Moved normalization, more tests.

[DictionaryPC.git] / src / com / hughes / android / dictionary / engine / DictionaryTest.java

diff --git a/src/com/hughes/android/dictionary/engine/DictionaryTest.java b/src/com/hughes/android/dictionary/engine/DictionaryTest.java
index f766e0ab50ace5d74ece89b47618f6dbb2fbaa72..52c3161b206e2712bc7281775dabc470d30f0170 100644 (file)
--- a/src/com/hughes/android/dictionary/engine/DictionaryTest.java
+++ b/src/com/hughes/android/dictionary/engine/DictionaryTest.java
@@ -18,6 +18,7 @@ import java.io.IOException;
 import java.io.RandomAccessFile;
 import java.util.Arrays;
 import java.util.List;
+import java.util.Random;
 import java.util.concurrent.atomic.AtomicBoolean;
 
 import junit.framework.TestCase;
@@ -178,17 +179,122 @@ public class DictionaryTest extends TestCase {
   }
 
   public void testMultiSearchBig() throws IOException {
-    final RandomAccessFile raf = new RandomAccessFile(OUTPUTS + "EN-IT_enwiktionary.quickdic", "r");
+    final RandomAccessFile raf = new RandomAccessFile(OUTPUTS + "DE-EN_chemnitz_enwiktionary.quickdic", "r");
+    final Dictionary dict = new Dictionary(raf);
+    final Index enIndex = dict.indices.get(1);
+
+    {
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("train", "station"), new AtomicBoolean(false));
+    System.out.println(CollectionUtil.join(rows, "\n  "));
+    assertTrue(rows.toString(), rows.size() > 0);
+    assertEquals("Bahnhof {{de-noun|g=m|genitive=Bahnhofs|genitive2=Bahnhofes|plural=Bahnhöfe}}\ttrain station", rows.get(0).toString());
+    }
+
+    {
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("a", "train", "station"), new AtomicBoolean(false));
+    System.out.println(CollectionUtil.join(rows, "\n  "));
+    assertTrue(rows.toString(), rows.size() > 0);
+    assertEquals("Bahnhofsuhr {{de-noun|g=f|plural=Bahnhofsuhren}}\tstation clock (at a train station)", rows.get(0).toString());
+    }
+
+    {
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("a", "station"), new AtomicBoolean(false));
+    System.out.println(CollectionUtil.join(rows, "\n  "));
+    assertTrue(rows.toString(), rows.size() > 0);
+    assertEquals("Kraftwerk {n}\tpower plant (a station built for the production of electric power) (noun)", rows.get(0).toString());
+    }
+
+    {
+    // Should print: Giving up, too many words with prefix: p
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("p", "eat"), new AtomicBoolean(false));
+    System.out.println(CollectionUtil.join(rows, "\n  "));
+    assertTrue(rows.toString(), rows.size() > 0);
+    assertTrue(rows.toString().contains("verschlingen; verputzen\tto dispatch (eat)"));
+    }
+
+    {
+    // Should print: Giving up, too many words with prefix: p
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("p", "p"), new AtomicBoolean(false));
+    assertTrue(rows.size() >= 1000);
+    }
+
+    {
+    // Should print: Giving up, too many words with prefix: a
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("a", "a"), new AtomicBoolean(false));
+    assertTrue(rows.size() >= 1000);
+    }
+
+    {
+    // Should print: Giving up, too many words with prefix: a
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("b", "ba"), new AtomicBoolean(false));
+    assertTrue(rows.size() >= 1000);
+    }
+
+    {
+    // Should print: Giving up, too many words with prefix: a
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("b", "ba"), new AtomicBoolean(false));
+    assertTrue(rows.size() >= 1000);
+    }
+
+    raf.close();
+  }
+
+  public void testMultiSearchBigAF() throws IOException {
+    final RandomAccessFile raf = new RandomAccessFile(OUTPUTS + "EN-AF_enwiktionary.quickdic", "r");
     final Dictionary dict = new Dictionary(raf);
     final Index enIndex = dict.indices.get(0);
 
     {
-    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("space", "station"), new AtomicBoolean(false));
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("pig", "eats"), new AtomicBoolean(false));
+    System.out.println(CollectionUtil.join(rows, "\n  "));
+    assertTrue(rows.toString(), rows.size() > 0);
+    assertEquals("pig (someone who overeats or eats rapidly) (noun)\tvark", rows.get(0).toString());
+    }
+
+    {
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("pig", "eat"), new AtomicBoolean(false));
+    System.out.println(CollectionUtil.join(rows, "\n  "));
+    assertTrue(rows.toString(), rows.size() > 0);
+    assertEquals("pig (someone who overeats or eats rapidly) (noun)\tvark", rows.get(0).toString());
+    }
+
+    {
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("pi", "ea"), new AtomicBoolean(false));
     System.out.println(CollectionUtil.join(rows, "\n  "));
     assertTrue(rows.toString(), rows.size() > 0);
+    assertTrue(rows.toString().contains("pig (someone who overeats or eats rapidly) (noun)\tvark"));
+    }
+
+    {
+    final List<RowBase> rows = enIndex.multiWordSearch(Arrays.asList("p", "eat"), new AtomicBoolean(false));
+    System.out.println(CollectionUtil.join(rows, "\n  "));
+    assertTrue(rows.toString(), rows.size() > 0);
+    assertTrue(rows.toString().contains("pig (someone who overeats or eats rapidly) (noun)\tvark"));
+    }
+
+    
+    raf.close();
+  }
+
+
+  public void testExactSearch() throws IOException {
+    final RandomAccessFile raf = new RandomAccessFile(OUTPUTS + "EN-ZH_enwiktionary.quickdic", "r");
+    final Dictionary dict = new Dictionary(raf);
+    final Index zhIndex = dict.indices.get(1);
+
+    final Random random = new Random(10);
+    
+    for (int i = 0; i < 1000; ++i) {
+      final int ii = random.nextInt(zhIndex.sortedIndexEntries.size());
+      final IndexEntry indexEntry = zhIndex.sortedIndexEntries.get(ii);
+      final IndexEntry found = zhIndex.findExact(indexEntry.token);
+      assertNotNull(found);
+      assertEquals(indexEntry.token, found.token);
+      assertEquals(indexEntry, found);  // Test of caching....
     }
     
     raf.close();
   }
 
+
 }