OPENNLP-1125: Change misclassified report layout for LanguageDetector

Author: wcolen

opennlp-tools/src/main/java/opennlp/tools/cmdline/langdetect/LanguageDetectorEvaluationErrorListener.java

@@ -44,11 +44,18 @@ public LanguageDetectorEvaluationErrorListener() {
    */
   public LanguageDetectorEvaluationErrorListener(OutputStream outputStream) {
     super(outputStream);
+    printStream.println("Expected\tPredicted\tContext");
   }
 
   @Override
   public void missclassified(LanguageSample reference, LanguageSample prediction) {
     printError(reference, prediction);
   }
 
+  @Override
+  protected void printError(LanguageSample referenceSample, LanguageSample predictedSample) {
+    printStream.println(String.join("\t", referenceSample.getLanguage().getLang(),
+        predictedSample.getLanguage().getLang(),
+        referenceSample.getContext()));
+  }
 }

opennlp-tools/src/test/java/opennlp/tools/langdetect/LanguageDetectorEvaluatorTest.java

@@ -17,11 +17,15 @@
 
 package opennlp.tools.langdetect;
 
+import java.io.ByteArrayOutputStream;
+import java.nio.charset.StandardCharsets;
 import java.util.concurrent.atomic.AtomicInteger;
 
 import org.junit.Assert;
 import org.junit.Test;
 
+import opennlp.tools.cmdline.langdetect.LanguageDetectorEvaluationErrorListener;
+
 
 public class LanguageDetectorEvaluatorTest {
 
@@ -33,6 +37,8 @@ public void processSample() throws Exception {
     final AtomicInteger correctCount = new AtomicInteger();
     final AtomicInteger incorrectCount = new AtomicInteger();
 
+    ByteArrayOutputStream outputStream = new ByteArrayOutputStream();
+
     LanguageDetectorEvaluator evaluator = new LanguageDetectorEvaluator(langdetector,
         new LanguageDetectorEvaluationMonitor() {
           @Override
@@ -46,7 +52,7 @@ public void missclassified(LanguageSample reference,
                                      LanguageSample prediction) {
             incorrectCount.incrementAndGet();
           }
-        });
+        }, new LanguageDetectorEvaluationErrorListener(outputStream));
 
     evaluator.evaluateSample(new LanguageSample(new Language("pob"),
         "escreve e faz palestras pelo mundo inteiro sobre anjos"));
@@ -63,6 +69,12 @@ public void missclassified(LanguageSample reference,
 
     Assert.assertEquals(3, evaluator.getDocumentCount());
     Assert.assertEquals(0.33, evaluator.getAccuracy(), 0.01);
+
+    String report = outputStream.toString(StandardCharsets.UTF_8.name());
+
+    Assert.assertEquals("Expected\tPredicted\tContext\n" +
+        "fra\tpob\tescreve e faz palestras pelo mundo inteiro sobre anjos\n" +
+        "fra\tpob\tescreve e faz palestras pelo mundo inteiro sobre anjos\n", report);
   }
 
 }