refactor: upgrade to SpringAI 1.0.0-M6

Author: bsbodden

demos/roms-vss/src/main/java/com/redis/om/vss/domain/Product.java

@@ -79,7 +79,7 @@ public class Product {
            schemaFieldType = SchemaFieldType.VECTOR, //
            algorithm = VectorAlgorithm.HNSW, //
            type = VectorType.FLOAT32, //
-           dimension = 768, //
+           dimension = 384, //
            distanceMetric = DistanceMetric.COSINE, //
            initialCapacity = 10
   )

redis-om-spring/pom.xml

@@ -78,9 +78,9 @@
         <elementary.version>2.0.1</elementary.version>
         <gson.version>2.10.1</gson.version>
         <djl.starter.version>0.26</djl.starter.version>
-        <djl.version>0.27.0</djl.version>
+        <djl.version>0.30.0</djl.version>
         <junit-bom.version>5.10.2</junit-bom.version>
-        <spring-ai.version>1.0.0-M2</spring-ai.version>
+        <spring-ai.version>1.0.0-M6</spring-ai.version>
     </properties>
 
     <dependencyManagement>
@@ -200,7 +200,7 @@
         </dependency>
         <dependency>
             <groupId>org.springframework.ai</groupId>
-            <artifactId>spring-ai-vertex-ai-palm2</artifactId>
+            <artifactId>spring-ai-vertex-ai-embedding</artifactId>
             <version>${spring-ai.version}</version>
             <optional>true</optional>
         </dependency>

redis-om-spring/src/main/java/com/redis/om/spring/RedisAiConfiguration.java

@@ -35,26 +35,29 @@
 import org.springframework.ai.openai.OpenAiEmbeddingOptions;
 import org.springframework.ai.openai.api.OpenAiApi;
 import org.springframework.ai.retry.RetryUtils;
-import org.springframework.ai.vertexai.palm2.VertexAiPaLm2EmbeddingModel;
-import org.springframework.ai.vertexai.palm2.api.VertexAiPaLm2Api;
+import org.springframework.ai.transformers.TransformersEmbeddingModel;
+import org.springframework.ai.vertexai.embedding.VertexAiEmbeddingConnectionDetails;
+import org.springframework.ai.vertexai.embedding.text.VertexAiTextEmbeddingModel;
+import org.springframework.ai.vertexai.embedding.text.VertexAiTextEmbeddingOptions;
 import org.springframework.beans.factory.annotation.Qualifier;
 import org.springframework.beans.factory.annotation.Value;
 import org.springframework.boot.autoconfigure.condition.ConditionalOnMissingBean;
 import org.springframework.boot.autoconfigure.condition.ConditionalOnProperty;
 import org.springframework.boot.context.properties.EnableConfigurationProperties;
 import org.springframework.context.ApplicationContext;
-import org.springframework.context.annotation.*;
+import org.springframework.context.annotation.Bean;
+import org.springframework.context.annotation.Configuration;
+import org.springframework.context.annotation.Primary;
 import org.springframework.lang.Nullable;
 import org.springframework.util.StringUtils;
-import org.springframework.web.client.RestClient;
 import software.amazon.awssdk.auth.credentials.AwsBasicCredentials;
 import software.amazon.awssdk.auth.credentials.AwsCredentials;
 import software.amazon.awssdk.auth.credentials.StaticCredentialsProvider;
 import software.amazon.awssdk.regions.Region;
 
 import java.io.IOException;
 import java.net.InetAddress;
-import java.time.*;
+import java.time.Duration;
 import java.util.Map;
 
 @ConditionalOnProperty(name = "redis.om.spring.ai.enabled")
@@ -70,8 +73,8 @@ public ImageFactory imageFactory() {
   }
 
   @Bean(name = "djlImageEmbeddingModelCriteria")
-  public Criteria<Image, byte[]> imageEmbeddingModelCriteria(RedisOMAiProperties properties) {
-    return Criteria.builder().setTypes(Image.class, byte[].class) //
+  public Criteria<Image, float[]> imageEmbeddingModelCriteria(RedisOMAiProperties properties) {
+    return Criteria.builder().setTypes(Image.class, float[].class) //
         .optEngine(properties.getDjl().getImageEmbeddingModelEngine())  //
         .optModelUrls(properties.getDjl().getImageEmbeddingModelModelUrls()) //
         .build();
@@ -123,7 +126,8 @@ public Criteria<Image, float[]> faceEmbeddingModelCriteria( //
       RedisOMAiProperties properties) {
 
     return Criteria.builder() //
-        .setTypes(Image.class, float[].class).optModelUrls(properties.getDjl().getFaceEmbeddingModelModelUrls()) //
+        .setTypes(Image.class, float[].class) //
+        .optModelUrls(properties.getDjl().getFaceEmbeddingModelModelUrls()) //
         .optModelName(properties.getDjl().getFaceEmbeddingModelName()) //
         .optTranslator(translator) //
         .optEngine(properties.getDjl().getFaceEmbeddingModelEngine()) //
@@ -142,8 +146,9 @@ public ZooModel<Image, float[]> faceEmbeddingModel(
   }
 
   @Bean(name = "djlImageEmbeddingModel")
-  public ZooModel<Image, byte[]> imageModel(
-      @Nullable @Qualifier("djlImageEmbeddingModelCriteria") Criteria<Image, byte[]> criteria) throws MalformedModelException, ModelNotFoundException, IOException {
+  public ZooModel<Image, float[]> imageModel(
+      @Nullable @Qualifier("djlImageEmbeddingModelCriteria") Criteria<Image, float[]> criteria)
+      throws MalformedModelException, ModelNotFoundException, IOException {
     return criteria != null ? ModelZoo.loadModel(criteria) : null;
   }
 
@@ -178,6 +183,28 @@ public HuggingFaceTokenizer sentenceTokenizer(RedisOMAiProperties properties) {
     }
   }
 
+  @Bean(name = "transformersEmbeddingModel")
+  public TransformersEmbeddingModel transformersEmbeddingModel(RedisOMAiProperties properties) {
+    TransformersEmbeddingModel embeddingModel = new TransformersEmbeddingModel();
+    if (properties.getTransformers().getTokenizerResource() != null) {
+      embeddingModel.setTokenizerResource(properties.getTransformers().getTokenizerResource());
+    }
+
+    if (properties.getTransformers().getModelResource() != null) {
+      embeddingModel.setModelResource(properties.getTransformers().getModelResource());
+    }
+
+    if (properties.getTransformers().getResourceCacheDirectory() != null) {
+      embeddingModel.setResourceCacheDirectory(properties.getTransformers().getResourceCacheDirectory());
+    }
+
+    if (!properties.getTransformers().getTokenizerOptions().isEmpty()) {
+      embeddingModel.setTokenizerOptions(properties.getTransformers().getTokenizerOptions());
+    }
+
+    return embeddingModel;
+  }
+
   @ConditionalOnMissingBean
   @Bean
   public OpenAiEmbeddingModel openAITextVectorizer(RedisOMAiProperties properties,
@@ -204,7 +231,7 @@ public OpenAiEmbeddingModel openAITextVectorizer(RedisOMAiProperties properties,
 
       // Rest of the configuration
       return new OpenAiEmbeddingModel(openAiApi, MetadataMode.EMBED,
-          OpenAiEmbeddingOptions.builder().withModel("text-embedding-ada-002").build(),
+          OpenAiEmbeddingOptions.builder().model("text-embedding-ada-002").build(),
           RetryUtils.DEFAULT_RETRY_TEMPLATE);
     } else {
       return null;
@@ -251,7 +278,7 @@ public OpenAIClient azureOpenAIClient(RedisOMAiProperties properties, //
 
   @ConditionalOnMissingBean
   @Bean
-  VertexAiPaLm2EmbeddingModel vertexAiPaLm2EmbeddingModel(RedisOMAiProperties properties, //
+  VertexAiTextEmbeddingModel vertexAiEmbeddingModel(RedisOMAiProperties properties, //
       @Value("${spring.ai.vertex.ai.api-key:}") String apiKey,
       @Value("${spring.ai.vertex.ai.ai.base-url:}") String baseUrl) {
     if (!StringUtils.hasText(apiKey)) {
@@ -281,9 +308,15 @@ VertexAiPaLm2EmbeddingModel vertexAiPaLm2EmbeddingModel(RedisOMAiProperties prop
     }
 
     if (StringUtils.hasText(apiKey) && StringUtils.hasText(baseUrl)) {
-      VertexAiPaLm2Api vertexAiApi = new VertexAiPaLm2Api(baseUrl, apiKey, VertexAiPaLm2Api.DEFAULT_GENERATE_MODEL,
-          VertexAiPaLm2Api.DEFAULT_EMBEDDING_MODEL, RestClient.builder());
-      return new VertexAiPaLm2EmbeddingModel(vertexAiApi);
+
+      VertexAiEmbeddingConnectionDetails connectionDetails = VertexAiEmbeddingConnectionDetails.builder()
+          .projectId(System.getenv("VERTEX_AI_GEMINI_PROJECT_ID")).location(System.getenv("VERTEX_AI_GEMINI_LOCATION"))
+          .build();
+
+      VertexAiTextEmbeddingOptions options = VertexAiTextEmbeddingOptions.builder()
+          .model(VertexAiTextEmbeddingOptions.DEFAULT_MODEL_NAME).build();
+
+      return new VertexAiTextEmbeddingModel(connectionDetails, options);
     } else {
       return null;
     }
@@ -346,7 +379,7 @@ BedrockCohereEmbeddingModel bedrockCohereEmbeddingModel(RedisOMAiProperties prop
     if (!StringUtils.hasText(model)) {
       model = properties.getBedrockCohere().getModel();
       if (!StringUtils.hasText(model)) {
-        model = CohereEmbeddingModel.COHERE_EMBED_MULTILINGUAL_V1.id();
+        model = CohereEmbeddingModel.COHERE_EMBED_MULTILINGUAL_V3.id();
         properties.getBedrockCohere().setModel(model);
       }
     }
@@ -439,19 +472,19 @@ BedrockTitanEmbeddingModel bedrockTitanEmbeddingModel(RedisOMAiProperties proper
   @Primary
   @Bean(name = "featureExtractor")
   public Embedder featureExtractor(
-      @Nullable @Qualifier("djlImageEmbeddingModel") ZooModel<Image, byte[]> imageEmbeddingModel,
+      @Nullable @Qualifier("djlImageEmbeddingModel") ZooModel<Image, float[]> imageEmbeddingModel,
       @Nullable @Qualifier("djlFaceEmbeddingModel") ZooModel<Image, float[]> faceEmbeddingModel,
       @Nullable @Qualifier("djlImageFactory") ImageFactory imageFactory,
       @Nullable @Qualifier("djlDefaultImagePipeline") Pipeline defaultImagePipeline,
-      @Nullable @Qualifier("djlSentenceTokenizer") HuggingFaceTokenizer sentenceTokenizer,
+      @Nullable @Qualifier("transformersEmbeddingModel") TransformersEmbeddingModel transformersEmbeddingModel,
       @Nullable OpenAiEmbeddingModel openAITextVectorizer, @Nullable OpenAIClient azureOpenAIClient,
-      @Nullable VertexAiPaLm2EmbeddingModel vertexAiPaLm2EmbeddingModel,
+      @Nullable VertexAiTextEmbeddingModel vertexAiTextEmbeddingModel,
       @Nullable BedrockCohereEmbeddingModel bedrockCohereEmbeddingModel,
       @Nullable BedrockTitanEmbeddingModel bedrockTitanEmbeddingModel,
       RedisOMAiProperties properties,
       ApplicationContext ac) {
     return new DefaultEmbedder(ac, imageEmbeddingModel, faceEmbeddingModel, imageFactory, defaultImagePipeline,
-            sentenceTokenizer, openAITextVectorizer, azureOpenAIClient, vertexAiPaLm2EmbeddingModel,
+        transformersEmbeddingModel, openAITextVectorizer, azureOpenAIClient, vertexAiTextEmbeddingModel,
             bedrockCohereEmbeddingModel, bedrockTitanEmbeddingModel, properties);
   }
 }

redis-om-spring/src/main/java/com/redis/om/spring/RedisOMAiProperties.java

@@ -5,13 +5,17 @@
 import org.springframework.boot.autoconfigure.condition.ConditionalOnProperty;
 import org.springframework.boot.context.properties.ConfigurationProperties;
 
+import java.util.HashMap;
+import java.util.Map;
+
 @ConditionalOnProperty(name = "redis.om.spring.ai.enabled")
 @ConfigurationProperties(
     prefix = "redis.om.spring.ai", ignoreInvalidFields = true
 )
 public class RedisOMAiProperties {
   private boolean enabled = false;
   private final Djl djl = new Djl();
+  private final Transformers transformers = new Transformers();
   private final OpenAi openAi = new OpenAi();
   private final AzureOpenAi azureOpenAi = new AzureOpenAi();
   private final VertexAi vertexAi = new VertexAi();
@@ -31,6 +35,10 @@ public Djl getDjl() {
     return djl;
   }
 
+  public Transformers getTransformers() {
+    return transformers;
+  }
+
   public OpenAi getOpenAi() {
     return openAi;
   }
@@ -55,6 +63,30 @@ public Ollama getOllama() {
     return ollama;
   }
 
+  // Transformer properties
+  public static class Transformers {
+    private String tokenizerResource;
+    private String modelResource;
+    private String resourceCacheDirectory;
+    private Map<String, String> tokenizerOptions = new HashMap<>();
+
+    public String getTokenizerResource() {
+      return tokenizerResource;
+    }
+
+    public String getModelResource() {
+      return modelResource;
+    }
+
+    public String getResourceCacheDirectory() {
+      return resourceCacheDirectory;
+    }
+
+    public Map<String, String> getTokenizerOptions() {
+      return tokenizerOptions;
+    }
+  }
+
   // DJL properties
   public static class Djl {
     private static final String DEFAULT_ENGINE = "PyTorch";
@@ -73,7 +105,7 @@ public static class Djl {
     @NotNull
     private String sentenceTokenizerModelMaxLength = "768";
     @NotNull
-    private String sentenceTokenizerModel = "sentence-transformers/all-mpnet-base-v2";
+    private String sentenceTokenizerModel = "sentence-transformers/msmarco-distilbert-dot-v5";
 
     // face detection
     @NotNull
@@ -91,6 +123,7 @@ public static class Djl {
     @NotNull
     private String faceEmbeddingModelModelUrls = "https://resources.djl.ai/test-models/pytorch/face_feature.zip";
 
+
     public Djl() {
     }
 
@@ -278,6 +311,24 @@ public static class VertexAi {
     private String apiKey;
     private String endPoint;
     private String model;
+    private String projectId;
+    private String location;
+
+    public String getProjectId() {
+      return projectId;
+    }
+
+    public void setProjectId(String projectId) {
+      this.projectId = projectId;
+    }
+
+    public String getLocation() {
+      return location;
+    }
+
+    public void setLocation(String location) {
+      this.location = location;
+    }
 
     public String getApiKey() {
       return apiKey;

redis-om-spring/src/main/java/com/redis/om/spring/annotations/EmbeddingProvider.java

@@ -1,6 +1,7 @@
 package com.redis.om.spring.annotations;
 
 public enum EmbeddingProvider {
+  TRANSFORMERS,
   DJL,
   OPENAI,
   OLLAMA,

redis-om-spring/src/main/java/com/redis/om/spring/annotations/Vectorize.java

@@ -4,7 +4,6 @@
 import org.springframework.ai.bedrock.titan.api.TitanEmbeddingBedrockApi.TitanEmbeddingModel;
 import org.springframework.ai.ollama.api.OllamaModel;
 import org.springframework.ai.openai.api.OpenAiApi.EmbeddingModel;
-import org.springframework.ai.vertexai.palm2.api.VertexAiPaLm2Api;
 
 import java.lang.annotation.*;
 
@@ -16,17 +15,17 @@
 
   EmbeddingType embeddingType() default EmbeddingType.SENTENCE;
 
-  EmbeddingProvider provider() default EmbeddingProvider.DJL;
+  EmbeddingProvider provider() default EmbeddingProvider.TRANSFORMERS;
 
   EmbeddingModel openAiEmbeddingModel() default EmbeddingModel.TEXT_EMBEDDING_ADA_002;
 
   OllamaModel ollamaEmbeddingModel() default OllamaModel.MISTRAL;
 
   String azureOpenAiDeploymentName() default "text-embedding-ada-002";
 
-  String vertexAiPaLm2ApiModel() default VertexAiPaLm2Api.DEFAULT_EMBEDDING_MODEL;
+  String vertexAiPaLm2ApiModel() default "text-embedding-004";
 
-  CohereEmbeddingModel cohereEmbeddingModel() default CohereEmbeddingModel.COHERE_EMBED_MULTILINGUAL_V1;
+  CohereEmbeddingModel cohereEmbeddingModel() default CohereEmbeddingModel.COHERE_EMBED_MULTILINGUAL_V3;
 
   TitanEmbeddingModel titanEmbeddingModel() default TitanEmbeddingModel.TITAN_EMBED_IMAGE_V1;
 }