add VerbForm=Fin

brown-uk · Sep 24, 2024 · 20660eb · 20660eb
1 parent c45bbda
commit 20660eb
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 5 deletions.
diff --git a/src/main/groovy/ua/net/nlp/tools/tag/UdModule.groovy b/src/main/groovy/ua/net/nlp/tools/tag/UdModule.groovy
@@ -88,7 +88,11 @@ class UdModule {
             if( tkn.lemma in NEGATIVES ) {
                 udTags << "Polarity=Neg"
             }
-
+
+            if( tkn.tags =~ /^verb(?!.*(:inf|:impers))/ ) {
+                udTags << "VerbForm=Fin"
+            }
+
             addPluralGender(tkn, udTags)
 
             def misc = []

diff --git a/src/test/groovy/ua/net/nlp/tools/tag/TagTextVerticalOutputTest.groovy b/src/test/groovy/ua/net/nlp/tools/tag/TagTextVerticalOutputTest.groovy
@@ -92,7 +92,7 @@ class TagTextVerticalOutputTest {
     public void testTxtFormatWithUD() {
         tagText.setOptions(new TagOptions(outputFormat: OutputFormat.conllu, semanticTags: true))
 
-        def text = "А треба далі воно - озеро Світязь яке я..."
+        def text = "А треба далі воно - озеро Світязь де я затримався..."
         TagResult tagged = tagText.tagText(text)
 
         def expected =
@@ -105,9 +105,10 @@ class TagTextVerticalOutputTest {
 5 - - PUNCT punct _ _ _ _
 6 озеро озеро NOUN noun:inanim:n:v_naz Animacy=Inan|Case=Nom|Gender=Neut|Number=Sing _ _ _
 7 Світязь Світязь PROPN noun:inanim:m:v_naz:prop:geo:xp1 Animacy=Inan|Case=Nom|Gender=Masc|NameType=Geo|Number=Sing _ _ SemTags=1:conc:loc
-8 яке який ADJ adj:n:v_naz:&pron:int:rel:def Case=Nom|Gender=Neut|Number=Sing|PronType=Int|PronType=Rel _ _ _
-9 я я PRON noun:anim:s:v_naz:&pron:pers:1 Animacy=Anim|Case=Nom|Number=Sing|Person=1|PronType=Prs _ _ SemTags=1:conc:hum:deictic|SpaceAfter=No
-10 ... ... PUNCT punct _ _ _ _
+8 де де ADV adv:&pron:int:rel PronType=Int|PronType=Rel _ _ _
+9 я я PRON noun:anim:s:v_naz:&pron:pers:1 Animacy=Anim|Case=Nom|Number=Sing|Person=1|PronType=Prs _ _ SemTags=1:conc:hum:deictic
+10 затримався затриматися VERB verb:rev:perf:past:m Aspect=Perf|Gender=Masc|Mood=Ind|Number=Sing|Reflex=Yes|Tense=Past|VerbForm=Fin _ _ SpaceAfter=No
+11 ... ... PUNCT punct _ _ _ _
 """.toString()
 
         assertEquals expected, adjustResult(tagged.tagged)