Fix a flaky unit test:testMultiFieldsKnnIndex, which was failing due …

…to inconsistent merge behaviors (opensearch-project#1924)

Signed-off-by: Navneet Verma <navneev@amazon.com>

Quantization Framework Implementation with 1bit, 2bit and 4bit Binary Quantizer

CHANGELOG.md

@@ -21,6 +21,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 ### Infrastructure
 ### Documentation
 ### Maintenance
+* Fix a flaky unit test:testMultiFieldsKnnIndex, which was failing due to inconsistent merge behaviors [#1924](https://github.com/opensearch-project/k-NN/pull/1924)
 ### Refactoring
 * Introduce KNNVectorValues interface to iterate on different types of Vector values during indexing and search [#1897](https://github.com/opensearch-project/k-NN/pull/1897)
 * Clean up parsing for query [#1824](https://github.com/opensearch-project/k-NN/pull/1824)

src/main/java/org/opensearch/knn/quantization/sampler/SamplingFactory.java

@@ -9,7 +9,15 @@
  * SamplingFactory is a factory class for creating instances of Sampler.
  * It uses the factory design pattern to encapsulate the creation logic for different types of samplers.
  */
-public class SamplingFactory {
+public final class SamplingFactory {
+
+    /**
+     * Private constructor to prevent instantiation of this  class.
+     * The class is not meant to be instantiated, as it provides static methods only.
+     */
+    private SamplingFactory() {
+
+    }
 
     /**
      * SamplerType is an enumeration of the different types of samplers that can be created by the factory.

src/main/java/org/opensearch/knn/quantization/util/QuantizationStateSerializer.java

@@ -8,7 +8,12 @@
 import org.opensearch.knn.quantization.models.quantizationParams.QuantizationParams;
 import org.opensearch.knn.quantization.models.quantizationState.QuantizationState;
 
-import java.io.*;
+import java.io.ByteArrayOutputStream;
+import java.io.ObjectOutputStream;
+import java.io.Serializable;
+import java.io.IOException;
+import java.io.ByteArrayInputStream;
+import java.io.ObjectInputStream;
 
 /**
  * QuantizationStateSerializer is a utility class that provides methods for serializing and deserializing

src/test/java/org/opensearch/knn/index/codec/KNNCodecTestCase.java

@@ -137,28 +137,27 @@ public void testMultiFieldsKnnIndex(Codec codec) throws Exception {
         Document doc = new Document();
         doc.add(vectorField);
         writer.addDocument(doc);
-        writer.close();
+        // ensuring the refresh happens, to create the segment and hnsw file
+        writer.flush();
 
         /**
          * Add doc with field "my_vector"
          */
-        IndexWriterConfig iwc1 = newIndexWriterConfig();
-        iwc1.setMergeScheduler(new SerialMergeScheduler());
-        iwc1.setCodec(ACTUAL_CODEC);
-        writer = new RandomIndexWriter(random(), dir, iwc1);
         float[] array1 = { 6.0f, 14.0f };
         VectorField vectorField1 = new VectorField("my_vector", array1, sampleFieldType);
         Document doc1 = new Document();
         doc1.add(vectorField1);
         writer.addDocument(doc1);
+        // ensuring the refresh happens, to create the segment and hnsw file
+        writer.flush();
         IndexReader reader = writer.getReader();
         writer.close();
         ResourceWatcherService resourceWatcherService = createDisabledResourceWatcherService();
         NativeMemoryLoadStrategy.IndexLoadStrategy.initialize(resourceWatcherService);
         List<String> hnswfiles = Arrays.stream(dir.listAll()).filter(x -> x.contains("hnsw")).collect(Collectors.toList());
 
         // there should be 2 hnsw index files created. one for test_vector and one for my_vector
-        assertEquals(hnswfiles.size(), 2);
+        assertEquals(2, hnswfiles.size());
         assertEquals(hnswfiles.stream().filter(x -> x.contains("test_vector")).collect(Collectors.toList()).size(), 1);
         assertEquals(hnswfiles.stream().filter(x -> x.contains("my_vector")).collect(Collectors.toList()).size(), 1);
 

src/test/java/org/opensearch/knn/quantization/factory/QuantizerFactoryTests.java

@@ -5,20 +5,14 @@
 
 package org.opensearch.knn.quantization.factory;
 
-<<<<<<< HEAD
 import org.junit.Before;
-=======
-import org.junit.BeforeClass;
-import org.junit.Test;
->>>>>>> b93c2a3 (Quantization Framework Implementation with 1bit, 2bit and 4bit Binary Quantizer)
 import org.opensearch.knn.KNNTestCase;
 import org.opensearch.knn.quantization.enums.SQTypes;
 import org.opensearch.knn.quantization.models.quantizationParams.SQParams;
 import org.opensearch.knn.quantization.quantizer.MultiBitScalarQuantizer;
 import org.opensearch.knn.quantization.quantizer.OneBitScalarQuantizer;
 import org.opensearch.knn.quantization.quantizer.Quantizer;
 
-<<<<<<< HEAD
 import java.lang.reflect.Field;
 
 public class QuantizerFactoryTests extends KNNTestCase {
@@ -36,35 +30,27 @@ public void test_Lazy_Registration() {
         Quantizer<?, ?> quantizer = QuantizerFactory.getQuantizer(params);
         assertTrue(quantizer instanceof OneBitScalarQuantizer);
         assertTrue(isRegisteredFieldAccessible());
-=======
-public class QuantizerFactoryTests extends KNNTestCase {
-
-    public void testLazyRegistration() {
-        SQParams params = new SQParams(SQTypes.ONE_BIT);
-        Quantizer<?, ?> quantizer = QuantizerFactory.getQuantizer(params);
-        assertTrue(quantizer instanceof OneBitScalarQuantizer);
->>>>>>> b93c2a3 (Quantization Framework Implementation with 1bit, 2bit and 4bit Binary Quantizer)
     }
 
-    public void testGetQuantizer_withOneBitSQParams() {
+    public void testGetQuantizer_withOneBitSQParams () {
         SQParams params = new SQParams(SQTypes.ONE_BIT);
         Quantizer<?, ?> quantizer = QuantizerFactory.getQuantizer(params);
         assertTrue(quantizer instanceof OneBitScalarQuantizer);
     }
 
-    public void testGetQuantizer_withTwoBitSQParams() {
+    public void testGetQuantizer_withTwoBitSQParams () {
         SQParams params = new SQParams(SQTypes.TWO_BIT);
         Quantizer<?, ?> quantizer = QuantizerFactory.getQuantizer(params);
         assertTrue(quantizer instanceof MultiBitScalarQuantizer);
     }
 
-    public void testGetQuantizer_withFourBitSQParams() {
+    public void testGetQuantizer_withFourBitSQParams () {
         SQParams params = new SQParams(SQTypes.FOUR_BIT);
         Quantizer<?, ?> quantizer = QuantizerFactory.getQuantizer(params);
         assertTrue(quantizer instanceof MultiBitScalarQuantizer);
     }
 
-    public void testGetQuantizer_withUnsupportedType() {
+    public void testGetQuantizer_withUnsupportedType () {
         SQParams params = new SQParams(SQTypes.UNSUPPORTED_TYPE);
         try {
             QuantizerFactory.getQuantizer(params);
@@ -73,11 +59,7 @@ public void testGetQuantizer_withUnsupportedType() {
             assertTrue(e.getMessage().contains("No quantizer registered for type identifier"));
         }
     }
-<<<<<<< HEAD
-
-=======
->>>>>>> b93c2a3 (Quantization Framework Implementation with 1bit, 2bit and 4bit Binary Quantizer)
-    public void testGetQuantizer_withNullParams() {
+    public void testGetQuantizer_withNullParams () {
         try {
             QuantizerFactory.getQuantizer(null);
             fail("Expected IllegalArgumentException");
@@ -86,12 +68,7 @@ public void testGetQuantizer_withNullParams() {
         }
     }
 
-<<<<<<< HEAD
-
-    public void test_Concurrent_Registration() throws InterruptedException {
-=======
-    public void testConcurrentRegistration() throws InterruptedException {
->>>>>>> b93c2a3 (Quantization Framework Implementation with 1bit, 2bit and 4bit Binary Quantizer)
+    public void testConcurrentRegistration () throws InterruptedException {
         Runnable task = () -> {
             SQParams params = new SQParams(SQTypes.ONE_BIT);
             QuantizerFactory.getQuantizer(params);
@@ -103,12 +80,10 @@ >>>>>>> b93c2a3 (Quantization Framework Implementation with 1bit, 2bit and 4bit
         thread2.start();
         thread1.join();
         thread2.join();
-
-<<<<<<< HEAD
         assertTrue(isRegisteredFieldAccessible());
     }
 
-    private boolean isRegisteredFieldAccessible() {
+    private boolean isRegisteredFieldAccessible () {
         try {
             Field field = QuantizerFactory.class.getDeclaredField("isRegistered");
             field.setAccessible(true);
@@ -117,8 +92,5 @@ private boolean isRegisteredFieldAccessible() {
             fail("Failed to access isRegistered field.");
             return false;
         }
-=======
-        // No need to check isRegistered flag, as it doesn't exist anymore
->>>>>>> b93c2a3 (Quantization Framework Implementation with 1bit, 2bit and 4bit Binary Quantizer)
     }
 }