The project is funded as a BRIDGE 2 project by the
It has started in May 2010.
In this project we aim at developing and implementing ontology-driven methods for domain-specific information extraction and retrieval.
Publically available resources developed (partly or in full) as part of this project:
- GATE plugin AppDoc - automatically create documentation for GATE pipeliens, JAPE and other components
- GATE plugin VirtualCorpus - make a directory of documents directly accessible as a corpus
- GATE plugin StringAnnotation - provides an improved and extended version of the GATE list gazetteer and a new processing resource for using Java regular expressions for annotating the document content
- GATE plugin VirtualDocuments - several processing resources for generating and processing "virtual views" of a document based on a various annotation features and/or the original document text
- GATE plugin JapeUtils - this plugin provides an extended version of JAPE that has several additional matching constraints pre-loaded (valueref, coextensive, startswith, endswith). In addition, it also provides an extensive library of methods to simplify the Java language rule right-hand-side code.
- GATE plugin Modular Pipelines - allows to use individual GATE applications/pipelines as modules and combining them by nesting them while still keeping their original identity. With this plugin it is possible to continue developing and updating each module separately without changing or updating any containing pipeline.
The plugins AppDoc, VirtualCorpus, VirtualDocuments, ModularPipelines, StringAnnotation and JapeUtils are mature enough to be in the list of plugins that is made available through OFAI's GATE plugin repository (http://www.ofai.at/~johann.petrak/GATE/gate-update-site.xml) which is preconfigured in the GATE update manager.
Resources that got superseded or merged/changed to new resources during the project:
- GATE plugin JAPEBackRefs - extend JAPE so that rules can match against previously matched values, similar to back-references in regular expressions. This plugin got merged into the JapeUtils plugin.
- GATE plugin JapeParms - allows to use a templating mechanism to parametrize a JAPE transducer grammar in order to make it portable and re-usable. This plugin got merged into the JapeUtils plugin.