We demonstrate a framework for automated information system generation from a given data ontology describing either an existing data set, or a data set to be filled up and maintained using the created information system. The IS data are stored into real or virtual RDF data stores and are accessed by means of SPARQL queries. The data ontology structure information and annotations allow automated generation of IS reflecting the data structure specified in the ontology.
The semantic web standards, including RDF [
We demonstrate here a system for automated information system user interface
creation from the structure of data described as annotated OWL ontology thus
allowing browsing the available RDF data, as well as editing the data, if the
corresponding RDF data store allows for data updates and data entry. There are two
basic use cases of the OBIS Framework: (i) to browse and analyze data in existing
1 Partially supported by Latvian State Research programme NexIT project No.1 'Technologies
of ontologies, semantic web and security'
real or virtual RDF databases, including the RDF databases that are conceptual
representations of relational database data [
We structure the presentation into two main sections – first is explained the basic
OBIS application generation, followed by application tuning possibility description.
The conceptual foundation of the OBIS system has been described in [
The ontology-based information system creation in OBIS starts with obtaining or
designing of data ontology reflecting the structure of the data underlying the
information system. Figure 1 contains an example mini-University ontology in the
OWLGrEd2 ontology editor notation [
Person personName personID nationality:Nationality
Nationality nCode nValue
FacultyLevel acadTitle With the OBIS framework up and running on a web-server the user creates a new application in the OBIS Application Manager. The user then fills in the application name as well as ontology namespace (from the data ontology) and data namespace (not relevant for read-only ontology data access), chooses the repository type (e.g. Virtuoso server3, a SPARQL-endpoint or a Jena TDB store) and connection
Student studentName{<personName} studentNumber student
1 parameters. After saving the created basic application information and loading the data ontology the concrete application based on the data ontology can be generated and run. Figure 2 shows mini-University example case of the obtained application working interface, consisting of a class tree browser with the option to select a class and obtain a table view of its instance attributes.
Fig. 2. Example: the generated mini-University information system: the instance table view for Course class. Further ontology data exploration options include the sorting and filtering, export to MS Excel as well as related data item view allowing the user to select one or several instance links defined for the class in the ontology as well as a list of instances and obtain the lists of items connected to the selected ontology instances via the specified links. Figure 3 demonstrate this facility for the courses ‘Semantic Web’ and ‘Quantum Computations’ in the mini-University example (note the properties includes and takes coming into the Course class in the data ontology in Figure 1).
There is also a detailed view form for instances with fields for attribute values and links in the ontology data coming in (the direct links) and going out (the inverse link), as in Figure 4. Show Permalink option helps sharing the instance property page among various developers working with the same OBIS server application.
OBIS can be used not only as the ontology data information viewer. In the case, if the underlying RDF data store supports SPARQL data updates, there is possibility of new information entry, as well as information update and delete in OBIS, as well. The class instance information entry and edit forms are similar to instance view forms of Figure 4 (with information save option enabled); the link information is updated via the Add option in the tables showing link relations.
Further options in OBIS are report creation that allows also execution of textual
SPARQL queries against the underlying RDF data store as well as form and report
information export to MS Excel. The internal report definition tool in OBIS is work in
progress. To ease writing textual SPARQL queries one can use e.g. a ViziQuer tool4
[
An information system user interface that is based just on the reproduction of the
structure of ontology classes and data and object properties (with an analogy with
RDB table, attribute and link structure) may appear to be too simplistic for a useful
information system. One could desire e.g. more specific table and form view
definitions for particular classes. In particular, the classes containing the classifier
data, or, in general, the links among classes with their target maximum cardinality not
exceeding 1 (i.e., the n:1 links) may benefit from special visual representation in the
user interface. There are three basic mechanisms of enhancing OBIS applications:
- specification of cardinality information in the ontology (this requires using the
standard means available in OWL, so the effects of cardinality information
availability can be considered basic OBIS application generation aspect);
- extending OWL ontology with specific user interface annotations, and
- customizing (configuring) the application after its initial generation.
The main user interface annotations, supported by OBIS, are the following5:
- obis:textPattern. Defines textual form for class instance presentation via
relations with maximum cardinality 1. The textual form of linked instances of
such classes are included into linking instance presentations both in table and
form and edit views (a combo-box is used for the edit view).
- obis:isEnumerated. Specifies a class to be a classifier class (an enumerated
class), so that a textual form of its instance presentation is available
disregarding an explicit obis:textPattern specification by the user.
- obis:defaultOrder. Specifies the default order of attributes (both data attributes
and classifier attributes) for class instances within a class table or form view.
The OWL ontology axiom structure does not foresee a possibility to specify
the order of the attributes for a class. The obis:defaultOrder annotations create
a compensating mechanism. An automated generation of these of axioms is
built in OWLGrEd editor extension OWLGrEd/OBIS [
May include both attributes ascribed for the class and its subclasses in the
ontology, as well as any other attributes (e.g. attributes whose domain (ascription) in the ontology is not defined). The obis:view annotations, where specified, take precedence over obis:defaultOrder annotations and it is in general considered that the obis:view annotations are to be entered manualy.
Figure 5 shows the mini-University ontology of Figure 1 extended with some
example user interface annotations added. The notation used is domain specific
ontology notation in the OWLGrEd ontology editor [
The option of fine-tuning the OBIS application after its initial generation is based on the principle of availability of the OBIS application configuration also as OBIS application itself, so OBIS itself can be used to perform its application fine-tuning. 4
The OBIS Framework demonstrates the feasibility of automated generation of
feature-rich information systems just from annotated data ontologies. Besides a
potential stand-alone usability for viewing or managing data structured in accordance
to the RDF format, the concepts from OBIS might appear useful in the context of
other editors of instances corresponding to structure defined by OWL ontology or
RDF Schema, such as Web Protégé [
Although the OBIS tool can be used to interpret data ontologies created by
different tools, its integration within the OWLGrEd tool [
There is work in progress towards creating the custom-built reports, extending the
validation rule set, as well as extending the set of annotations, interpretable by OBIS,
however, trying to keep the annotations to be specified by the user to a
comprehensible minimum. The advances towards creating user-comprehensible UI
annotations that allow generation of fully functional UI from the annotated data
model could be of value also for UI generation from UML-style conceptual models.
Another possible venue of the OBIS extension is via integration of the visual
SPARQL query creation tool ViziQuer [