OPENNLP-989: Fix validation of CONT after START with different type This closes #126
Project: http://git-wip-us.apache.org/repos/asf/opennlp/repo Commit: http://git-wip-us.apache.org/repos/asf/opennlp/commit/ed246d8a Tree: http://git-wip-us.apache.org/repos/asf/opennlp/tree/ed246d8a Diff: http://git-wip-us.apache.org/repos/asf/opennlp/diff/ed246d8a Branch: refs/heads/parser_regression Commit: ed246d8a65a3f6ebe6e3ffae257f7fe6571bfdb8 Parents: d77b5e2 Author: Peter Thygesen <[email protected]> Authored: Fri Feb 17 15:17:13 2017 +0100 Committer: Jörn Kottmann <[email protected]> Committed: Thu Apr 20 12:40:22 2017 +0200 ---------------------------------------------------------------------- .../opennlp/tools/namefind/NameFinderSequenceValidator.java | 5 +++-- .../test/java/opennlp/tools/eval/Conll02NameFinderEval.java | 8 ++++---- .../tools/namefind/NameFinderSequenceValidatorTest.java | 2 -- 3 files changed, 7 insertions(+), 8 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/opennlp/blob/ed246d8a/opennlp-tools/src/main/java/opennlp/tools/namefind/NameFinderSequenceValidator.java ---------------------------------------------------------------------- diff --git a/opennlp-tools/src/main/java/opennlp/tools/namefind/NameFinderSequenceValidator.java b/opennlp-tools/src/main/java/opennlp/tools/namefind/NameFinderSequenceValidator.java index d42e8c5..5143468 100644 --- a/opennlp-tools/src/main/java/opennlp/tools/namefind/NameFinderSequenceValidator.java +++ b/opennlp-tools/src/main/java/opennlp/tools/namefind/NameFinderSequenceValidator.java @@ -35,8 +35,9 @@ public class NameFinderSequenceValidator implements return false; } else if (outcomesSequence[li].endsWith(NameFinderME.OTHER)) { return false; - } else if (outcomesSequence[li].endsWith(NameFinderME.CONTINUE)) { - // if it is continue, we have to check if previous match was of the same type + } else if (outcomesSequence[li].endsWith(NameFinderME.CONTINUE) || + outcomesSequence[li].endsWith(NameFinderME.START)) { + // if it is continue or start, we have to check if previous match was of the same type String previousNameType = NameFinderME.extractNameType(outcomesSequence[li]); String nameType = NameFinderME.extractNameType(outcome); if (previousNameType != null || nameType != null ) { http://git-wip-us.apache.org/repos/asf/opennlp/blob/ed246d8a/opennlp-tools/src/test/java/opennlp/tools/eval/Conll02NameFinderEval.java ---------------------------------------------------------------------- diff --git a/opennlp-tools/src/test/java/opennlp/tools/eval/Conll02NameFinderEval.java b/opennlp-tools/src/test/java/opennlp/tools/eval/Conll02NameFinderEval.java index d1a71cf..abe53aa 100644 --- a/opennlp-tools/src/test/java/opennlp/tools/eval/Conll02NameFinderEval.java +++ b/opennlp-tools/src/test/java/opennlp/tools/eval/Conll02NameFinderEval.java @@ -287,9 +287,9 @@ public class Conll02NameFinderEval { TokenNameFinderModel maxentModel = train(dutchTrainingFile, LANGUAGE.NL, params, combinedType); - eval(maxentModel, dutchTestAFile, LANGUAGE.NL, combinedType, 0.6728164867517175d); + eval(maxentModel, dutchTestAFile, LANGUAGE.NL, combinedType, 0.6673209028459275d); - eval(maxentModel, dutchTestBFile, LANGUAGE.NL, combinedType, 0.6985893619774816d); + eval(maxentModel, dutchTestBFile, LANGUAGE.NL, combinedType, 0.6984085910208306d); } @Test @@ -507,9 +507,9 @@ public class Conll02NameFinderEval { TokenNameFinderModel maxentModel = train(spanishTrainingFile, LANGUAGE.ES, params, combinedType); - eval(maxentModel, spanishTestAFile, LANGUAGE.ES, combinedType, 0.706765154179857d); + eval(maxentModel, spanishTestAFile, LANGUAGE.ES, combinedType, 0.707400023454908d); - eval(maxentModel, spanishTestBFile, LANGUAGE.ES, combinedType, 0.7583580194667795d); + eval(maxentModel, spanishTestBFile, LANGUAGE.ES, combinedType, 0.7576868829337094d); } @Test http://git-wip-us.apache.org/repos/asf/opennlp/blob/ed246d8a/opennlp-tools/src/test/java/opennlp/tools/namefind/NameFinderSequenceValidatorTest.java ---------------------------------------------------------------------- diff --git a/opennlp-tools/src/test/java/opennlp/tools/namefind/NameFinderSequenceValidatorTest.java b/opennlp-tools/src/test/java/opennlp/tools/namefind/NameFinderSequenceValidatorTest.java index 35752c1..1b2f6ed 100644 --- a/opennlp-tools/src/test/java/opennlp/tools/namefind/NameFinderSequenceValidatorTest.java +++ b/opennlp-tools/src/test/java/opennlp/tools/namefind/NameFinderSequenceValidatorTest.java @@ -17,7 +17,6 @@ package opennlp.tools.namefind; import org.junit.Assert; -import org.junit.Ignore; import org.junit.Test; /** @@ -55,7 +54,6 @@ public class NameFinderSequenceValidatorTest { } - @Ignore @Test public void testContinueAfterStartAndNotSameType() {
