Prototype Implementation Progress

Transcription

1 Prototype Implementation Progress Group: Hoang Huu Hanh Nguyen Huu Tinh 17/05/ /06/2004

2 Content Results Code fragments of the Monitor module Demo Explore some features of Joseki 2.1 Demo Next Step 03/05/2004 Group: Hanh & Tinh 2

3 Previous Meeting Developing module for monitoring and selective parsing the new messages into RDF model. Building user interface for the system. Creating Inference layer for this prototype supporting for semantic querying. 03/05/2004 Group: Hanh & Tinh 3

4 Results (1/2) Accessing to mailing systems, periodic checking for new messages and retrieving them from Parsing the new messages to RDF model Uniting the new model to existing model persisted in a MySQL database. Accessing and reading the RDF model stored in a MySQL database. Making union of two models. Persisting the union model to a MySQL database 03/05/2004 Group: Hanh & Tinh 4

5 Results (2/2) Explore Joseki (version 2.1) Joseki Introduction, Installation & Run Add a model to RDF Server From a RDF file From a MySQL database Fetch and Query a RDF model from Joseki from a Web browser Fetch and Query a RDF model from Java program 03/05/2004 Group: Hanh & Tinh 5

6 Monitor Module Code Fragments (1/4) Accessing mailing system, monitoring for new messages and parsing them to RDF data. Store store = session.getstore("imap"); "); store.connect(host,, username, password); Folder folder = store.getfolder(inbox); folder.open(folder.read_write); // Add messagecountlistener to listen for new messages folder.addmessagecountlistener addmessagecountlistener(new MessageCountAdapter() { public void messagesadded(messagecountevent ev) ) { Message[] msgs = ev.getmessages(); System.out.println("Got " + msgs.length + " new messages"); for (int( i = 0; i < msgs.length; ; i++) { try { <... Codes of scattering messages info to arrays...> } } }}); 03/05/2004 Group: Hanh & Tinh 6

7 Monitor Module Code Fragments (2/4) Creating a new model from info stored in arrays Model m = createmm(msgid,, from, nfrom,, to, nto, subj, mimetype, sdate,, count); // dump the new model out m.write(system.out); System.out.print("\n"); "); Merging the new model into existing model stored in a database loadrdb("mailrdf", ", "root", "", "MailMessages" MailMessages", m); Cheking for new messages in every ms int freq = Integer.parseInt(200000); for (; ;) { Thread.sleep(freq); // sleep for freq milliseconds // This is to force the IMAP server to send us // EXISTS notifications. folder.getmessagecount(); } 03/05/2004 Group: Hanh & Tinh 7

8 Monitor Module Code Fragments (3/4) Creating a new model from info stored in arrays Function createmm() Model model = ModelFactory.createDefaultModel(); Property _sn_ = model.createproperty(smmuri, ssname); Property _fr_ = model.createproperty(smmuri, sfrom); Property _rn_ = model.createproperty(smmuri, srname); Property _to = model.createproperty(smmuri, sto); Property _sb_ = model.createproperty(smmuri, ssubj); Property _mime = model.createproperty(smmuri, smime); try { for (int( i=0; i < msgcount; ; i++){ // Create a new resource for each message with URI is msgid Resource msg = model.createresource(msgid[i]); // add values of properties of a message msg.addproperty(_fr, from[i]).addproperty(_sn,, ((nfrom[i nfrom[i]!= ]!=null)?nfrom[i]:"no-name")) name")).addproperty(_to, to[i]).addproperty(_rn,, ((nto[i nto[i]!= ]!=null)?nto[i]:"no-name")) name")).addproperty(_sb,, ((subj[i subj[i]!= ]!=null)?subj[i]:"no-subject")) subject")).addproperty(dc.date, sdate[i]).addproperty(_mime, mimetype[i]); } 03/05/ }

9 Monitor Module Code Fragments (4/4) Uniting the new model to existing model in a database Function loadrdb() // Read a model from the database ModelRDB em = ModelRDB.open(conn, dbmodel); // uniting two models and write out a RDF file Model agg_model = em.union(newmodel); agg_model.write(new FileOutputStream("messages.rdf"), "), "RDF/XML-ABBREV"); // remove the old model and persist the aggregated model into a database em.remove(); em = ModelRDB.createModel(conn, dbmodel); em.read(new FileInputStream("messages.rdf"), "), "RDF/XML-ABBREV"); 03/05/2004 Group: Hanh & Tinh 9

10 Demo...

11 Joseki 2.1 Joseki is a server for publishing RDF models on the web. Models have URLs and they can be accessed by query using HTTP GET Joseki provides a RDF WebAPI that is based on extracting a subgraph from the published RDF. The extracted RDF can then be processed locally with the API provided by Jena Joseki is extensible: new query languages and new operations can be added without modifying the core system Joseki is open source and has a BSD license 03/05/2004 Group: Hanh & Tinh 11

12 Run Joseki RDF Server Set the classpath: : put all the JAR files in the lib/ directory on our classpath From Windows : cd <the root directory> set JOSEKIROOT=<the root directory> bin\joseki_path set CLASSPATH=%CP% bin\rdfserver The default Joseki configuration file is etc/joseki.n3. We will need to write our own to identify the models we wish to publish We are able to write a configuration file and run the server: java joseki.rdfserver <ConfigurationFile> 03/05/2004 Group: Hanh & Tinh 12

13 Publish a Model to Joseki Declare a model in etc/joseki.n3 (default) file Model from a RDF file (Text file) < a joseki:attachedmodel ; joseki:attachedmodel <file:data/messages.rdf>; joseki:hasqueryoperation joseki:bindingfetchclosure; joseki:hasqueryoperation joseki:bindingrdql; joseki:hasqueryoperation joseki:bindingget; joseki:namespaceprefix [ joseki:prefix "mm" ; joseki:nsuri < ]; rdfs:comment "A collection of messages" ;. 03/05/2004 Group: Hanh & Tinh 13

14 Publish a Model to Joseki Model from a MySQL database < maildb> a joseki:attachedmodel ; joseki:attachedmodel <jdbc:mysql://localhost/mailrdf?autoreconnect=true> ; joseki:user "root" ; joseki:password "" ; joseki:dbmodelname "MailMessages"" ; joseki:dbtype "MySQL"" ; joseki:dbdriver "com.mysql.jdbc.driver"" ; joseki:hasqueryoperation joseki:bindingrdql ; joseki:hasoperation joseki:bindingping ; joseki:hasoperation joseki:bindingquerymodel ; joseki:namespaceprefix [ joseki:prefix "mm" ; joseki:nsuri < ] ;. 03/05/2004 Group: Hanh & Tinh 14

15 Fetch a Model From Web browsers Using HTTP protocol From Java program Using built-in in method: HttpFetch fetch = New HttpFetch(modelURI, resourceuri); Model obj = fetch.exec(); 03/05/2004 Group: Hanh & Tinh 15

16 Query a Model From Web browsers Using following syntax: where: lang: : specify query language query: : specify query statement 03/05/2004 Group: Hanh & Tinh 16

17 Query a Model with Java A fragment code String host = argv[0]; String modeluri = " + host + "/" + argv[1]; String querystring = "SELECT?x,?sentname? sentname,,?from " + "WHERE (?x, <mm:sentname< mm:sentname>,?sentname? sentname), (?x, <mm:from< mm:from>,?from) " + "USING mm FOR < try { Query query = new Query(queryString) ) ; QueryExecution qe = new = new QueryEngineHTTP(query, modeluri) ) ; } // Get results and print them out QueryResults results = qe.exec() ; for ( Iterator iter = results ; iter.hasnext() ; ) { // print result out } results.close() ; 03/05/2004 Group: Hanh & Tinh 17

18 Demo

19 Prototype Architecture Client INTERFACE / ANALYSIS / EVENT-HANDLER Monitoring Module Javamail API The Internet Network API (Joseki) Query RDQL (other QL) Inference OWL API / RDFS API RDF API (Writer/Reader) Jena Semantic Web Framework In-Memory RDBMS Stores POP3/IMAP 03/05/2004 Mail Server 19

20 Next Step Creating Inference layer for this prototype supporting for semantic querying. Building user interface for the system. 03/05/2004 Group: Hanh & Tinh 20

21 Thank you!