CNDB-13192: Bump the default cassandra.sai.latest.version from DC to EC so we support BM25 by default. (#1737)

We do not need to update JVECTOR_VERSION post CNDB-14301

Author: ekaterinadimitrova2

src/java/org/apache/cassandra/config/CassandraRelevantProperties.java

@@ -358,7 +358,7 @@ public enum CassandraRelevantProperties
     CUSTOM_SSTABLE_WATCHER("cassandra.custom_sstable_watcher"),
 
     /** The current version of the SAI on-disk index format. */
-    SAI_CURRENT_VERSION("cassandra.sai.latest.version", "dc"),
+    SAI_CURRENT_VERSION("cassandra.sai.latest.version", "ec"),
 
     /** Controls the maximum top-k limit for vector search */
     SAI_VECTOR_SEARCH_MAX_TOP_K("cassandra.sai.vector_search.max_top_k", "1000"),

test/unit/org/apache/cassandra/index/sai/disk/v1/InvertedIndexSearcherTest.java

@@ -24,6 +24,8 @@
 import java.util.stream.Collectors;
 
 import org.agrona.collections.Int2IntHashMap;
+import org.apache.cassandra.index.sai.*;
+import org.junit.Before;
 import org.junit.BeforeClass;
 import org.junit.Test;
 
@@ -32,10 +34,6 @@
 import org.apache.cassandra.cql3.Operator;
 import org.apache.cassandra.db.marshal.UTF8Type;
 import org.apache.cassandra.dht.Murmur3Partitioner;
-import org.apache.cassandra.index.sai.IndexContext;
-import org.apache.cassandra.index.sai.QueryContext;
-import org.apache.cassandra.index.sai.SAITester;
-import org.apache.cassandra.index.sai.SSTableContext;
 import org.apache.cassandra.index.sai.disk.MemtableTermsIterator;
 import org.apache.cassandra.index.sai.disk.format.IndexComponents;
 import org.apache.cassandra.index.sai.disk.format.IndexDescriptor;
@@ -61,6 +59,8 @@ public class InvertedIndexSearcherTest extends SaiRandomizedTest
     // Use a shared index context to prevent creating too many metrics unnecessarily
     private final IndexContext indexContext = SAITester.createIndexContext("meh", UTF8Type.instance);
 
+    private final Version version;
+
     @ParametersFactory()
     public static Collection<Object[]> data()
     {
@@ -70,7 +70,11 @@ public static Collection<Object[]> data()
         return Version.ALL.stream().map(v -> new Object[]{v}).collect(Collectors.toList());
     }
 
-    private final Version version;
+    @Before
+    public void setCurrentSAIVersion()
+    {
+        SAIUtil.setCurrentVersion(version);
+    }
 
     public InvertedIndexSearcherTest(Version version)
     {
@@ -214,7 +218,7 @@ private IndexSearcher buildIndexAndOpenSearcher(int terms, List<InvertedIndexBui
 
     private List<InvertedIndexBuilder.TermsEnum> buildTermsEnum(Version version, int terms, int postings)
     {
-        return InvertedIndexBuilder.buildStringTermsEnum(version, terms, postings, () -> randomSimpleString(3, 5), () -> nextInt(0, Integer.MAX_VALUE));
+        return InvertedIndexBuilder.buildStringTermsEnum(version, terms, postings, () -> randomSimpleString(3, 5), () -> nextInt(0, terms * postings * 2));
     }
 
     private Int2IntHashMap createMockDocLengths(List<InvertedIndexBuilder.TermsEnum> termsEnum)

test/unit/org/apache/cassandra/index/sai/disk/v1/TermsReaderTest.java

@@ -18,13 +18,11 @@
 package org.apache.cassandra.index.sai.disk.v1;
 
 import java.io.IOException;
-import java.util.ArrayList;
 import java.util.Collection;
-import java.util.HashMap;
 import java.util.List;
-import java.util.Map;
 import java.util.stream.Collectors;
 
+import org.junit.Before;
 import org.junit.Test;
 
 import com.carrotsearch.randomizedtesting.annotations.ParametersFactory;
@@ -37,7 +35,6 @@
 import org.apache.cassandra.index.sai.SAIUtil;
 import org.apache.cassandra.index.sai.disk.MemtableTermsIterator;
 import org.apache.cassandra.index.sai.disk.PostingList;
-import org.apache.cassandra.index.sai.disk.RAMStringIndexer;
 import org.apache.cassandra.index.sai.disk.TermsIterator;
 import org.apache.cassandra.index.sai.disk.format.IndexComponentType;
 import org.apache.cassandra.index.sai.disk.format.IndexComponents;
@@ -50,11 +47,9 @@
 import org.apache.cassandra.index.sai.utils.SaiRandomizedTest;
 import org.apache.cassandra.index.sai.utils.TypeUtil;
 import org.apache.cassandra.io.util.FileHandle;
-import org.apache.cassandra.utils.ByteBufferUtil;
 import org.apache.cassandra.utils.Pair;
 import org.apache.cassandra.utils.bytecomparable.ByteComparable;
 import org.apache.cassandra.utils.bytecomparable.ByteSourceInverse;
-import org.apache.lucene.util.BytesRef;
 
 import static org.apache.cassandra.index.sai.disk.v1.InvertedIndexBuilder.buildStringTermsEnum;
 import static org.apache.cassandra.index.sai.metrics.QueryEventListeners.NO_OP_TRIE_LISTENER;
@@ -63,6 +58,8 @@ public class TermsReaderTest extends SaiRandomizedTest
 {
     public static final ByteComparable.Version VERSION = TypeUtil.BYTE_COMPARABLE_VERSION;
 
+    private final Version version;
+
     @ParametersFactory()
     public static Collection<Object[]> data()
     {
@@ -72,7 +69,11 @@ public static Collection<Object[]> data()
         return Version.ALL.stream().map(v -> new Object[]{v}).collect(Collectors.toList());
     }
 
-    private final Version version;
+    @Before
+    public void setCurrentSAIVersion()
+    {
+        SAIUtil.setCurrentVersion(version);
+    }
 
     public TermsReaderTest(Version version)
     {
@@ -239,6 +240,8 @@ private Int2IntHashMap createMockDocLengths(List<InvertedIndexBuilder.TermsEnum>
 
     private List<InvertedIndexBuilder.TermsEnum> buildTermsEnum(Version version, int terms, int postings)
     {
-        return buildStringTermsEnum(version, terms, postings, () -> randomSimpleString(4, 10), () -> nextInt(0, Integer.MAX_VALUE));
+        // We use terms * postings * 2 as the upper bound on row ids used in the postings list to allow for a somewhat
+        // sparse mapping but not one that is too sparse.
+        return buildStringTermsEnum(version, terms, postings, () -> randomSimpleString(4, 10), () -> nextInt(0, terms * postings * 2));
     }
 }