You are here

Search-On

Introduction

The Search-On environment has been designed as an interface –based on Semantic Web technologies– to enable domain experts and ontology engineers to explore the data accessible over a SPARQL endpoint.

This environment reduces the complexity of using the semantic technologies by hiding the ontology constructs by means of an easy to use interface based on semi natural language inputs. The web interface entails different methods to navigate through semantic data sets, including:

‒keyword-based semantic search
‒graph visualization
‒faceted interfaces
‒sentence construction based on natural language
 

The user-friendliness of the interface has been a key objective in the design of the environment. In the course of this, the annotations/metadata of the ontology elements play an important role in the presentation of data to the user, as well as in the interpretation if the users’ queries.

The environment has been validated through their application in the integrated data of the Semanco project. However, the prototype environment is generic enough to be applied to multiple domains and cases.

The environment will be made available to the scientific and academic communities soon.

 

Search-on features

The following section describe the main features of the existing prototype of the tool. The tool will be made available to the scientific and academic communities soon.


    Interested people in testing the tool please contact Álvaro Sicilia (asicilia@salleurl.edu)
 
Search method

The input search is a phrase composed of two types of elements: the concepts which are ontology classes (in terms of OWL) and the relations which are the object properties that connect the classes.

The sentence introduced by the user is the input for a verification method whose purpose is twofold:

  • to confirm that the sentence adheres to the ontology stored in the endpoint
  • to generate an SPARQL query to retrieve data from the endpoint

Each word of the sentence is matched to a class or an object property of the ontology. The pattern followed by the sentence is:

        concept > relation > concept 

 
Interface: Searching

A suggest feature has been implemented as a drop-down list which is activated while the user is typing the input. The drop-down list is filled with the concepts of the ontology which describes the data source. In the example the Land concept is selected. The user can complete the sentence with the support of the environment. The search pattern is highlighted with the colors “blue” for concepts and “red” for object properties. Each time a sentence is correct can be executed and the results are shown.

 

 

Interface: Displaying results

When the user searches the input sentence is processed and shown to the user to ensure that the user input is the same as the query engine processed. The results are shown following the input query sentence.

Other option is to show the results in a table layout. The instances are organized in four columns in the detailed mode displaying all the object and data properties.

 

Interface: Durther exploring

The instances can be further explored. A pop-up box is shown to the user with the expanded visualization of the selected concept and its relations visualized as a graph structure. When a relation is clicked its details are shown.

 

Technical details

The Search-On tool has been developed in PHP using the framework Code Igniter. The graph representation has been implement using the VivaGraphJS library and ARC to parse RDF files.

  - CodeIgniter, Open source PHP web application framework – http://codeigniter.com
  - Graph drawing library for JavaScript – https://github.com/anvaka/VivaGraphJS
  - ARC, Appmosphere RDF classes – https://github.com/semsol/arc2/wiki

 

References

Deliverable 4.3: User interfaces for domain experts interacting with SEIF by Michael Wolters, German Nemirovski, Joan Pleguezuelos, Álvaro Sicilia
The report is a description of two environments created in the project to support the ontology design and building, as follows. 1. An ontology editor, a tool to facilitate cooperative ontology design involving ontology designers and domain experts, such as building engineers and energy consultants. 2. A semantic data explorer to enable domain experts to explore the data exposed by the Semantic Energy Information Framework (SEIF).  

 

Copyright (C) 2014 

ARC Enginyeria i Arquitectura La Salle - Universitat Ramon Llull
    Joan Pleguezuelos
    Leandro Madrazo (madrazo@salleurl.edu)
    Álvaro Sicilia (asicilia@salleurl.edu)