Add metadataFieldsToFilter property for MongoDB store

Introduce a new property for the MongoDB vector store:
spring.ai.vectorstore.mongodb.metadata-fields-to-filter

This property accepts comma-separated values specifying which metadata
fields can be used for filtering when querying the vector store. It
ensures that metadata indexes are created if they don't already exist.

This addition enhances query performance and flexibility by allowing
users to define filterable fields in advance.

Co-authored-by: Eddú Meléndez <eddu.melendez@gmail.com>

spring-ai-docs/src/main/antora/modules/ROOT/pages/api/vectordbs/mongodb.adoc

@@ -120,6 +120,7 @@ spring.ai.vectorstore.mongodb.collection-name=vector_store
 spring.ai.vectorstore.mongodb.initialize-schema=true
 spring.ai.vectorstore.mongodb.path-name=embedding
 spring.ai.vectorstore.mongodb.indexName=vector_index
+spring.ai.vectorstore.mongodb.metadata-fields-to-filter=foo
 ----
 
 
@@ -131,6 +132,7 @@ spring.ai.vectorstore.mongodb.indexName=vector_index
 |`spring.ai.vectorstore.mongodb.initialize-schema`| whether to initialize the backend schema for you | `false`
 |`spring.ai.vectorstore.mongodb.path-name`| The name of the path to store the vectors. | `embedding`
 |`spring.ai.vectorstore.mongodb.indexName`| The name of the index to store the vectors. | `vector_index`
+|`spring.ai.vectorstore.mongodb.metadata-fields-to-filter` | comma separated values that specifies which metadata fields can be used for filtering when querying the vector store.  Needed so that metadata indexes are created if they already don't exist | empty list
 |===
 
 

spring-ai-spring-boot-autoconfigure/src/main/java/org/springframework/ai/autoconfigure/vectorstore/mongo/MongoDBAtlasVectorStoreAutoConfiguration.java

@@ -41,6 +41,7 @@
  * @author Eddú Meléndez
  * @author Christian Tzolov
  * @author Soby Chacko
+ * @author Ignacio López
  * @since 1.0.0
  */
 @AutoConfiguration
@@ -72,6 +73,9 @@ MongoDBAtlasVectorStore vectorStore(MongoTemplate mongoTemplate, EmbeddingModel
 		if (StringUtils.hasText(properties.getIndexName())) {
 			builder.withVectorIndexName(properties.getIndexName());
 		}
+		if (!properties.getMetadataFieldsToFilter().isEmpty()) {
+			builder.withMetadataFieldsToFilter(properties.getMetadataFieldsToFilter());
+		}
 		MongoDBAtlasVectorStore.MongoDBVectorStoreConfig config = builder.build();
 
 		return new MongoDBAtlasVectorStore(mongoTemplate, embeddingModel, config, properties.isInitializeSchema(),

spring-ai-spring-boot-autoconfigure/src/main/java/org/springframework/ai/autoconfigure/vectorstore/mongo/MongoDBAtlasVectorStoreProperties.java

@@ -18,9 +18,12 @@
 import org.springframework.ai.autoconfigure.vectorstore.CommonVectorStoreProperties;
 import org.springframework.boot.context.properties.ConfigurationProperties;
 
+import java.util.List;
+
 /**
  * @author Eddú Meléndez
  * @author Christian Tzolov
+ * @author Ignacio López
  * @since 1.0.0
  */
 @ConfigurationProperties(MongoDBAtlasVectorStoreProperties.CONFIG_PREFIX)
@@ -43,6 +46,11 @@ public class MongoDBAtlasVectorStoreProperties extends CommonVectorStoreProperti
 	 */
 	private String indexName;
 
+	/**
+	 * Name of the metadata fields to use as filters.
+	 */
+	private List<String> metadataFieldsToFilter = List.of();
+
 	public String getCollectionName() {
 		return this.collectionName;
 	}
@@ -67,4 +75,12 @@ public void setIndexName(String indexName) {
 		this.indexName = indexName;
 	}
 
+	public List<String> getMetadataFieldsToFilter() {
+		return this.metadataFieldsToFilter;
+	}
+
+	public void setMetadataFieldsToFilter(List<String> metadataFieldsToFilter) {
+		this.metadataFieldsToFilter = metadataFieldsToFilter;
+	}
+
 }

spring-ai-spring-boot-autoconfigure/src/test/java/org/springframework/ai/autoconfigure/vectorstore/mongo/MongoDBAtlasVectorStoreAutoConfigurationIT.java

@@ -21,6 +21,7 @@
 import java.time.Duration;
 import java.util.Collections;
 import java.util.List;
+import java.util.Map;
 import java.util.stream.Collectors;
 
 import org.junit.jupiter.api.Disabled;
@@ -32,6 +33,8 @@
 import org.springframework.ai.observation.conventions.VectorStoreProvider;
 import org.springframework.ai.vectorstore.SearchRequest;
 import org.springframework.ai.vectorstore.VectorStore;
+import org.springframework.ai.vectorstore.filter.Filter;
+import org.springframework.ai.vectorstore.filter.FilterExpressionBuilder;
 import org.springframework.ai.vectorstore.observation.VectorStoreObservationContext;
 import org.springframework.boot.autoconfigure.AutoConfigurations;
 import org.springframework.boot.autoconfigure.data.mongo.MongoDataAutoConfiguration;
@@ -40,6 +43,7 @@
 import org.springframework.boot.test.context.runner.ApplicationContextRunner;
 import org.springframework.context.annotation.Bean;
 import org.springframework.context.annotation.Configuration;
+import org.springframework.data.mongodb.core.MongoTemplate;
 import org.testcontainers.containers.GenericContainer;
 import org.testcontainers.containers.wait.strategy.Wait;
 import org.testcontainers.junit.jupiter.Container;
@@ -51,6 +55,7 @@
  * @author Eddú Meléndez
  * @author Christian Tzolov
  * @author Thomas Vitale
+ * @author Ignacio López
  */
 @Testcontainers
 @EnabledIfEnvironmentVariable(named = "OPENAI_API_KEY", matches = ".+")
@@ -70,7 +75,13 @@ class MongoDBAtlasVectorStoreAutoConfigurationIT {
 			new Document("Hello World Hello World Hello World Hello World Hello World Hello World Hello World"),
 			new Document(
 					"Great Depression Great Depression Great Depression Great Depression Great Depression Great Depression",
-					Collections.singletonMap("meta2", "meta2")));
+					Collections.singletonMap("meta2", "meta2")),
+			new Document(
+					"Testcontainers Testcontainers Testcontainers Testcontainers Testcontainers Testcontainers Testcontainers",
+					Collections.singletonMap("foo", "bar")),
+			new Document(
+					"Testcontainers Testcontainers Testcontainers Testcontainers Testcontainers Testcontainers Testcontainers",
+					Collections.singletonMap("foo", "baz")));
 
 	private final ApplicationContextRunner contextRunner = new ApplicationContextRunner()
 		.withUserConfiguration(Config.class)
@@ -123,6 +134,35 @@ public void addAndSearch() {
 
 			List<Document> results2 = vectorStore.similaritySearch(SearchRequest.query("Great").withTopK(1));
 			assertThat(results2).isEmpty();
+
+			context.getBean(MongoTemplate.class).dropCollection("test_collection");
+		});
+	}
+
+	@Test
+	public void addAndSearchWithFilters() {
+		contextRunner.withPropertyValues("spring.ai.vectorstore.mongodb.metadata-fields-to-filter=foo").run(context -> {
+
+			VectorStore vectorStore = context.getBean(VectorStore.class);
+			vectorStore.add(documents);
+
+			Thread.sleep(5000); // Await a second for the document to be indexed
+
+			List<Document> results = vectorStore.similaritySearch(SearchRequest.query("Testcontainers").withTopK(2));
+			assertThat(results).hasSize(2);
+			results.forEach(doc -> assertThat(doc.getContent().contains("Testcontainers")).isTrue());
+
+			FilterExpressionBuilder b = new FilterExpressionBuilder();
+			results = vectorStore.similaritySearch(
+					SearchRequest.query("Testcontainers").withTopK(2).withFilterExpression(b.eq("foo", "bar").build()));
+
+			assertThat(results).hasSize(1);
+			Document resultDoc = results.get(0);
+			assertThat(resultDoc.getId()).isEqualTo(documents.get(3).getId());
+			assertThat(resultDoc.getContent().contains("Testcontainers")).isTrue();
+			assertThat(resultDoc.getMetadata()).containsEntry("foo", "bar");
+
+			context.getBean(MongoTemplate.class).dropCollection("test_collection");
 		});
 	}