant to be used in Clojure code to simplify the creation of RDF data. Here is an example of how clj-turtle can be used to generate RDF data to feed to the OSF Crud: Create web service endpoint via the clj-osf DSL:

[cc lang=”lisp”][raw]
[require ‘[clj-osf.crud :as crud])

(crud/create
(crud/dataset “http://test.com/datasets/foo”)
(crud/document
(rdf
(iri link) (a) (bibo :Article)
(iri link) (iron :prefLabel) (literal “Some article”)))
(crud/is-rdf-n3)
(crud/full-indexation-mode))
[/raw][/cc]

Using the turtle alias operator

Depending on your taste, it is possible to use the (turtle) operator instead of the (rdf) one to generate the RDF/Turtle code:

[cc lang=”lisp”][raw]
(turtle
(foo “bar”) (iron :prefLabel) (literal “Robert” :lang :fr))
[/raw][/cc]

Output:

[cce]
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#prefLabel> “””Robert”””@fr .
[/cce]

Merging clj-turtle

Depending the work you have to do in your Clojure application, you may have to generate the Turtle data using a more complex flow of operations. However, this is not an issue for clj-turtle since the only thing you have to do is to concatenate the triples you are creating. You can do so using a simple call to the str function, or you can create more complex processing using loopings, mappings, etc that end up with a (apply str) to generate the final Turtle string.

[cc lang=”lisp”][raw]
(str
(rdf
(foo “bar”) (a) (owl “Thing”))
(rdf
(foo “bar”) (iron :prefLabel) (literal “Bar”)
(foo “bar”) (iron :altLabel) (literal “Foo”)))
[/raw][/cc]

Output:

[cce]
<http://purl.org/ontology/foo#bar> <http://www.w3.org/1999/02/22-rdf-syntax-ns#> <http://www.w3.org/2002/07/owl#Thing> .
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#prefLabel> “””Bar””” .
<http://purl.org/ontology/foo#bar> <http://purl.org/ontology/iron#altLabel> “””Foo””” .
[/cce]

Conclusion

As you can see now, this is a really simple DSL for generating RDF/Turtle code. Even if simple, I find it quite effective by its simplicity. However, even if it quite simple and has a minimum number of operators, this is flexible enough to be able to describe any kind of RDF data. Also, thanks to Clojure, it is also possible to write all kind of code that would generate DSL code that would be executed to generate the RDF data. For example, we can easily create some code that iterates a collection to produce one triple per item of the collection like this:

[cc lang=”lisp”][raw]
(->> {:label “a”
:label “b”
:label “c”}
(map
(fn [label]
(rdf
(foo :bar) (iron :prefLabel) (literal label))))
(apply str))
[/raw][/cc]

That code would generate 3 triples (or more if the input collection is bigger). Starting with this simple example, we can see how much more complex processes can leverage clj-turtle for generating RDF data.

A future enhancement to this DSL would be to add a syntactic rule that gives the opportunity to the user to only have to specify the suject of a triple the first time it is introduced to mimic the semi-colon of the Turtle syntax.

Major UMBEL Release: 1.10

After more than 2 years, we are now finally releasing a new version of the UMBEL ontology and reference concept structure. One might think that we haven’t worked on the project all that time, but it is not strictly true. umbel_logo_260_160

We did improve the mapping to external vocabularies/ontologies, we worked much on linking Wikipedia pages to the UMBEL structure, but we haven’t had time to release a new version… until now!

For people new to the ontology, UMBEL is a general reference structure of about 28,000 reference concepts, which provides a scaffolding to link and interoperate other datasets and domain vocabularies. Its main purpose is to have a coherent conceptual structure that we can use to link and interoperate unrelated data sources. But it can also be used as a conceptual structure to be used to describe information like any other ontologies.

What is new with the ontology?

The major change in UMBEL is not the structure itself, but the piece of software used to generate it. In fact, the previous system we developed for generating UMBEL was about 7 years old. It was a bit clunky and really not that easy to work with.

Based on our prior experience with UMBEL, we choose to dump it and to create a brand new UMBEL reference structure generator. This new generator has been developed in Clojure and uses the latest version of the OWL API. It makes the management of the structure much simpler, which means that it will help in releasing new UMBEL version more regularly. We also have a suite of tools to analyze the structure and to pinpoint possible issues.

Other than that, we updated the Schema.org, DBpedia Ontology and Geonames Ontology mappings to UMBEL. This is a major effort undertaken by Mike for this new version. The mappings are composed of:

  • 754 rdfs:subClassOf relationships between Schema.org classes and UMBEL reference concepts
  • 688 rdfs:subClassOf relationships between DBpedia Ontology classes and UMBEL reference concepts
  • 682 rdfs:subClassOf relationships between Geonames Ontology classes and UMBEL reference concepts

These new mappings will help manage data instances that use these external ontologies/schemas in a broader conceptual structure (which is UMBEL). This enables us to be able to reason over this external data using the UMBEL conceptual structure even if these external data sources didn’t originally use UMBEL to describe their data. That is one of the main features of UMBEL.

We also managed to add a few hundred UMBEL reference concepts. Most of them were added to create these new linkages with the external ontologies. Others have been added because they were improving the overall structure.

A few weeks back, we found an issue with the umbel:superClassOf assignations, which has also now been resolved in version 1.10.

In the previous versions of UMBEL, the preferred labels were not unique. There were a few hundred of the concepts that were having the same preferred labels. This was not an issue in itself, but this was not a best practice to create an ontology. We managed to remove all these non-distinct preferred labels and to make all of them unique.

We added a few skos:broader and skos:narrower relationships between some of the reference concepts. In the previous versions, all the relationships were skos:broaderTransitive and skos:narrowerTransitive properties only.

Finally we made sure that the entire UMBEL reference structure (Core + the Geo module) was absent of any inconsistencies and that it was satisfiable.

What is new with the portal and web services?

This new version of UMBEL also led us to create a few new features to the UMBEL website. The most apparent feature is the new External Linkage section that may appear at the top of a reference concept page (obviously, it will not appear if there are no external links for a given reference concept). This section shows you the linkage between the UMBEL reference concept and other external classes:

umbel_linkage_module

Another feature that you will notice on this screenshot is the Core blue tag at the right of the URI of the reference concept. This tag is used to tell you from where the reference concept is coming. Another tag that you may encounter is the green Geo tag, which tells you that the reference concept comes from the UMBEL Geo module. The same tags appear in the search resultsets:

search_modules

What is next?

Because UMBEL is an ontology, by nature it will always evolve over time. Things change, and the way we see the World can always improve.

For the next version of UMBEL, we will analyze the entire UMBEL reference concept structure using different algorithms, heuristics and other techniques to analyze the conceptual structure and to find conceptual gaps in it. The goal of this analysis is to tighten the structure, to have a better conceptual hierarchy and a more fine-grained one.

Other things we want to do in other coming versions are to improve the Super Types structure of UMBEL. As you may know, many of the Super Types are non disjoint because some of the concepts belong to multiple Super Type classes. What we want to do here is to create new Super Types classes that are the intersection between two, or more, Super Types that will be used to categorize these concepts that belong to multiple Super Types. That way, we will end-up with a better classification of the UMBEL reference concepts from a Super Types standpoint.

Another thing we want to do related to the UMBEL web services is to update them such that you can query the linkage to the external ontologies. For now, you can see the linkage when querying the sub-classes and super-classes of a reference concept. But you cannot query the web services this way: give me all the sub-classes-of the http://schema.org/FireStation class, for example.

As you can see, the UMBEL ontology and web services will continue to evolve over time to enable new ways to leverage the conceptual structure and external data sources.

Graph Analysis of a Big Structure: UMBEL

In this blog post, I will show you how we can use basic graph analysis metrics to analyze Big Structures. I will do the analysis of the UMBEL reference concept structure to demonstration how graph and network analysis can be used to better understand the nature and organization of possible issues with Big Structures such as UMBEL. umbel.graph

I will first present the formulas that have been used to perform the UMBEL analysis. To better understand this section, you will need some math knowledge, but there is nothing daunting here. You can probably safely skip that section if you desire.

The second section is the actual analysis of a Big Structure (UMBEL) using the graph measures I presented in the initial section. The results will be presented, analyzed and explained.

After reading this blog post, you should better understand how graph and network analysis techniques can be used to understand, use and leverage Big Structures to help integrate and interoperate disparate data sources.

[extoc]

Big Structure: Big in Size

A Big Structure is a network (a graph) of inter-related concepts which is composed of thousands or even hundred of thousands of such concepts. One characteristic of a Big Structure is its size. By nature, a Big Structure is too big to manipulate by hand and it requires tools and techniques to understand and assess the nature and the quality of the structure. It is for this reason that we have to leverage graph and network measures to help us in manipulating these Big Structures.

In the case of UMBEL, the Big Structure is a scaffolding of reference concepts used to link external (unrelated) structures to help data integration and to help unrelated systems inter-operate. In a World where the Internet Of Things is the focus of big companies and where there are more than 400 standards, such techniques and technologies are increasingly important, otherwise it will end-up being the Internet of [Individual] Things. Such a Big Structure can also be used for other tasks such as helping machine learning techniques to categorize and disambiguate pieces of data by leveraging such a structure of types.

UMBEL as a Graph

UMBEL is an RDF and OWL ontology of a bit more than 26 000 reference concepts. Because the structure is represented using RDF, it means that it is a directed graph. All of UMBEL’s vertices are classes, and all of the edges are properties.

The most important fact to keep in mind until the end of this blog post is that we are manipulating a directed graph. This means that all of the formulas used to analyze the UMBEL graph are formulas applicable to directed graphs only.

I will keep the normal graph analysis language that we use in the literature, however keep in mind that a vertice is a class or a named individual and that a edge is a property.

The UMBEL structure we are using is composed of the classes view and the individuals view of the ontology. That means that all the concepts are there, where some of them only have a class view, others have an individual view and others have both (because they got punned).

Graph Analysis Metrics

In this section, I will present the graph measures that we will use to perform the initial UMBEL graph analysis. In the next section, we will make the analysis of UMBEL using these measures, and we will discuss the results.

This section uses math notations. It could be skipped, but I suggest to try to take time some time to understand each measure since it will help to understand the analysis.

Some Notations

In this blog post, a graph is represented as G = (V,E) where G is the graph, V is the set of all the vertices and E is the set of all the edges of the same type that relates vertices.

The UMBEL analysis focuses on one of the following transitive properties:  rdfs:subClassOf, umbel:superClassOf, skos:broaderTransitive, skos:narrowerTransitive and rdf:type. When we do perform the analysis, we are picking-up a subgraph that is composed of all the connections between the vertices that are linked by this  edge (property).

Density

The first basic measure is the density of a graph. The density measures how many edges are in set E compared to the maximum possible number of edges between vertices in set V. The density is measured with:


 D = \frac{\raisebox{1.5mm}{$\vert E \vert$}}{\raisebox{-1.5mm}{$\vert V \vert ( \vert V \vert \mathbin{-} 1 )$}}
 

where D is the density, E is the number of properties (edges) and V is the number of classes (vertices).

The density is a ratio of the number of edges that exists, and the number of edges that could exists in the graph. \vert E \vert is the number of edges in the graph and \vert V \vert ( \vert V \vert \mathbin{-} 1 ) is the number of possible maximum number of edges.

The maximum density is 1, and the minimum density is 0. The density of a graph gives us an idea about the number of connections that exists between the vertices.

Average Degree

The degree of a vertex is the number of edges that connect that vertex to other vertices. The average degree of a graph G is another measure of how many edges are in set E compared to number of vertices in set V.


\bar D = \frac{\raisebox{1.5mm}{$\vert E \vert$}}{\raisebox{-1.5mm}{$\vert V \vert$}}
 

where \bar D is the average degree, E is the number of properties (edges) and V is the number of classes (vertices).

This measure tells the average number of nodes to which any given node is connected.

Diameter

The diameter of a graph G is the longest shortest path between two vertices in the graph. This means that this is the longest path that excludes all detours, loops, etc. between two vertices.

Let d(v_{i},v_{j}) be the length of the shortest path between v_{i} and v_{j}. And v_{i}, v_{j} in V. The diameter of the graph is defined as:


d = max(d(v_{i},v_{j}))
 

This metric gives us an assessment of the size of the graph. It is useful to understand the kind of graph we are playing with. We will also relate it with the average path length to assess the span of the graph and the distribution of path lengths.

Average Path Length

The average path length is the average of the shortest path length, averaged over all pairs of vertices. Let d(v_{i},v_{j}) be the length of the shortest path between v_{i} and v_{j}. And v_{i}, v_{j} in V.


\bar P =\frac{\raisebox{1.5mm}{$ 1 $}}{\raisebox{-1.5mm}{$ N \cdot (N \mathbin{-} 1) $}} \cdot \sum_{i \neq j}^{} d(v_{i},v_{j})
 

where, N is the number of vertices in the graph G; where, N \cdot (N \mathbin{-} 1) is the number of pairs of distinct vertices. Note that the number of pairs of distinct vertices is equal to the number of shortest paths between all pairs of vertices if we pick just one in case of a tie (two shortest paths with the same length).

In the context of ontology analysis, I would compare this metric as the speed of the ontology. What I mean by that is that one of the main tasks we do with an ontology is to infer new facts from known facts. Many inferring activities requires traversing the graph of an ontology. This means that the smaller the average path length between two classes, the more performant these inferencing activities should be.

Average Local Clustering Coefficient

The local clustering coefficient quantifies how well connected the neighborhood vertices of a given vertex are. It is the ratio of the edges that exists between all of the neighborhood vertices of a given vertex and the maximum number of possible edges between these same neighborhood vertices.


C_{i} = \frac{\raisebox{1.5mm}{$\vert \lbrace e_{jk} : v_{j}, v_{k} \in N_{i}, e_{jk} \in E \rbrace \vert$}}{\raisebox{-1.5mm}{$k_{i} (k_{i} \mathbin{-} 1 $}}
 

where k_{i} is the number of neighborhood vertices, $k_{i} (k_{i} \mathbin{-} 1 $}) is the maximum number of edges between the neighborhood vertices and \vert \lbrace e_{jk} : v_{j}, v_{k} \in N_{i}, e_{jk} \in E \rbrace \vert is the set of all the neighborhood vertices for a given vertex e_{jk}.

The local clustering coefficient \bar C is represented by the sum of the clustering coefficient of all the vertices of a graph G divided by the number of vertices in G. It is given by:


\bar C = \sum_{i=1}^{n} \frac{C_{i}}{n}

Betweenness Centrality

Betweenness centrality is a measure of importance of a node in a graph. It is represented by the number of times a node participates in the shortest path between other nodes. If a node participates in the shortest path of multiple other nodes, then it means that it is more important than other nodes in the graph. It acts like a conduit.


g(v) = \sum_{s \neq v \neq t}^{} \frac{\raisebox{1.5mm}{$\sigma_{st}(v)$}}{\raisebox{-1.5mm}{$\sigma_{st}$}}
 

where \sigma_{st} is the total number of shortest paths from node s to node t and \sigma_{st}(v) is the number of those paths that pass through v.

In the context of ontology analysis, the betweenness centrality will tell us which of the classes that participates the more often in the shortest paths of a given transitive property between other classes. This measure is interesting to help us understand how a subgraph is constructed. For example, if we take a transitive property such as rdfs:subClassOf, then the graphs generated by a subgraph composed of this relationship only should be more hierarchic by the semantic nature of the property. This means that the nodes (classes) with the highest betweenness centrality value should be classes that participate in the upper portion of the ontology (the more general classes). However, if we think about the foaf:knows transitive property between named individuals, then the results should be quite different and suggest a different kind of graph.

Initial UMBEL Graph Analysis

Now that we have a good understanding of some core graph analysis measures, we will use them to analyze the graph of relationship between the UMBEL classes and reference concepts using the subgraphs generated by the properties: rdfs:subClassOf, umbel:superClassOf, skos:broaderTransitive, skos:narrowerTransitive and rdf:type.

Density

The maximum number of edges in UMBEL is: 26 345 * (26 345 - 1) = 694 032 680 which is about two thirds of a billion of edges. This is quite a lot of edges, but it is important to keep in mind since most of the following ratios are based on this maximum number of edges (connections) between the nodes of the UMBEL graph.

Here is the table that shows the density of each subgraph generated by each property:

Class view Individual view
Metric sub class of super class of broader narrower type
Number of edges 39 410 116 792 36 016 36 322 271 810
Density 0.0000567 0.0001402 0.0000518 0.0000523 0.0001021

As you can see, the density of any of the UMBEL subgraphs is really low considering that the maximum density of the graph is 1. However this gives us a picture of the structure: Most of the concepts have no more than few connections between each node for any of the analyzed properties.

This makes sense, since a conceptual structure is meant to model relationships between concepts that represent concepts of the real world, and in the real world, the concepts that we created are far from being connected to every other concept.

This is actually what we want: we want a conceptual structure which has a really low density. This suggest that the concepts are unambiguously related and hierarchized.

Having a high density (let’s say, 0.01, which would mean that there are nearly 7 million connections between the 26 345 concepts for a given property) may suggest that the concepts are too highly connected which could suggest that using the UMBEL ontology for tagging, classifying and reasoning over its concepts won’t be an optimal choice because of the nature of the structure and its connectivity (and possible lack of hierarchy).

Average Degree

The average degree shows the average number of UMBEL nodes that are connected to any other node for one of the given property.

Class view Individual view
Metric sub class of super class of broader narrower type
Number of Edges 39406 97323 36016 36322 70921
Number of Vertices 26345 26345 26345 26345 26345
Average degree 1.4957 3.6941 1.3670 1.3787 2.6920

As you can see, the number are quite low: from 1.36 to 3.69. This is consistent with what we saw with the density measure above. This helps confirm our assumption that all of these properties create mostly hierarchical subgraphs.

However there seems to be one anomaly with these results, the average degree 3.69 of the umbel:superClassOf property. Intuitively, its degree should be near the one of the rdfs:subClassOf but it is far from this: it is more than twice its average degree. Looking at the OWL serialization of the UMBEL version 1.05 reveals that most umbel:RefConcept do have 3 triples:

  umbel:superClassOf skos:Collection ,
                     skos:ConceptScheme ,
                     skos:OrderedCollection .

This makes no sense that the umbel:RefConcept are super classes of these skos classes. I suspect that this got introduced via punning at some point in the history of UMBEL and got unnoticed until today. This issue will be fixed in a coming maintenance version of UMBEL.

If we check back the density measure of the graph, we notice that we have a density of 0.0001402 for the umbel:superClassOf property versus 0.0000567 for the rdfs:subClassOf property which has about the same ratio. So we could have noticed the same anomaly by taking a better look at the density measure.

But in any case, this shows how this kind of graph analysis can be used to find such issues in Big Structures (structures too big to find all these issues by scrolling the code only).

Diameter

The diameter of the UMBEL graph is like the worse case scenario. It tells us what is the longest shortest path for a given property’s subgraph.

Class view Individual view
Metric sub class of super class of broader narrower type
Diameter 19 19 19 19 4

In this case, this tell us that the longest shortest path between two given nodes for the rdfs:subClassOf property is 19. So in the worse case, if we infer something between these two nodes, then our algorithms will require maximum 19 steps (think in terms of a breath first search, etc).

Average Path Length Distribution

The average path length distribution shows us the a number paths that have x in length. Because of the nature of UMBEL and its relationships, I think we should expect a normal distribution. An interesting observation we can do is the the average path length is situated around 6, which is the six degree of separation.

We saw that the “worse case scenario” was a shortest path of 19 for all the property except rdf:type. Now we know that the average is around 6.

rdfs:SubClassOf

umbel-sub-class-of-transitive-paths-distribution

umbel:superClassOf

umbel-super-class-of-transitive-paths-distribution

Here we can notice an anomaly in the expected normal distribution of the path lengths. Considering the other analysis we did, we can consider that the anomaly is related to the umbel:superClassOf issue we found. What we will have to re-check this metric once we fix the issue. I expect we will see a return to a normal distribution.

skos:broaderTransitive

umbel-skos-broader-transitive-paths-distribution

 

skos:narrowerTransitive

umbel-narrower-transitive-transitive-paths-distribution

 

rdf:type

umbel-type-transitive-paths-distribution

 

Average Local Clustering Coefficient

The average local clustering coefficient will tell us how clustered the UMBEL subgraphs are.

Class view Individual view
Metric sub class of super class of broader narrower type
Average local clustering coefficient 0.0001201 0.00000388 0.03004554 0.00094251 0.77191429

As we can notice, UMBEL does not have the small-world effect with its small clustering coefficients depending on the properties we are looking at. It means that there is not a big number of hubs in the network and so that the number of steps from going from a class or a reference concept to another is higher than in other kinds of networks, like airport networks. This makes sense by looking at the average path length and the path length distributions we observed above.

At the same time, this is the nature of the UMBEL graph: it is meant to be a a well structured set of concepts with multiple different specificity layers.

To understand its nature, we could consider the robustness of the network. Normally, networks with high average clustering coefficients are known to be robust, which means that if a random node is removed, it shouldn’t impact the average clustering coefficient or the average path length of that network. However, in a network that doesn’t have the small-world effect, then they are considered less robust which means that if a node is removed, it could greatly impact the clustering coefficients (which would be lower) and the average path length (which would be higher).

This makes sense in such a conceptual network: if we remove a concept from the structure, it will most than likely impact the connectivity of the other concepts of the network.

One interesting thing to notice is the clustering coefficient of 0.03 for the skos:broaderTransitive property and 0.00012 for the rdfs:subClassOf property. I have no explanation for this discrepancy at the moment, but this should be investigated after the fix as well since intuitively these two coefficient should be close.

Betweenness Centrality

As I said above, In the context of ontology analysis, the betweenness centrality tells us which of the classes participates more often in the shortest paths of a given transitive property between other classes. This measure is useful to help us understand how a subgraph is constructed.

If we check the results below, we can see that all the top nodes are nodes that we could easily classify as being part of the upper portion of the UMBEL ontology (the general concepts). Another interesting thing to notice is that the issue we found with the umbel:superClassOf property doesn’t seem to have any impact on the betweenness centrality of this subgraph.

rdfs:subClassOf
PartiallyTangible 0.1853398395065167
HomoSapiens 0.1184076250864128
EnduringThing_Localized 0.1081317879905902
SpatialThing_Localized 0.092787668995485
HomoGenus 0.07956810084399618

You can download the full list from here

umbel:superClassOf
PartiallyTangible 0.1538140444614064
HomoSapiens 0.1053345447606491
EnduringThing_Localized 0.08982813055659158
SuperType 0.08545549956197795
Animal 0.06988077993945754

You can download the full list from here

skos:broaderTransitive
PartiallyTangible 0.2051286996513612
EnduringThing_Localized 0.1298479341295156
HomoSapiens 0.09859060900526667
Person 0.09607892589570508
HomoGenus 0.08806468362881092

You can download the full list from here

skos:narrowerTransitive
PartiallyTangible 0.2064665713447941
EnduringThing_Localized 0.1294085106511612
HomoSapiens 0.1019893654646639
Person 0.09366129700219764
HomoGenus 0.09047063411213117

You can download the full list from here

rdf:type
owl:Class 0.3452564052004049
RefConcept 0.3424169525837704
ExistingObjectType 0.0856962574437039
ObjectType 0.01990245954437302
TemporalStuffType 0.01716817183946576

You can download the full list from here

Analysis with Inferred UMBEL

Now, let’s push the analysis further. Remember that I mentioned that all the properties we are analyzing in this blog post are transitive? Let’s perform exactly the same metrics analysis, but this time we will use the transitive closure of the subgraphs.

The transitivity characteristic of a property is simple to understand. Let’s consider this tiny graph A \xrightarrow{p^{t}} B \xrightarrow{p^{t}} C where the property \xrightarrow{p^{t}} is a transitive property. Since \xrightarrow{p^{t}} is transitive, there is also a relationship A \xrightarrow{p^{t}} C.

Given A \xrightarrow{p^{t}} B \xrightarrow{p^{t}} C we inferred A \xrightarrow{p^{t}} C using the transitive relation \xrightarrow{p^{t}}.

Now, let’s use the power of these transitive properties and let’s analyze the transitive closure of the subgraphs that we are using to compute the metrics. The transitive closure is simple to understand. From the input subgraph, we are generating a new graph where all these transitive relations are explicit.

Let’s illustrate that using this small graph: A \xrightarrow{p^{t}} B \xrightarrow{p^{t}} C, B \xrightarrow{p^{t}} D. The transitive clojure would create a new graph: A \xrightarrow{p^{t}} B \xrightarrow{p^{t}} C, B \xrightarrow{p^{t}} D, A \xrightarrow{p^{t}} C, A \xrightarrow{p^{t}} D

This is exactly what we will be doing with the sub-graphs created by the properties we are analyzing in this blog post. The end result is that we will be analyzing a graph with many more edges than we previously had with the non transitive closure versions of the subgraphs.

What we will analyze now is the impact of considering the transitive closure upon the ontology metrics analysis.

Density

Remember that the maximum number of edges in UMBEL is: 26 345 * (26 345 - 1) = 694 032 680, which is about two thirds of a billion edges.

Class view Individual view
Metric sub class of super class of broader narrower type
Number of Edges 789 814 922 328 674 061 661 629 76 074
Density 0.0011380 0.0013289 0.0009712 0.0009533 0.0001096

As we can see, we have many more edges now with the transitive closure. The density of the graph is higher as well since we inferred new relationships between nodes from the transitive nature of the properties. However it is still low considering the number of possible edges between all nodes of the UMBEL graph.

Average Degree

We now see the impact of transitive closure on the average degree of the subgraphs. Now each node of the subgraphs are connected by 25 to 35 other nodes in average.

Class view Individual view
Metric sub class of super class of broader narrower type
Number of Edges 789 814 922 328 674 061 661 629 76 074
Number of Vertices 26345 26345 26345 26345 26345
Average degree 29.97965 35.00960 25.58591 25.11402 2.887606

One interesting fact is that the anomaly disappears with this transitive closure subgraph for the umbel:superClassOf property. There is still a glitch with it, but I don’t think it would raise suspicion at first. This is important to note since we won’t have noticed this issue with the current version of the UMBEL ontology if we would have analyzed the transitive closure of the subgraph only.

Diameter

Class view Individual view
Metric sub class of super class of broader narrower type
Diameter 2 2 2 2 2

As expected, the diameter of any of the transitive closure subgraphs is 2. It is the case since we made explicit a fact (a edge) between two nodes that was not explicit at first. This is good, but this is no quite useful from the perspective of ontology analysis.

This would only be helpful if the number were not 2 which would suggest some errors in the way you computed the diameter of the graph.

However what we can see here is that the speed of the ontology (as defined in the Average Path Length section above) is greatly improved. Since we forward-chained the facts in the transitive closure sub-graphs, it means that knowing if a class A is a sub-class of a class B is much faster, since we have a single lookup to do instead of an average of 6 for the non transitive closure version of the subgraphs.

Average Path Length Distribution

All of the path length distributions will be the same as this one:

umbel-sub-class-of-transitive-paths-distribution-inferredSince the diameter is 2, then we have a full lot of paths at 2, and 26 345 at 1. This is not really that helpful from the standpoint of ontology analysis.

Average Local Clustering Coefficient

Class view Individual view
Metric sub class of super class of broader narrower type
Average local clustering coefficient 0.3281777 0.0604906 0.2704963 0.0138592 0.7590267

Some of the properties like the rdfs:subClassOf property shows a much stronger coefficient than with the non-transitive closure version. This is normal since all the nodes are connected to the other nodes down the paths. So, if a node in between disappears, then it won’t affect the connectivity of the subgraph since all the linkage that got inferred still remains.

This analysis also suggests that the transitive closure version of the subgraphs are much stronger (which makes sense too).

However I don’t think this metric is that important of a characteristic to check when we analyze reference ontologies since they do not need to be robust. They are not airport or telephonic networks that need to cope with disappearing nodes in the network.

Betweenness Centrality

What the betweenness centrality measure does with the transitive closure of the subgraphs is that it highlight the real top concepts of the ontology like Thing, Class, Individual, etc. Like most of the other measures, it blurs the details of the structure (which is not necessarily a good thing).

rdfs:subClassOf
SuperType 0.03317996389023238
AbstractLevel 0.02810408526564482
Thing 0.02772171675862925
Individual 0.02747482318621853
TopicsCategories 0.02638342698407473

You can download the full list from here

umbel:superClassOf

The interesting thing here is that this measure actual shows the actual concepts for which we discovered an issue with above.

skos:Concept 0.02849203320293865
owl:Thing 0.02849094898994718
SuperType 0.02848878056396423
skos:ConceptScheme 0.02825567477079737
skos:OrderedCollection 0.02825567477079737

You can download the full list from here

skos:broaderTransitive
Thing 0.03225428380683926
Individual 0.03196350419108375
Location_Underspecified 0.0261924189600178
Region_Underspecified 0.02618796825161338
TemporalThing 0.02594169571990208

You can download the full list from here

skos:narrowerTransitive
Thing 0.03298126713602109
Location_Underspecified 0.02680852092899528
Region_Underspecified 0.02680398659044947
TemporalThing 0.02648809433842489
SomethingExisting 0.02582305801837314

You can download the full list from here

rdf:type
owl:Class 0.3452564052004049
RefConcept 0.342338078899975
ExistingObjectType 0.0856962574437039
ObjectType 0.01990245954437302
TemporalStuffType 0.01716817183946576

You can download the full list from here

Conclusion

This blog post shows that simple graph analysis metrics applied to Big Structures can be quite helpful to understand their nature, how they have been constructed, what is their size, their impact on some algorithms that could use them, and to find potential issues in the structure.

One thing we found is that the correlation between the properties rdfs:subClassOf and skos:broaderTransitive are nearly identical. They nearly have the same values for each metrics. If you were new to the UMBEL ontology you wouldn’t have known this fact without doing this kind of analysis or by spending much time looking at the serialized OWL file. It doesn’t tell us anything about how similar the relations are, but it does tell us that they have the same impact on the ontology’s graph structure.

Performing this analysis also led us to discover a few anomalies with the umbel:superClassOf property, suggesting an issue with the current version of the ontology. This issue would have been hard to notice, and understand, without performing such a graph analysis to the structure.

However, I also had the intuition that the analysis of the transitive closure of the subgraphs would have led to more interesting results. At best that analysis did confirm a few things, but in most of the cases it only blurred the specificities of most of the metrics.

These analysis metrics will soon be made available as standard Web services, so that they may be applied against any arbitrary graph or ontology.