Thank you, James I don't know the best way to solve it. For now I added to DetailedFMeasureListener a method called createReport(Locale), where it is possible to specify the desired locale. It is overloaded by createReport(), that uses the default locale. The toString will use createReport() with the default locale.
On Fri, Aug 19, 2011 at 10:34 PM, James Kosin <[email protected]> wrote: > William, > > The format for the % seems to be local specific, the tests still fail for > me with the only difference of a '.' character instead of the ','. If this > is correct, then the validation should also be converted to the same local > format. > > James > > > On 8/19/2011 5:38 AM, [email protected] wrote: > >> Thanks. >> >> On Fri, Aug 19, 2011 at 1:57 AM, James Kosin<[email protected]> >> wrote: >> >> Never mind, >>> >>> New error now with latest. >>> ---- >>> Failed tests: >>> testEvaluator(opennlp.tools.****chunker.**** >>> ChunkerDetailedFMeasureListene****rTest): >>> >>> null expected:<[processed 3 samples with 32 entities; found: 35 entities; >>> correct: 28. >>> TOTAL: precision: 80,00%; recall: 87,50%; F1: 83,58%. >>> NP: precision: 90,00%; recall: 94,74%; F1: 92,31%. [target: >>> 19; tp: 18; fp: 2] >>> VP: precision: 75,00%; recall: 75,00%; F1: 75,00%. [target: >>> 8; tp: 6; fp: 2] >>> PP: precision: 57,14%; recall: 100,00%; F1: 72,73%. [target: >>> 4; tp: 4; fp: 3] >>> SBAR: precision: 0,00%; recall: 0,00%; F1: 0,]00%. [target: >>> 1; t...> but was:<[Evaluated 3 samples with 32 entities; found: 35 >>> entities; >>> correct: 28. >>> TOTAL: precision: 80.00%; recall: 87.50%; F1: 83.58%. >>> NP: precision: 90.00%; recall: 94.74%; F1: 92.31%. >>> [target: >>> 19; tp: 18; fp: 2] >>> VP: precision: 75.00%; recall: 75.00%; F1: 75.00%. >>> [target: >>> 8; tp: 6; fp: 2] >>> PP: precision: 57.14%; recall: 100.00%; F1: 72.73%. >>> [target: >>> 4; tp: 4; fp: 3] >>> SBAR: precision: 0.00%; recall: 0.00%; F1: 0.]00%. >>> [target: 1; t...> >>> >>> Tests run: 254, Failures: 1, Errors: 0, Skipped: 0 >>> >>> ---- >>> I'll check back later... >>> James >>> >>> On 8/19/2011 12:44 AM, James Kosin wrote: >>> >>> William& Jorn, >>>> >>>> I'm getting null pointer exceptions now when running the evaluator for >>>> the >>>> name finder. >>>> The "Testing..." comment line is from the batch file I'm using... >>>> everything after that is from the application run with "%JCMD% >>>> TokenNameFinderEvaluator -encoding utf8 -model enNameFinder.model -data >>>> testa.txt" >>>> >>>> ---- >>>> Testing... on eng.testa >>>> Loading Token Name Finder model ... done (0.419s) >>>> Exception in thread "main" java.lang.NullPointerException >>>> at opennlp.tools.util.eval.****Evaluator.**** >>>> notifyMissclassified(** >>>> Evaluator.java >>>> :114) >>>> at opennlp.tools.util.eval.****Evaluator.evaluateSample(** >>>> Evaluator.java:70) >>>> at opennlp.tools.util.eval.****Evaluator.evaluate(Evaluator.**** >>>> java:86) >>>> at opennlp.tools.cmdline.****namefind.** >>>> TokenNameFinderEvaluatorTool.****run(Token >>>> NameFinderEvaluatorTool.java:****104) >>>> at opennlp.tools.cmdline.CLI.****main(CLI.java:191) >>>> current: 1.0 sent/s avg: 1.0 sent/s total: 1 sent >>>> current: 0.0 sent/s avg: 0.5 sent/s total: 1 sent >>>> current: 0.0 sent/s avg: 0.3 sent/s total: 1 sent >>>> current: 0.0 sent/s avg: 0.2 sent/s total: 1 sent >>>> current: 0.0 sent/s avg: 0.2 sent/s total: 1 sent >>>> current: 0.0 sent/s avg: 0.2 sent/s total: 1 sent >>>> current: 0.0 sent/s avg: 0.1 sent/s total: 1 sent >>>> current: 0.0 sent/s avg: 0.1 sent/s total: 1 sent >>>> current: 0.0 sent/s avg: 0.1 sent/s total: 1 sent >>>> >>>> ---- >>>> James >>>> >>>> >>> >
