GERBIL 1.1.0

Features:

• Added a better error handling to the DataIDGenerator. Now this class won't throw a NullpointerException if an experiment type or a matching is missing. refs #73
• Added the mapping for the google analytics file. refs #72
• Enhanced the performance problem of the database adapter if it has to load the lates experiments.
• Added a way to use always the same color for a single adapter. Moved the spider diagram java script files into the js folder.
• Ordered experiment types, annotators and datasets alphabetically.
• Added the version of the program as property to the properties file.
  • Since the version is set automatically by maven, the property files have been moved into an own folder.Thus, the filtering done by maven will not influence other resource files. refs issue #56

Bug Fixes:

• Added new URL of gerbil_data.zip
  •  Added the URL of the gerbil_data zip file pointing to version 1.0.0 of GERBIL on GitHub
• Fixed macrof1 description in json-ld format
• Added the missing C2W experiment type to the GERBIL vocabulary.
• Changed the mapping classes.
  • They are adding a suffix to uploaded datasets and NIF based web services that have been added using the configuration screen.
• Added the missing parent pom for jena.
• Added the missing jena-iri jar to the local repo.
• Added a clean shutdown of the database.
• Added the missing close method to the SimpleLoggingDAO.
• Added the creation of parent folder for cache files if they are missing.
  • Added a JUnit test for the SingletonWikipediaApi class. refs issue #70
• Added the logo and icon to the project (instead loading it from another URL).

GERBIL 1.0.0

This version of GERBIL provides a running web service enabling the evaluation, benchmarking, publishing and archiving of entity annotation experiments. GERBIL is a community effort.

Currently, we support the following annotators

• Wikipedia Miner
• Spotlight
• TagMe 2
• KEA
• WAT
• AGDISTIS
• Babelfy
• NERD-ML
• Dexter
• NIF-based Annotator (this adapter can be used for any NIF based web service)

and the following datasets

• ACE2004
• AQUAINT
• MSNBC
• IITB
• Meij
• AIDA/CoNLL
• N3 - reuters-128
• N3 - rss-500
• KORE50
• Spotlight Corpus
• Microposts.

Furthermore, GERBIL supports six experiment types, three matchings and 6 metrics. You will never again have to write your own entity annotation benchmark system. Just write a wrapper for your annotator or dataset and let it run against state-of-the-art systems and corpora.

Join us!

Find more information at our or at http://gerbil.aksw.org.

If you have any trouble or suggestions, please file an GitHub issue.

Thank you a lot!
The GERBIL team

Version 1.0.0 of the gerbil.nif.transfer library

This is the first release of the gerbil.nif.transfer library. This simple lib translates a document from java into NIF, transfers it to a webservice or client and retranslates it into a java object. It is part of the GERBIL project and can be used by others to implement a NIF based web service for their own annotators.