[OPENNLP-549] Inconsistent handling of lower-/upper- case POS tags in the JWNLDictionary.getLemmas method

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: tools-1.5.2-incubating
• Fix Version/s: tools-1.5.3
• Component/s: Coref
• Assignee: Aliaksandr Autayeu
• Reporter: Gleb Alexeyev
• Created: 16/Nov/12 11:05
• Updated: 19/Dec/12 22:19
• Resolved: 22/Nov/12 17:48

Description

JWNLDictionary.getLemmas method has a code that maps string-valued POS tags to net.didion.jwnl.data.POS instances, and this code clearly has a bug (for example, it maps tags starting with "N" both to POS.NOUN and POS.VERB). Please see the patch below (the patch assumes Penn TreeBank tags).

Index: src/main/java/opennlp/tools/coref/mention/JWNLDictionary.java
===================================================================
— src/main/java/opennlp/tools/coref/mention/JWNLDictionary.java (revision 1410284)
+++ src/main/java/opennlp/tools/coref/mention/JWNLDictionary.java (working copy)
@@ -84,10 +84,10 @@
if (tag.startsWith("N") || tag.startsWith("n"))

{ pos = POS.NOUN; }

• else if (tag.startsWith("N") || tag.startsWith("v")) {
  + else if (tag.startsWith("V") || tag.startsWith("v")) { pos = POS.VERB; }
• else if (tag.startsWith("J") || tag.startsWith("a")) {
  + else if (tag.startsWith("J") || tag.startsWith("j")) { pos = POS.ADJECTIVE; }

  else if (tag.startsWith("R") || tag.startsWith("r")) {