OBDA plugin for NeOn Toolkit User Guide
Contents
1. Cancel 3 Verify that when opening the perspective a new menu item called OBDA is added to the main menu bar Fie Edt Navigate Search Project cena Run Window Help 4 Verifiy that a folder called OBDA which contains two views called Individuals Queries View and OBDA View was added the under Window Show View ab Individual Query View E amp OBDA BR OBDA view Updating an existing installation To update a previous version of the OBDA plugin simply remove the folders of the old version from the plugins folder of your NeOn Toolkit installation and repeat the installation steps for the new versions Remove the OBDA Plugin To delete the OBDA plugin from your current NeOn Toolkit installation simply remove the 3 plugins included in the package from the plugins folder of you installation 1 2 OBDA files The OBDA plugin uses so called obda files to store and load information about data sources mappings and queries These files are created automatically for each ontology More over each time the active ontology is changes all OBDA data for the current ontology is stored in its obda file and the stored data for the newly selected ontology is loaded from its obda file The obda files are located in a folder named obda located in each project folder with in the workspace of the NeOn Toolkit installation The obda files are named as the owl file of the ontologies related to them For example if we have an owl file ca2. SELECT WHERE 4 x r SELECT fn in securit SELECT fname in sec C v 13 1 The OBDA extension to DIG 1 1 The OBDA Plugin for NeOn implements a combination of several standards for interac tion with DL reasoners specifically they implement the DIG 1 1 Interface the DIG 1 2 extensions and the OBDA extensions for DIG 1 1 The later is an extension to the original DIG protocol developed with in the TONES project and aimed at standardizing access to reasoners which offer OBDA related functionality The OBDA extension to DIG 1 1 was designed to accommodate for the different characterizations of the OBDA scenario Specifically the standardization proposal was formulated with the following requirements in mind e To based on already available and used standards e To separate of the general OBDA related functionality and the details related to specific types of sources and mappings e To be flexible with respect to the types of data sources allowed To be flexible with respect to syntax and semantics for mapping formalisms To be open standard which is easily available and extensible The resulting extension was presented in and a full description is available at its websitet Ihttp obda inf unibz it dig 11 obda 14
3. That is a MySQL console client convention and should be avoided Data source inspection If the data source is an RDBMS data source one can use the Source Schema Inspector and Direct Source Query tabs in the data source manager to interact with the RDBMS e With the Direct source query tab indside the Data Source Inspector component you can issue SQL queries to the selected data source To verify the body of the mappings for example gt Data source inspector Source Schema Inspector Direct source querry SQL Querry select from language Results isocode name EN English IT italiano DE deutsch e With the Source Schema Inspector you can review the schema of the selected data source Data source inspector Source Schema Inspector Direct source querry Relations Attributes Relation Name Row Count A Attribute Autoincrement audio id False copy 3 Firstname False country lastname false customer street False district districtname false dvd districtzip False Function dateofbirth False lanquage lt 1 6 The IndividualQueriesView The IndividualQueriesView provides a way to send and visualize the result of non standard ABox individual centric queries to a reasoner It also provides means to man age these queries save organize plus visualize the results obtained from the reasoner At the moment the plugin supports Union of Conjunctive Queries UCQs exp
4. OBDA Run Window Help mS a iP iwi ES BR OBDAPerspec dh Ontology Navigator 23 E BR OBDA View 3 0 Individual Query View a I E NewOntologyProject OWL I E NewOntologyProject1 OWL Data source manager 4 1S NewOntologyProject2 OWL dh http fobda inf unibeitfstockex Select project NewOntologyProject2 refresh Select ontology http obde inf unibe it st_ E Classes Address LJ ig a FinantialInstrument S Datesources person om ge JDBC URL abcimysal Mocalhost StockExchangeList a e E Transaction tH D Object Properties Database Username nto 2 Data Properties e C Annotation Properties E Datatypes JDBC Driver com mysql jdbc Driver Source ID Database Name stockclient1 Database Password Data source mappings Mapping 1D Head Body a Mappings mapping addresses S t mapping brokers mapping broker address E console 23 a o Reading Document 09 40 03 Parsing document 09 40 04 Finished sending Document 09 40 04 Synch result true Figure 1 1 The OBDA perspective The information about RDBMS data sources and their mappings is transmitted to the reasoner during the synchronization stage in which it gets translated into DIG in the way defined in the RDBMS Module for the OBDA extension to DIG 1 1 see Appendix The OBDA plugin also adds facilities to inspect RDBMS data sources and to query them directly These two features are int
5. OBDA plugin for NeOn Toolkit User Guide Mariano Rodr guez Muro Manfred Gerstgrasser Free University of Bozen Bolzano 31 of October 2008 Contents 1 User Guide 3 1A Installation uu ei e ogee A a 3 1 1 1 Requirements tcs ai a ee tE ad 3 1 1 2 Obtaining the software o o o 3 12 OBDA Diles nar 3 took ee Geli a wh ees e Ba ie eS 5 1 3 OBDA Perspective e 6 14 gt OBDA VIEW us ic a e Re a Pe os 7 1 5 Visual guide for the OBDA view 2 0200 8 1 6 The IndividualQueriesView 0 0 00002 eee ee 10 1 7 Visual guide for the IndividualQueriesView 11 1 The OBDA extension to DIG 1 1 o e 13 1 User Guide 1 1 Installation 1 1 1 Requirements 1 Java 1 5 or above 2 NeOn Toolkit 1 2 3 An OBDA Enabled Reasoner Note The OBDA Plugin requires an OBDA enabled reasoner for query answering support An OBDA enabled reasoner is a DL reasoner that supports the OBDA archi tecture and implements the OBDA extension to DIG used by the OBDA Plugin At the moment of writing this manual the only the DIG Server for QuOnto works with this architecture The DIG server for the QuOnto reasoner is available for academic use with agree ment to the authors Please contact them directly to obtain a copy of the reasoner Information about QuOnto can be found at http www dis uniromal it quonto 1 1 2 Obtaining the software The OBDA Plugin for th
6. as workspace 2 Open the file workspace build build xml with any text editor and edit the 3 first properties as specified in the file The property dir workspace is the folder where the sources where unpacked The property dir eclipseplugins is the folder that hold libraries of the eclipse plugins required to build the NeOn toolkit e g the contents of the plugins folder of the ntk1 2 0 2008 10 02 datamodel package of the NeOn toolkit sources The property dir neoneclipseplugins is the folder that holds the libraries of the main plugins of the NeOn toolkit e g the contents of the workspace folder of the ntk1 2 0 2008 10 02 plugins package of the NeOn toolkit sources 3 Open a terminal window and CD into the workspace build folder 4 Build the sources by invoking the command ant After building the sources one can proceed with the binary installation steps Binary Installation steps 1 Close the NeOn Toolkit 2 Copy the folders e dig plugin 1 0 0 e inf unibz it obda neon 1 0 0 and e inf unibz it obda neon api_1 0 0 into the plugin folder of your NeOn Toolkit installation Verifying the installation 1 Start your NeOn Toolkit installation 2 Verify that a new perspective called OBDA perspective was added to the perspec tive lists and open it N Open Perspective q F Logic default FE F Logic source 3 Genetic Optimizer 3 Integration BRopDAPerspective Elow p Rule Query ES Rule debug Show all
7. ce manager Select project w refresh Select ontology w refresh a Source ID Data sources O Randestockexchange JDBC URL jdbe mysql flocalhost 8880 Database Name stockclient1 Database Username root Database Password eoccce JDBC Driver com mysql jdbc Driver E Update e To configure the connection parameters select the data source edit its details in the parameters editor and finally press the UPDATE button e To add a new data source click on the PLUS button and enter a name for it e To delete a data source select it and than press the MINUS button and the data source will be removed Create Update mappings with the mapping manager e Add a new mapping by clicking on the PLUS button e To delete a mapping first select it and press the MINUS button e To duplicate a mapping first select it and then press the DUPLICATE button e To update a mapping first select it then modify its ID Head and Body in the appropriate textfields and finally press the UPDATE button Mapping ID Head update Body l a Mappings mapping addresses E mapping brokers le mapping broker address Note The red bar in the icon of the mapping heas means that it contains some errors and it should be reviewed The mapping head is a CQ wheras the mapping body must be a valid SQL query DO NOT USE AT THE END OF THE SQL QUERIES
8. e NeOn toolkit can be downloaded from http obda inf unibz it neon plugin The package is composed by e The OBDA PLUGIN An open source plugin for the NeOn toolkit which provides facilities to create an associate RDBMS data sources and source ontology mappings to ontologies It is built on top of the OBDA API PLUGIN and the DIG PLUGIN e The OBDA API An open source library that allows for the manipulation of data sources and mappings e The OBDA API PLUGIN An open source Neon plugin which exposes the OBDA API classes to the Neon Toolkit enabling third parties to use the API in their own projects e The DIG PLUGIN A Neon Toolkit plugin which enables interaction with DIG reasoners Developed by the University of Karlsruhe in cooperation with the Free University of Bozen Bolzano The package can be downloaded in binary format as well as source code In the following sections we will explain the procedure for compiling the sources into binaries and the binary installation procedure Compiling the sources Requirements Compilation of the plugins from the sources requires e Java JDK 1 5 or greater e A properly installed java ANT system e The contents of the ntk1 2 0 2008 10 02 datamodel package from the NeOn toolkit developers package e The contents of the ntk1 2 0 2008 10 02 plugins package from the NeOn toolkit developers package 1 Unpack the source package onto any folder From now on we will refer to this folder
9. end to help users during the modeling process We clarify that the reasoner doesnt intervene when using these features instead the plugin will attempt to contact the RDBMS directly and execute the users query This is helpful to verify the connection parameters specified to inspect to databases schema during or to test queries which later could be used for mappings Syntax of RDBMs mappings RDBMS mappings are composed by a mapping body and a mapping head The mapping body accepted for RDBMS sources is any valid SQL query Note that the plugin doesn t parse this query therefore the user must make sure that the query is really a valid SQL query in order to avoid problems during the synchronization or query answering stages The mapping head accepted for RDBMS sources is in the form of a Conjunctive Query over the ontology expressed in a datalog syntax For the query to be valid it must only refer to predicates Classes or Properties already defined in the ontology The semantics of the mappings is not defined in the plugin In that sense the plugin adopts the same position as the RDBMS module for the OBDA extension to DIG 1 1 in which the semantics of the mappings is left undefined so that implementors of OBDA reasoners are able to define the specific semantics they use In short it is crucial to understand the semantics implemented by the reasoner as it defines the behavior of the OBDA system modeled with the plugin Some valid sample mapp
10. enubar This menu allows you to a trigger a synchronization action with the server b to enable an disable the OBDA synchronization c to change the server s connection parameters and d to disable or enable reasoning in the server during query answering In case you changed something in the perspective and you want have back its initial look you can reset the perspective by right clicking on the OBDA perspective button in the top right corner of the window and selecting reset in the appearing popup menu 1 4 OBDA View Overview The OBDA View is the core GUI component of the OBDA plugin It allows you to define one or more data sources to specify their access parameters and to create mappings which relate the data in those data sources to the Classes and Properties of the ontology The default installation of the plugin allows you to create RDBMS data sources ac cessible through JDBC Associated to a RDMBS data source one can define a set of semantic mappings These are on the form described in the OBDA extension to DIG 1 1 that is a query over the ontology and a query over the data source Specifically in the case for RDBMS data sources a mapping is composed by a SQL query over the RDBMS source the mapping s body and a the query over the ontology is Conjunctive Query CQ in data log syntax the mappings head 60 OBDAPerspective NeOn Toolkit C Program Files NeOn NeOnToolkit21 workspace File Edit Navigate Search Project
11. ings The following are a few examples of valid mappings We do not describe the ontologies nor the schemas of the databases as are self explanatory We stress that the exact meaning of these mappings depends solely on the semantics implemented in the reasoner Example 1 4 1 A simple mapping for a class HEAD Person getPerson id BODY SELECT id FROM employee Example 1 4 2 A mapping for an object property involving the a join between of two tables and renaming of variables HEAD worksFor getPerson employeeid getPerson managerid BODY SELECT employee id AS employeeid area managerid AS managerid FROM employee area WHERE employee area area id Example 1 4 3 A simple mapping for a datatype property HEAD hasID getPerson id id BODY SELECT id FROM employee Example 1 4 4 A complex class property mapping asserting membership in a class and participation in an object property at the same time HEAD Woman getPerson childid Daugther getPerson childid daughterOf getPerson childid getPerson parentid BODY SELECT childid parentid FROM childof 1 5 Visual guide for the OBDA view Data source manager The data source manager allows one to add delete and modify data sources related to an ontology The first step is to select the ontology to work with Once the ontology is selected all already existing obda data related to this ontology is loaded automatically and than one can start modifying it Data sour
12. lled ontologyl owl the corresponding obda file will be ontology1_owl obda If you already have an obda file and you want to use it within the NeOn Toolkit you can copy it into the obda folder of the project to which the ontology belongs Note the obda file has to explicitly relate the data sources to the ontology for which you want to use it This can be easily checked by opening the file and comparing the DatasourceUri of the data source elements with the actual URI of the ontology 1 3 OBDA Perspective The default installation of the OBDA plugin adds a new perspective called OBDA Perspective Switching to it opens all the needed components for working with the OBDA Plugin Specifically it opens the OBDA View the Individual Query View and an output console The OBDA View is the main GUI component of the plugin and allows you to define one or more data sources to specify their access parameters and to create mappings which relate the data in those data sources to the Classes and Properties of the ontology The Individual Queries View provides a way to send and visualize the result of non standard ABox individual centric queries to a reasoner It also provides means to man age these queries save organize plus visualize the results obtained from the reasoner The output console is mainly used for showing the progress of different operations Furthermore the perspective adds an additional menu item called OBDA to the main m
13. omen which have a child which is itself a women with a child SELECT x WHERE 4 x hasChild y x rdf type Woman y hasChild z y rdf type Woman Retrieve all women which have a child OR all women which work SELECT x WHERE 4 4 x hasChild y x rdf type Woman x rdf type Woman x rdf type Worker 1 7 Visual guide for the IndividualQueriesView Union of Conjunctive Queries The basic query facility offered by the plugin is the SPARQL query tab Through this tab Union of Conjunctive Queries can be issued to the reasoner The UCQs are written in a limited SPARQL syntax Write the query in the input box and press the EXECUTE button This will open a console window and where the progress of the operation is shown Once the dialog indicates that the responses were retrieved go back to the Individual Query View where the results are shown 11 SPARQL EOL Query SPARQL x addressID id x sincity city x inState state x inStreet street SELECT WHERE xrdf type Address x x inCountry country x shasNumber number Query Results Once a query was succesfully executed the results if there are any are shown in the Query Results part of the Individual Query View Query Results Results x Query manager the project getAddressObj 991 getAddressObj 992 getAddressObj 995 getAddressObj 996 getAddressObj 993 getAddres
14. ressed in a SPARQL syntax Union of Conjunctive queries are a class of queries which is con sidered rather powerful and which are recently catching the attention of researchers in the Description Logics DL and Ontology communities Several reasoners have already incorporated support for UCQs as an additional reasoning service e g QuOnto Racer 10 Cerebra Engine Pellet KAON2 SHER An important remark is that answering con junctive queries about individuals over description logics ontologies is a complex problem which in general requires complex and resource time memory consuming algorithms therefore reasoners impose restrictions regarding the semantics and the structure of queries It is of great importance to understand these as the answers obtained from the reasoner depend directly on them UCQs written in this tab are sent to the reasoner through DIG In order to acomplish this the plugin translates the queries into equivalent DIG 1 2 conjunctive queries This enables the OBDA plugin to interact with any reasoner offering UCQ answering services as long as they implement the DIG 1 2 CQ extensions Some sample UCQs The following are a couple of UCQs written as SPARQL queries and their intended meaning under traditional semantics Retrieve all parents and their children SELECT x y WHERE x hasChild y Retrieve all women and their children SELECT x y WHERE x hasChild y x rdf type Woman Retrieve all w
15. sObj 997 street Road street ia Marconi Huberg Strasse via Piani di Bolzano Romer Street Samara road city Chonala Bolzano Bolzano Marconi Malaga Puebla country Mexico Italy Italy Italy Spain Mexico state Veracruz Bolzano Bolzano Trentino Malaga Puebla It is possible to save frequently used queries using the save button on the right bottom corner of the query area next to the execute button A small popup window will apear were you can enter a unique identifier for this query and may add it to an exiting or new group By rightclicking on the Query Manager panel a popup menu will apear and you can delete the selected querie or group The queries saved there are included in the obda file for permanent storage which means they will be available with Saved Queries a S A Simple queries ab Addresses aw Persons Addres ab StockTraders 5 AH cos GAH EQL SELE iv 12 Query history In the query hisotry panel all executed queries are recorded By rightclicking on it a popup menu will apear with the delete button one can remove the selected query the configure button opens a dialog window where one can change the maximum number of recorded query in the history Query manager Query history Query history a SELECT WHERE x r SELECT fname in sed
Download Pdf Manuals
Related Search