From: Roman C. <rom...@gm...> - 2011-02-25 09:52:17
|
Hi, Unless it was updated, the KEA from Gate contrib will not work - I don't remember all the changes that I had to do, but they were many. Also the old plugin was using the old KEA and the new KEA is so much better that you really don't want the old version. You can get the new ported version of the new KEA plugin at: http://www.roman-chyla.net/temp/kea.tgz (it includes all dependencies and test documents with taxonomies, 40MB) -- If you use it, please add CERN to the copyright. I once (actually twice) asked if there was interest of getting the new version into GATE, but got no reply, but of course, feel free to do so if that might be helpful to others now. Best, Roman On Thu, Feb 24, 2011 at 8:58 PM, VERMILLION, Bonnie <YVE...@lm...> wrote: > Turns out the .key and .txt extensions are needed in the KEA Corpus > Importer. The importer now loads the .txt file after merging with the .key > file uploaded .txt file and is tagged with keyphrases. Subsequently, the KEY > Corpus Pipeline was run against the .txt file within the KEA Corpus that > gets loaded with the .txt file and with trainingMode set to true in the > attached KEA Keyphrase Extractor processing resource, the application runs. > However, now when the KEA Keyphrase Extractor is added to a normal corpus > pipeline and the trainingMode set to false, the following error occurs: > > > > (the documentation does not indicate setting an output instance format - > does anyone know what the solution is?) > > gate.creole.ExecutionException: java.lang.NullPointerException: No output > instance format defined > > at gate.creole.kea.Kea.finishTraining(Kea.java:438) > > at gate.creole.kea.Kea.execute(Kea.java:147) > > at > gate.util.Benchmark.executeWithBenchmarking(Benchmark.java:299) > > at > gate.creole.SerialController.runComponent(SerialController.java:222) > > at > gate.creole.SerialController.executeImpl(SerialController.java:153) > > at > gate.creole.SerialAnalyserController.executeImpl(SerialAnalyserController.java:115) > > at > gate.creole.AbstractController.execute(AbstractController.java:62) > > at > gate.util.Benchmark.executeWithBenchmarking(Benchmark.java:299) > > at > gate.gui.SerialControllerEditor$RunAction$1.run(SerialControllerEditor.java:1552) > > at java.lang.Thread.run(Unknown Source) > > Caused by: java.lang.NullPointerException: No output instance format defined > > at weka.filters.Filter.numPendingOutput(Unknown Source) > > at kea.KEAFilter.batchFinished(KEAFilter.java:626) > > at gate.creole.kea.Kea.finishTraining(Kea.java:436) > > ... 9 more > > Thank you, > > Bonnie Vermillion > > > > From: VERMILLION, Bonnie [mailto:YVE...@lm...] > Sent: Thursday, February 24, 2011 1:34 PM > To: gat...@li... > Subject: Re: [gate-users] KEA Extractor Processing Resource > > > > I removed the default file extensions from the KEA Corpus Importer and now > the documents load. But after creating a Corpus Pipeline and running the > application using the KEA Processing Resource pointed to the Corpus > containing the .key and .txt files , this error displays ( there are 275 > phrases in the .key file): > > > > No keyphrases in document: 4140_key_phrases.key > > gate.creole.ExecutionException: java.lang.NullPointerException > > at gate.creole.kea.Kea.execute(Kea.java:141) > > at > gate.util.Benchmark.executeWithBenchmarking(Benchmark.java:299) > > at > gate.creole.SerialController.runComponent(SerialController.java:222) > > at > gate.creole.SerialController.executeImpl(SerialController.java:153) > > at > gate.creole.SerialAnalyserController.executeImpl(SerialAnalyserController.java:115) > > at > gate.creole.AbstractController.execute(AbstractController.java:62) > > at > gate.util.Benchmark.executeWithBenchmarking(Benchmark.java:299) > > at > gate.gui.SerialControllerEditor$RunAction$1.run(SerialControllerEditor.java:1552) > > at java.lang.Thread.run(Unknown Source) > > Caused by: java.lang.NullPointerException > > at kea.KEAFilter.convertInstance(KEAFilter.java:915) > > at kea.KEAFilter.input(KEAFilter.java:596) > > at gate.creole.kea.Kea.execute(Kea.java:139) > > ... 8 more > > > > Thank you, > Bonnie Vermillion > > > > From: VERMILLION, Bonnie [mailto:YVE...@lm...] > Sent: Wednesday, February 23, 2011 4:28 PM > To: gat...@li... > Subject: [gate-users] KEA Extractor Processing Resource > > > > I’m getting the error: there were problems creating keyphrases for both the > .txt and .key files when attempting to import these files with the KEA > Corpus importer in order to perform annotation based on a keyword list and > then apply this list to the document to be trained. > > > > Thank you in advance > > Bonnie Vermillion > > ------------------------------------------------------------------------------ > Free Software Download: Index, Search & Analyze Logs and other IT data in > Real-Time with Splunk. Collect, index and harness all the fast moving IT > data > generated by your applications, servers and devices whether physical, > virtual > or in the cloud. Deliver compliance at lower cost and gain new business > insights. http://p.sf.net/sfu/splunk-dev2dev > _______________________________________________ > GATE-users mailing list > GAT...@li... > https://lists.sourceforge.net/lists/listinfo/gate-users > > |