Download presentation
Presentation is loading. Please wait.
1
Laurea Magistrale in Scienze di Internet
Project Andrea Nuzzolese Laurea Magistrale in Scienze di Internet KMDM a.a. 2010/2011 Bologna, 18 Aprile 2011
2
General rules The project has to be developed by groups
max 3 persons per group (2 are recommended) each group has a name Project and group validation the group must be communicated to the Google group three weeks by the deadline selected for delivery E.g. a group that wants to deliver on June 15th has to be communicated before May 25th the for communicating the group must contain group name members’ names description of the project (max 700 words) the proposed project and the group will be validated within three days If the project is rejected it has to be revised and re-submitted it is highly recommended to interact on the Google group during the development of the project proposal deadlines for project delivery June 15 access to oral exam from June 22nd July 13 access to oral exam from July 21st
3
Basic objectives of the project
understanding how to manage knowledge problems related to knowledge management and how to solve them learning how to use semantic technologies for KM realizing a Web-based platform which allows semantic mash-up
4
Semantic Mash-up a semantic mash-up aggregates data from heterogeneous (and possibly non semantic) sources it aggregates data/information to the aim of delivering knowledge to a user/application it connects data and make sense of them based on some knowledge-oriented criteria
6
Types of projects LOD search engine Relation finder
7
LOD search engine input: keywords
output: RDF graph describing the entity as from several LOD datasets (presented through user-friendly interface) user interaction includes keyword-based search on linked data pattern-based filtering pattern-based aggregation and transformation pattern-based presentation example: SIG.MA Chiarisci che Sig.ma non fa nessuna di quelle cose e che nel progetto gli aspetti di interazione menzionati sono tutti obbligatori
9
Relation finder input: text(+), keywords
output: RDF graph describing entities mentioned in the text and their relationships only relevant relationships based on pattern-lenses (presented through user-friendly interface) example: RelFinder Perché vuoi mettere il testo obbligatorio? Abbiamo detto che possono farlo dal testo oppure inserendo un numero di keyword > 1 Al limite possiamo dire I progetti che permettono l’inserimento/upload di un testo verranno valutati con un +1
11
Technology to use Javascript Apache Stanbol Java (maybe)
a modular software stack and reusable set of components for semantic content management Java (maybe)
12
Why Apache Stanbol Apache Stanbol provides HTTP RESTful services which allow Named entity recognition and linkage Entity caching and dataset configuration RDF graph refactoring Knowledge Pattern management …much more, but too much for your project
13
How to Build Apache Stanbol
Download the source code from tha Apache SVN Build the code using Maven $mvn –DskipTests install
14
How to run Stanbol go to stanbolroot/launchers/lite/target and launch
$java –jar –Xmx512m org.apache.stanbol.enhancer.launchers.full-0.9-SNAPSHOT.jar open your browser at
16
All the components are accessible via HTTP REST services
Components to use Enhancer recognizes entities in text documents and links those entities to LOD resources Entity Hub allows to configure a network of trusted datasets cache entities Refactor RDF graph transformation allows semantic harmonization OntoNet manage your ontology network All the components are accessible via HTTP REST services
17
Stanbol Enhancer Input Output
“Bob Marley was a famous musician from Jamaica.” Output GeoNames DBpedia
18
How the Enhancer works in: unstructured content (i.e. full text) +
the enhanced RDF meta-data (initially empty) Enhancement Engine 1 Enhancement Engine 2 Job Manager … out: the RDF meta-data enhanced from the unstructured content Enhancement Engine N
19
How the Enhancer works in: unstructured content (i.e. full text) +
the enhanced RDF meta-data (out of the previous engine) Enhancement Engine 1 Enhancement Engine 2 Job Manager … out: the RDF meta-data enhanced from the unstructured content + previous metadata Enhancement Engine N
20
How the Enhancer works in: unstructured content (i.e. full text) +
the enhanced RDF meta-data (out of the previous engine) Enhancement Engine 1 Enhancement Engine 2 Job Manager … out: the RDF meta-data enhanced from the unstructured content + previous metadata Enhancement Engine N
21
Enhancer’s REST services (cURL)
standard request in cURL $ curl -X POST -H "Content-type: text/plain" --data "Paris is a beautiful city." you can force a specific RDF serialization scheme by setting the "Accept" HTTP header $ curl -X POST -H "Content-type: text/plain" -H "Accept: application/rdf+xml" --data "Paris is a beautiful city." you can upload the content of a file $ curl -X POST -H "Content-type: text/plain" -H "Accept: text/rdf+nt"
22
Enhancer’s REST services (cURL)
store the enhanced RDF graph curl -i -X POST -H "Content-Type:text/plain” --data "The Stanbol enhancer can detect famous cities such as Paris.” HTTP/ CreatedLocation: 0Server: Jetty(6.1.x) store the enhanced RDF graph from file curl -i -X PUT -H "Content-Type:text/plain" -T $file
23
Enhancer’s REST services (Javascript)
$.ajax({ type: "POST", url: " data: "Paris is a beautiful city.", dataType: ”application/rdf+xml", success: function(result) { … }, error: function(result) { } }); RDF: application/rdf+xml RDF JSON: application/rdf+json JSON LD: application/json text plain: text/plain turtle: text/turtle N triple: text/rdf+nt N3: text/rdf+n3
24
EntityHub’s REST services (cURL)
if you don’t know the dataset ID $ curl -X GET if you want the output as RDF/XML curl -X GET -H "Accept: application/rdf+xml" if you know the dataset ID curl -X GET
25
EntityHub’s REST services (Javascript)
$.ajax({ type: ”GET", url: " id= dataType: ”application/rdf+json", success: function(result) { … }, error: function(result) { } }); RDF: application/rdf+xml RDF JSON: application/rdf+json JSON LD: application/json text plain: text/plain turtle: text/turtle N triple: text/rdf+nt N3: text/rdf+n3
26
How to configure the Entity Hub
go to username: admin password: admin look in the table for “Apache Stanbol Entityhub Referenced Site Configuration” add a new site configure an existing site
27
How to configure the Entity Hub
Similar presentations
© 2024 SlidePlayer.com Inc.
All rights reserved.