Fix column creation, when adding additional normal and embedding colums), and make index name unique (for when there are multiple vector indexes in the same keyspace)

And change stream to for-loop when converting List<Double> to Float[] for performance

Author: michaelsembwever

spring-ai-spring-boot-autoconfigure/src/main/java/org/springframework/ai/autoconfigure/vectorstore/cassandra/CassandraVectorStoreProperties.java

@@ -33,7 +33,7 @@ public class CassandraVectorStoreProperties {
 
 	private String table = CassandraVectorStoreConfig.DEFAULT_TABLE_NAME;
 
-	private String indexName = CassandraVectorStoreConfig.DEFAULT_INDEX_NAME;
+	private String indexName = null;
 
 	private String contentColumnName = CassandraVectorStoreConfig.DEFAULT_CONTENT_COLUMN_NAME;
 

spring-ai-spring-boot-autoconfigure/src/test/java/org/springframework/ai/autoconfigure/vectorstore/cassandra/CassandraVectorStorePropertiesTests.java

@@ -34,7 +34,7 @@ void defaultValues() {
 		assertThat(props.getTable()).isEqualTo(CassandraVectorStoreConfig.DEFAULT_TABLE_NAME);
 		assertThat(props.getContentColumnName()).isEqualTo(CassandraVectorStoreConfig.DEFAULT_CONTENT_COLUMN_NAME);
 		assertThat(props.getEmbeddingColumnName()).isEqualTo(CassandraVectorStoreConfig.DEFAULT_EMBEDDING_COLUMN_NAME);
-		assertThat(props.getIndexName()).isEqualTo(CassandraVectorStoreConfig.DEFAULT_INDEX_NAME);
+		assertThat(props.getIndexName()).isNull();
 		assertThat(props.getDisallowSchemaCreation()).isFalse();
 		assertThat(props.getFixedThreadPoolExecutorSize())
 			.isEqualTo(CassandraVectorStoreConfig.DEFAULT_ADD_CONCURRENCY);

vector-stores/spring-ai-cassandra/src/main/java/org/springframework/ai/vectorstore/CassandraVectorStore.java

@@ -206,7 +206,7 @@ public Optional<Boolean> delete(List<String> idList) {
 	@Override
 	public List<Document> similaritySearch(SearchRequest request) {
 		Preconditions.checkArgument(request.getTopK() <= 1000);
-		var embedding = this.embeddingClient.embed(request.getQuery()).stream().map(Double::floatValue).toList();
+		var embedding = toFloatArray(this.embeddingClient.embed(request.getQuery()));
 		CqlVector<Float> cqlVector = CqlVector.newInstance(embedding);
 
 		String whereClause = "";
@@ -350,4 +350,13 @@ private String getDocumentId(Row row) {
 		return this.conf.primaryKeyTranslator.apply(primaryKeyValues);
 	}
 
+	private static Float[] toFloatArray(List<Double> embeddingDouble) {
+		Float[] embeddingFloat = new Float[embeddingDouble.size()];
+		int i = 0;
+		for (Double d : embeddingDouble) {
+			embeddingFloat[i++] = d.floatValue();
+		}
+		return embeddingFloat;
+	}
+
 }

vector-stores/spring-ai-cassandra/src/main/java/org/springframework/ai/vectorstore/CassandraVectorStoreConfig.java

@@ -44,9 +44,12 @@
 import com.datastax.oss.driver.api.querybuilder.schema.CreateTableStart;
 import com.datastax.oss.driver.shaded.guava.common.annotations.VisibleForTesting;
 import com.datastax.oss.driver.shaded.guava.common.base.Preconditions;
+
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
+import org.springframework.lang.Nullable;
+
 /**
  * Configuration for the Cassandra vector store.
  *
@@ -71,7 +74,7 @@ public final class CassandraVectorStoreConfig implements AutoCloseable {
 
 	public static final String DEFAULT_ID_NAME = "id";
 
-	public static final String DEFAULT_INDEX_NAME = "embedding_index";
+	public static final String DEFAULT_INDEX_SUFFIX = "idx";
 
 	public static final String DEFAULT_CONTENT_COLUMN_NAME = "content";
 
@@ -186,7 +189,7 @@ public static class Builder {
 
 		private List<SchemaColumn> clusteringKeys = List.of();
 
-		private String indexName = DEFAULT_INDEX_NAME;
+		private String indexName = null;
 
 		private String contentColumnName = DEFAULT_CONTENT_COLUMN_NAME;
 
@@ -257,6 +260,8 @@ public Builder withClusteringKeys(List<SchemaColumn> clusteringKeys) {
 			return this;
 		}
 
+		/** defaults (if null) to '<table_name>_<embedding_column_name>_idx' **/
+		@Nullable
 		public Builder withIndexName(String name) {
 			this.indexName = name;
 			return this;
@@ -324,6 +329,9 @@ public Builder withPrimaryKeyTranslator(PrimaryKeyTranslator primaryKeyTranslato
 		}
 
 		public CassandraVectorStoreConfig build() {
+			if (null == this.indexName) {
+				this.indexName = String.format("%s_%s_%s", this.table, this.embeddingColumnName, DEFAULT_INDEX_SUFFIX);
+			}
 			for (SchemaColumn metadata : this.metadataColumns) {
 
 				Preconditions.checkArgument(
@@ -530,7 +538,7 @@ private void ensureTableColumnsExist(int vectorDimension) {
 				// special case for embedding column, bc JAVA-3118, as above
 				StringBuilder alterTableStmt = new StringBuilder(((BuildableQuery) alterTable).asCql());
 				if (newColumns.isEmpty() && !addContent) {
-					alterTableStmt.append(" ADD ");
+					alterTableStmt.append(" ADD (");
 				}
 				else {
 					alterTableStmt.setLength(alterTableStmt.length() - 1);
@@ -539,7 +547,7 @@ private void ensureTableColumnsExist(int vectorDimension) {
 				alterTableStmt.append(this.schema.embedding)
 					.append(" vector<float,")
 					.append(vectorDimension)
-					.append(">");
+					.append(">)");
 
 				logger.debug("Executing {}", alterTableStmt.toString());
 				this.session.execute(alterTableStmt.toString());

vector-stores/spring-ai-cassandra/src/test/java/org/springframework/ai/vectorstore/CassandraRichSchemaVectorStoreIT.java

@@ -134,7 +134,8 @@ void ensureSchemaNoCreation() {
 	@Test
 	void ensureSchemaPartialCreation() {
 		this.contextRunner.run(context -> {
-			for (int i = 0; i < 4; ++i) {
+			int PARTIAL_FILES = 5;
+			for (int i = 0; i < PARTIAL_FILES; ++i) {
 				executeCqlFile(context, format("test_wiki_partial_%d_schema.cql", i));
 				var wrapper = createStore(context, List.of(), false, false);
 				try {
@@ -148,6 +149,10 @@ void ensureSchemaPartialCreation() {
 					wrapper.store().close();
 				}
 			}
+			// make sure there's not more files to test
+			Assertions.assertThrows(IOException.class, () -> {
+				executeCqlFile(context, format("test_wiki_partial_%d_schema.cql", PARTIAL_FILES));
+			});
 		});
 	}
 

vector-stores/spring-ai-cassandra/src/test/resources/test_wiki_partial_4_schema.cql

@@ -0,0 +1,10 @@
+CREATE KEYSPACE IF NOT EXISTS test_wikidata WITH replication = {'class': 'SimpleStrategy', 'replication_factor': 1};
+
+CREATE TABLE IF NOT EXISTS test_wikidata.articles (
+    wiki text,
+    language text,
+    title text,
+    chunk_no int,
+    messages text,
+    PRIMARY KEY ((wiki, language, title), chunk_no)
+);