[uk] improve adj/noun inflection suggestions

languagetool-org · Dec 23, 2016 · 54d35cf · 54d35cf
1 parent bd11e23
commit 54d35cf
Show file tree

Hide file tree

Showing 3 changed files with 28 additions and 17 deletions.
diff --git a/...agetool-language-modules/uk/src/main/java/org/languagetool/rules/uk/InflectionHelper.java b/...agetool-language-modules/uk/src/main/java/org/languagetool/rules/uk/InflectionHelper.java
@@ -104,7 +104,10 @@ static List<Inflection> getAdjInflections(List<AnalyzedToken> adjTokenReadings)
         animTag = matcher.group(3).substring(2);	// :rinanim/:ranim
       }
 
-      masterInflections.add(new Inflection(gen, vidm, animTag));
+      Inflection inflection = new Inflection(gen, vidm, animTag);
+      if( ! masterInflections.contains(inflection) ) {
+        masterInflections.add(inflection);
+      }
     }
     return masterInflections;
   }
@@ -125,7 +128,10 @@ static List<Inflection> getNounInflections(List<AnalyzedToken> nounTokenReadings
       String vidm = matcher.group(3);
       String animTag = matcher.group(1);
 
-      slaveInflections.add(new Inflection(gen, vidm, animTag));
+      Inflection inflection = new Inflection(gen, vidm, animTag);
+      if( ! slaveInflections.contains(inflection) ) {
+        slaveInflections.add(inflection);
+      }
     }
     return slaveInflections;
   }

diff --git a/...uage-modules/uk/src/main/java/org/languagetool/rules/uk/TokenInflectionAgreementRule.java b/...uage-modules/uk/src/main/java/org/languagetool/rules/uk/TokenInflectionAgreementRule.java
@@ -206,7 +206,7 @@ else if ( nounPosTag.equals(JLanguageTool.SENTENCE_END_TAGNAME)
             slaveTokenReadings.get(0).getToken() + ": " + slaveInflections+ " // " + slaveTokenReadings));
         }
 
-        String msg = String.format("Потенційна помилка: прикметник неузгоджений з іменником: \"%s\": [%s] і \"%s\": [%s]", 
+        String msg = String.format("Потенційна помилка: прикметник не узгоджений з іменником: \"%s\": [%s] і \"%s\": [%s]", 
             adjTokenReadings.get(0).getToken(), formatInflections(masterInflections, true),
             slaveTokenReadings.get(0).getToken(), formatInflections(slaveInflections, false));
 
@@ -230,7 +230,7 @@ else if ( nounPosTag.equals(JLanguageTool.SENTENCE_END_TAGNAME)
 
 
             if( ! adjInflection._case.equals("v_kly")
-                && (adjInflection._case.equals("p")
+                && (adjInflection.gender.equals("p")
                 || PosTagHelper.hasPosTagPart(slaveTokenReadings, genderTag)) ) {
               for(AnalyzedToken nounToken: slaveTokenReadings) {
 
@@ -251,29 +251,28 @@ else if ( nounPosTag.equals(JLanguageTool.SENTENCE_END_TAGNAME)
                 }
               }
             }
-
         }
 
         for (Inflection nounInflection : slaveInflections) {
           String genderTag = ":"+nounInflection.gender+":";
           String vidmTag = nounInflection._case;
 
-                if( nounInflection.animMatters() ) {
-                    vidmTag += ":r" + nounInflection.animTag;
-                }
+          if( nounInflection.animMatters() ) {
+            vidmTag += ":r" + nounInflection.animTag;
+          }
 
-              for(AnalyzedToken adjToken: adjTokenReadings) {
-                String newAdjTag = adjToken.getPOSTag().replaceFirst(":.:v_...(:r(in)?anim)?", genderTag + vidmTag);
+          for(AnalyzedToken adjToken: adjTokenReadings) {
+            String newAdjTag = adjToken.getPOSTag().replaceFirst(":.:v_...(:r(in)?anim)?", genderTag + vidmTag);
 
-                String[] synthesized = ukrainianSynthesizer.synthesize(adjToken, newAdjTag, false);
+            String[] synthesized = ukrainianSynthesizer.synthesize(adjToken, newAdjTag, false);
 
-                for (String s : synthesized) {
-                  String suggestion = s + " " + tokenReadings.getToken();
-                  if( ! suggestions.contains(suggestion) ) {
-                    suggestions.add(suggestion);
-                  }
-                }
+            for (String s : synthesized) {
+              String suggestion = s + " " + tokenReadings.getToken();
+              if( ! suggestions.contains(suggestion) ) {
+                suggestions.add(suggestion);
               }
+            }
+          }
 
         }
 

diff --git a/...-modules/uk/src/test/java/org/languagetool/rules/uk/TokenInflectionAgreementRuleTest.java b/...-modules/uk/src/test/java/org/languagetool/rules/uk/TokenInflectionAgreementRuleTest.java
@@ -71,6 +71,12 @@ public void testRule() throws IOException {
     assertEquals(1, rule.match(langTool.getAnalyzedSentence("кволого тюльпан")).length);
     assertEquals(1, rule.match(langTool.getAnalyzedSentence("цинічна винахідливості")).length);
 
+    RuleMatch[] matches0 = rule.match(langTool.getAnalyzedSentence("4 російських винищувача"));
+    assertEquals(1, matches0.length);
+    assertTrue("Message is wrong: " + matches0[0].getMessage(),
+        matches0[0].getMessage().contains("[ч.р.: родовий, знахідний]"));
+    assertEquals(Arrays.asList("російських винищувачів", "російських винищувачах", "російського винищувача"), matches0[0].getSuggestedReplacements());
+
     // from real examples
 
     // і-и