Skip to content

liangkai/AutoSPARQL

 
 

Repository files navigation

AutoSPARQL is a graphical user interface, which allows to answer natural
language queries over RDF knowledge bases. It is based on algorithms
implemented in the DL-Learner Semantic Web machine learning framework.
Currently, there are two user interfaces 1.) a full-fledged UI closely
aligned to the unterlying algorithms and 2.) a simple UI which should be
easy to use for anyone.

USAGE
You need to clone the git repository and have Maven 3 or higher and Java 7 or higher installed on your machine. Then you need to run "./compile" and then "./run" on Linux or "compile.bat" and then "run.bat" on Windows (not tested on Windows, please create an issue if the Windows batch needs rework).
There is also "./createcompillelog" for easier error reporting.

Manually: Do "mvn install -N" in the folders in that order: autosparql, commons, algorithm-tbsl. Then go into autosparql-tbsl and run "mvn jetty:run".

Then, go into your browser and access "http://localhost:8080/autosparql-lite".

REMARKS
Symmetric properties now properly work again.

HOW TO ADD YOUR OWN DATASET
Using your own datasource instead of DBpedia or Oxford is nontrivial.
It needs several days of work in addition to the time needed to familiarize yourself with the code base.

You need to:

- fork AutoSPARQL TBSL
- add your own domain dependent lexicon as an LTAG grammar at algorithm-tbsl/src/main/resources/tbsl/lexicon (see http://pub.uni-bielefeld.de/publication/2002961 and http://pub.uni-bielefeld.de/publication/2278529 as well as the existing files in that folder)
- add your knowledge base:
 1. as a local model to algorithm-tbsl/src/main/resources/models/yourmodel (preferred for small knowledge bases as it is much faster and more reliable)
 2. as a SPARQL (version 1.0 is enough) endpoint along with a SOLR server instance
- create a singleton for your knowledge base (see package org.aksw.autosparql.tbsl.algorithm.knowledgebase)
- extend org.aksw.autosparql.tbsl.algorithm.learning.TBSL with TbslYourKnowledgeBase
- finally, in case the dataset isn't a private model, please do a pull request so that it can be integrated in the main project

-------------------------------------------------------
Package                                   Maven Module
org.aksw.autosparql.tbsl.algorithm        algorithm-tbsl
org.aksw.autosparql.commons               commons
org.aksw.autosparql.tbsl.gui.vaadin       autosparql-tbsl
-------------------------------------------------------

About

AutoSPARQL allows to create SPARQL queries over RDF knowledge bases from natural language with low effort.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Java 64.4%
  • Web Ontology Language 23.8%
  • CSS 8.1%
  • Lex 1.9%
  • JavaScript 1.0%
  • Makefile 0.6%
  • Other 0.2%