RAG the Easy Way with Oracle AI Vector Search, LangChain4J, OracleEmbeddingStore, and Oracle Database 23ai

A minimal Java example demonstrating how to build an embedding store using LangChain4J’s Oracle extension, store embeddings in Oracle Database 23ai, and perform vector similarity retrieval for a basic RAG workflow.

RAG the Easy Way with Oracle AI Vector Search, LangChain4J, OracleEmbeddingStore, and Oracle Database 23ai

---

Prerequisites

• Java Development Kit (JDK) 17+
• Oracle Database Free Release 23ai (via container image)
• Oracle JDBC Driver 23ai (version 23.9.0.25.07)
• Preferred Java IDE (Eclipse, IntelliJ IDEA, or VS Code)
• Build tool: Maven or Gradle

---

Step 1: Configure Your Java Project

Edit your pom.xml to include:

<properties>
  <java.version>17</java.version>
  <maven.compiler.source>17</maven.compiler.source>
  <maven.compiler.target>17</maven.compiler.target>
  <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
  <jdbc.version>23.9.0.25.07</jdbc.version>
  <langchain4j.version>1.3.0‑beta9</langchain4j.version>
</properties>

<dependencies>
  <!-- Oracle Database 23ai JDBC / UCP -->
  <dependency>
    <groupId>com.oracle.database.jdbc</groupId>
    <artifactId>ojdbc17</artifactId>
    <version>${jdbc.version}</version>
  </dependency>
  <dependency>
    <groupId>com.oracle.database.jdbc</groupId>
    <artifactId>ucp17</artifactId>
    <version>${jdbc.version}</version>
  </dependency>

  <!-- LangChain4J Oracle Extension -->
  <dependency>
    <groupId>dev.langchain4j</groupId>
    <artifactId>langchain4j‑oracle</artifactId>
    <version>${langchain4j.version}</version>
  </dependency>
  <dependency>
    <groupId>dev.langchain4j</groupId>
    <artifactId>langchain4j‑embeddings‑all‑minilm‑l6‑v2‑q</artifactId>
    <version>${langchain4j.version}</version>
  </dependency>
</dependencies>

---

Step 2: Java Implementation with OracleEmbeddingStore

Create a Java class (e.g., OracleEmbeddingStoreExample.java) to demonstrate embedding storage and retrieval:

package com.oracle.dev.jdbc.langchain4j;

import java.sql.SQLException;
import javax.sql.DataSource;
import dev.langchain4j.data.embedding.Embedding;
import dev.langchain4j.data.segment.TextSegment;
import dev.langchain4j.model.embedding.EmbeddingModel;
import dev.langchain4j.model.embedding.onnx.allminilml6v2q.AllMiniLmL6V2QuantizedEmbeddingModel;
import dev.langchain4j.rag.content.Content;
import dev.langchain4j.rag.content.retriever.ContentRetriever;
import dev.langchain4j.rag.content.retriever.EmbeddingStoreContentRetriever;
import dev.langchain4j.rag.query.Query;
import dev.langchain4j.store.embedding.EmbeddingStore;
import dev.langchain4j.store.embedding.oracle.CreateOption;
import dev.langchain4j.store.embedding.oracle.OracleEmbeddingStore;

public class OracleEmbeddingStoreExample {

  public static void main(String[] args) throws SQLException {

    DataSource dataSource = OracleDBUtils.getPooledDataSource();

    EmbeddingStore<TextSegment> embeddingStore = OracleEmbeddingStore.builder()
      .dataSource(dataSource)
      .embeddingTable("test_content_retriever", CreateOption.CREATE_OR_REPLACE)
      .build();

    EmbeddingModel embeddingModel = new AllMiniLmL6V2QuantizedEmbeddingModel();

    ContentRetriever retriever = EmbeddingStoreContentRetriever.builder()
      .embeddingStore(embeddingStore)
      .embeddingModel(embeddingModel)
      .maxResults(2)
      .minScore(0.5)
      .build();

    TextSegment segment1 = TextSegment.from("I like soccer.");
    Embedding embedding1 = embeddingModel.embed(segment1).content();
    embeddingStore.add(embedding1, segment1);

    TextSegment segment2 = TextSegment.from("I love Stephen King.");
    Embedding embedding2 = embeddingModel.embed(segment2).content();
    embeddingStore.add(embedding2, segment2);

    Content match = retriever.retrieve(Query.from("What is your favourite writer?")).get(0);

    System.out.println(match.textSegment());
  }
}

---

Step 3: DataSource Utility — OracleDBUtils.java

Provide a helper class to manage the Oracle JDBC connection:

package com.oracle.dev.jdbc.langchain4j;

import java.sql.Connection;
import java.sql.SQLException;
import javax.sql.DataSource;
import java.util.Properties;
import oracle.ucp.jdbc.PoolDataSource;
import oracle.ucp.jdbc.PoolDataSourceFactory;

public class OracleDBUtils {

  private final static String URL = "jdbc:oracle:thin:@localhost:1521/FREEPDB1";
  private final static String USERNAME = System.getenv("DB_23AI_USERNAME");
  private final static String PASSWORD = System.getenv("DB_23AI_PASSWORD");

  public static DataSource getPooledDataSource() throws SQLException {
    PoolDataSource pds = PoolDataSourceFactory.getPoolDataSource();
    pds.setConnectionFactoryClassName("oracle.jdbc.pool.OracleDataSource");
    pds.setURL(URL);
    pds.setUser(USERNAME);
    pds.setPassword(PASSWORD);

    Properties prop = new Properties();
    prop.setProperty("oracle.jdbc.vectorDefaultGetObjectType", "String");
    pds.setConnectionProperties(prop);

    pds.setInitialPoolSize(10);
    return pds;
  }
}

---

Step 4: Running the Sample

1. Compile and run the OracleEmbeddingStoreExample in your IDE of choice.
2. Expect output corresponding to whichever segment best matches the query "What is your favourite writer?" — in this case, likely "I love Stephen King.".
3. Try experimenting with different queries (e.g., "What is your favourite sport?") to see semantic retrieval in action.

---

Summary

You’ve successfully built a minimal Java RAG workflow leveraging:

• Oracle Database 23ai as a vector store,
• LangChain4J Oracle extension for seamless integration,
• Lightweight ONNX-based embedding model,
• Simple JDBC setup and retrieval logic.

With this foundation, you're well-positioned to expand into more complex RAG scenarios, integrate other embedding models, or connect with LLMs for generative applications.