fix annotation head for upos

INL · Nov 18, 2024 · 6352c33 · 6352c33
1 parent 604cb31
commit 6352c33
Show file tree

Hide file tree

Showing 2 changed files with 42 additions and 7 deletions.
diff --git a/server/src/main/kotlin/org/ivdnt/galahad/data/layer/Term.kt b/server/src/main/kotlin/org/ivdnt/galahad/data/layer/Term.kt
@@ -85,8 +85,8 @@ data class Term(
                 return annotation.split('-')[1]
             }
         }
-        // for POS
-        else if (annotationType == AnnotationType.POS) {
+        // for POS & UPOS
+        else if (listOf(AnnotationType.POS, AnnotationType.UPOS).contains(annotationType)) {
             return if (isMulti(annotationType)) {
                 // Split on + and transform each part
                 annotation.split("+").map { singlePosToHead(it) }.joinToString("+")

diff --git a/server/src/test/kotlin/org/ivdnt/galahad/data/layer/TermTest.kt b/server/src/test/kotlin/org/ivdnt/galahad/data/layer/TermTest.kt
@@ -14,25 +14,60 @@ class TermTest {
         )
     )
 
-    private val multiPos = Term("", "PD(type=art)+NOU-C(num=sg)", mutableListOf())
-    private val singlePos = Term("", "NOU-C(num=sg)", mutableListOf())
-    private val headOnlyPos = Term("", "NOU-C", mutableListOf())
+    // upos
+    private val singleUPos = Term(
+        mapOf(
+            AnnotationType.UPOS to "NOU-C(num=sg)"
+        ), mutableListOf()
+    )
+
+    // ner
+    private val singleNer = Term(
+        mapOf(
+            AnnotationType.NER to "B-LOC"
+        ), mutableListOf()
+    )
+
+    // pos
+    private val singlePos = Term(
+        mapOf(
+            AnnotationType.POS to "NOU-C(num=sg)"
+        ), mutableListOf()
+    )
 
+    private val multiPos = Term(
+        mapOf(
+            AnnotationType.POS to "PD(type=art)+NOU-C(num=sg)"
+        ), mutableListOf()
+    )
+
+    private val headOnlyPos = Term(
+        mapOf(
+            AnnotationType.POS to "NOU-C"
+        ), mutableListOf()
+    )
 
 
     @Test
     fun `PoS head and features`() {
         // single pos
         assertEquals("NOU-C", singlePos.annotationHead(AnnotationType.POS))
-        assertEquals("num=sg", Term.features(singlePos.pos))
+        assertEquals("num=sg", Term.features(singlePos.annotations.pos))
+
+        // upos
+        assertEquals("NOU-C", singleUPos.annotationHead(AnnotationType.UPOS))
+        assertEquals("num=sg", Term.features(singleUPos.annotations.upos))
+
+        // ner
+        assertEquals("LOC", singleNer.annotationHead(AnnotationType.NER))
 
         // multi pos
         assertEquals("PD+NOU-C", multiPos.annotationHead(AnnotationType.POS))
         // no features for now...
 
         // head only pos
         assertEquals("NOU-C", headOnlyPos.annotationHead(AnnotationType.POS))
-        assertEquals(null, Term.features(headOnlyPos.pos))
+        assertEquals(null, Term.features(headOnlyPos.annotations.pos))
     }
 
     @Test