Adding VoyageAI integration

PACKAGES.md

@@ -39,6 +39,9 @@ A BOM is provided that can be used to define the versions of all Semantic Kernel
 `semantickernel-aiservices-openai`
 : Provides a connector that can be used to interact with the OpenAI API.
 
+`semantickernel-aiservices-voyageai`
+: Provides connectors for VoyageAI's embedding and reranking services, including text embeddings, contextualized embeddings, multimodal embeddings, and document reranking.
+
 ## Example Configurations
 
 ### Example: OpenAI + SQLite
@@ -72,5 +75,36 @@ POM XML for a simple project that uses OpenAI.
 </project>
 ```
 
+### Example: VoyageAI Embeddings and Reranking
+
+POM XML for a project that uses VoyageAI for embeddings and reranking.
+
+```xml
+
+<project>
+    <dependencyManagement>
+        <dependencies>
+            <dependency>
+                <groupId>com.microsoft.semantic-kernel</groupId>
+                <artifactId>semantickernel-bom</artifactId>
+                <version>${semantickernel.version}</version>
+                <scope>import</scope>
+                <type>pom</type>
+            </dependency>
+        </dependencies>
+    </dependencyManagement>
+    <dependencies>
+        <dependency>
+            <groupId>com.microsoft.semantic-kernel</groupId>
+            <artifactId>semantickernel-api</artifactId>
+        </dependency>
+        <dependency>
+            <groupId>com.microsoft.semantic-kernel</groupId>
+            <artifactId>semantickernel-aiservices-voyageai</artifactId>
+        </dependency>
+    </dependencies>
+</project>
+```
+
 
 

aiservices/voyageai/pom.xml

@@ -0,0 +1,89 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
+    <modelVersion>4.0.0</modelVersion>
+    <parent>
+        <groupId>com.microsoft.semantic-kernel</groupId>
+        <artifactId>semantickernel-parent</artifactId>
+        <version>1.4.4-RC3-SNAPSHOT</version>
+        <relativePath>../../pom.xml</relativePath>
+    </parent>
+
+    <artifactId>semantickernel-aiservices-voyageai</artifactId>
+    <name>Semantic Kernel VoyageAI Services</name>
+    <description>VoyageAI services for Semantic Kernel</description>
+
+    <dependencies>
+        <dependency>
+            <groupId>com.microsoft.semantic-kernel</groupId>
+            <artifactId>semantickernel-api</artifactId>
+        </dependency>
+        <dependency>
+            <groupId>com.microsoft.semantic-kernel</groupId>
+            <artifactId>semantickernel-api-builders</artifactId>
+        </dependency>
+        <dependency>
+            <groupId>com.microsoft.semantic-kernel</groupId>
+            <artifactId>semantickernel-api-ai-services</artifactId>
+        </dependency>
+        <dependency>
+            <groupId>com.microsoft.semantic-kernel</groupId>
+            <artifactId>semantickernel-api-textembedding-services</artifactId>
+        </dependency>
+        <dependency>
+            <groupId>com.microsoft.semantic-kernel</groupId>
+            <artifactId>semantickernel-api-exceptions</artifactId>
+        </dependency>
+        <dependency>
+            <groupId>com.microsoft.semantic-kernel</groupId>
+            <artifactId>semantickernel-api-localization</artifactId>
+        </dependency>
+
+        <dependency>
+            <groupId>com.fasterxml.jackson.core</groupId>
+            <artifactId>jackson-databind</artifactId>
+            <scope>compile</scope>
+        </dependency>
+        <dependency>
+            <groupId>com.fasterxml.jackson.core</groupId>
+            <artifactId>jackson-core</artifactId>
+            <scope>compile</scope>
+        </dependency>
+        <dependency>
+            <groupId>com.fasterxml.jackson.core</groupId>
+            <artifactId>jackson-annotations</artifactId>
+            <scope>compile</scope>
+        </dependency>
+
+        <!-- HTTP Client -->
+        <dependency>
+            <groupId>com.squareup.okhttp3</groupId>
+            <artifactId>okhttp</artifactId>
+            <version>4.12.0</version>
+        </dependency>
+
+        <!-- Reactive Streams -->
+        <dependency>
+            <groupId>io.projectreactor</groupId>
+            <artifactId>reactor-core</artifactId>
+        </dependency>
+
+        <!-- Logging -->
+        <dependency>
+            <groupId>org.slf4j</groupId>
+            <artifactId>slf4j-api</artifactId>
+        </dependency>
+
+        <!-- Test Dependencies -->
+        <dependency>
+            <groupId>org.junit.jupiter</groupId>
+            <artifactId>junit-jupiter</artifactId>
+            <scope>test</scope>
+        </dependency>
+        <dependency>
+            <groupId>org.mockito</groupId>
+            <artifactId>mockito-core</artifactId>
+            <scope>test</scope>
+        </dependency>
+    </dependencies>
+
+</project>

aiservices/voyageai/src/main/java/com/microsoft/semantickernel/aiservices/voyageai/contextualizedembedding/VoyageAIContextualizedEmbeddingGenerationService.java

@@ -0,0 +1,204 @@
+// Copyright (c) Microsoft. All rights reserved.
+package com.microsoft.semantickernel.aiservices.voyageai.contextualizedembedding;
+
+import com.microsoft.semantickernel.aiservices.voyageai.core.VoyageAIClient;
+import com.microsoft.semantickernel.aiservices.voyageai.core.VoyageAIModels;
+import com.microsoft.semantickernel.orchestration.PromptExecutionSettings;
+import com.microsoft.semantickernel.services.textembedding.Embedding;
+import com.microsoft.semantickernel.services.textembedding.TextEmbeddingGenerationService;
+import org.slf4j.Logger;
+import org.slf4j.LoggerFactory;
+import reactor.core.publisher.Mono;
+
+import javax.annotation.Nullable;
+import java.util.ArrayList;
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.Comparator;
+import java.util.List;
+import java.util.stream.Collectors;
+
+/**
+ * VoyageAI contextualized embedding generation service.
+ * Generates embeddings that capture both local chunk details and global document-level metadata.
+ * Supports models like voyage-3.
+ */
+public class VoyageAIContextualizedEmbeddingGenerationService implements TextEmbeddingGenerationService {
+
+    private static final Logger LOGGER = LoggerFactory.getLogger(VoyageAIContextualizedEmbeddingGenerationService.class);
+
+    private final VoyageAIClient client;
+    private final String modelId;
+    private final String serviceId;
+
+    /**
+     * Creates a new instance of VoyageAI contextualized embedding generation service.
+     *
+     * @param client    VoyageAI client
+     * @param modelId   Model ID (e.g., "voyage-3")
+     * @param serviceId Optional service ID
+     */
+    public VoyageAIContextualizedEmbeddingGenerationService(
+        VoyageAIClient client,
+        String modelId,
+        @Nullable String serviceId) {
+
+        if (client == null) {
+            throw new IllegalArgumentException("Client cannot be null");
+        }
+        if (modelId == null || modelId.trim().isEmpty()) {
+            throw new IllegalArgumentException("Model ID cannot be null or empty");
+        }
+
+        this.client = client;
+        this.modelId = modelId;
+        this.serviceId = serviceId != null ? serviceId : PromptExecutionSettings.DEFAULT_SERVICE_ID;
+    }
+
+    @Override
+    public String getServiceId() {
+        return serviceId;
+    }
+
+    @Override
+    public String getModelId() {
+        return modelId;
+    }
+
+    /**
+     * Generates contextualized embeddings for document chunks.
+     *
+     * @param inputs List of lists where each inner list contains document chunks
+     * @return A Mono containing a list of embeddings for all chunks across all documents
+     */
+    public Mono<List<Embedding>> generateContextualizedEmbeddingsAsync(List<List<String>> inputs) {
+        if (inputs == null || inputs.isEmpty()) {
+            return Mono.just(Collections.emptyList());
+        }
+
+        LOGGER.debug("Generating contextualized embeddings for {} document groups using model {}",
+            inputs.size(), modelId);
+
+        VoyageAIModels.ContextualizedEmbeddingRequest request =
+            new VoyageAIModels.ContextualizedEmbeddingRequest();
+        request.setInputs(inputs);
+        request.setModel(modelId);
+
+        return client.sendRequestAsync(
+            "contextualizedembeddings",
+            request,
+            VoyageAIModels.ContextualizedEmbeddingResponse.class)
+            .map(response -> {
+                List<Embedding> embeddings = new ArrayList<>();
+                // Parse nested data structure: {"data":[{"data":[{"embedding":[...]}]}]}
+                for (VoyageAIModels.ContextualizedEmbeddingDataList dataList : response.getData()) {
+                    for (VoyageAIModels.EmbeddingDataItem item : dataList.getData()) {
+                        embeddings.add(new Embedding(item.getEmbedding()));
+                    }
+                }
+
+                LOGGER.debug("Received {} contextualized embeddings from VoyageAI", embeddings.size());
+                return embeddings;
+            });
+    }
+
+    /**
+     * Generates embeddings for the given text.
+     * For standard text embedding, wraps the data as a single input.
+     *
+     * @param data The text to generate embeddings for
+     * @return A Mono that completes with the embedding
+     */
+    @Override
+    public Mono<Embedding> generateEmbeddingAsync(String data) {
+        return generateEmbeddingsAsync(Arrays.asList(data))
+            .flatMap(embeddings -> {
+                if (embeddings.isEmpty()) {
+                    return Mono.empty();
+                }
+                return Mono.just(embeddings.get(0));
+            });
+    }
+
+    /**
+     * Generates embeddings for the given texts.
+     * Each text is treated as a separate document for contextualized embeddings.
+     *
+     * @param data The texts to generate embeddings for
+     * @return A Mono that completes with the list of embeddings
+     */
+    @Override
+    public Mono<List<Embedding>> generateEmbeddingsAsync(List<String> data) {
+        if (data == null || data.isEmpty()) {
+            return Mono.just(Collections.emptyList());
+        }
+
+        // Convert each string to a single-element list for contextualized embeddings
+        List<List<String>> inputs = new ArrayList<>();
+        for (String text : data) {
+            inputs.add(Arrays.asList(text));
+        }
+
+        return generateContextualizedEmbeddingsAsync(inputs);
+    }
+
+    /**
+     * Creates a builder for VoyageAI contextualized embedding generation service.
+     *
+     * @return A new builder instance
+     */
+    public static Builder builder() {
+        return new Builder();
+    }
+
+    /**
+     * Builder for {@link VoyageAIContextualizedEmbeddingGenerationService}.
+     */
+    public static class Builder {
+        private VoyageAIClient client;
+        private String modelId;
+        private String serviceId;
+
+        /**
+         * Sets the VoyageAI client.
+         *
+         * @param client VoyageAI client
+         * @return This builder
+         */
+        public Builder withClient(VoyageAIClient client) {
+            this.client = client;
+            return this;
+        }
+
+        /**
+         * Sets the model ID.
+         *
+         * @param modelId Model ID (e.g., "voyage-3")
+         * @return This builder
+         */
+        public Builder withModelId(String modelId) {
+            this.modelId = modelId;
+            return this;
+        }
+
+        /**
+         * Sets the service ID.
+         *
+         * @param serviceId Service ID
+         * @return This builder
+         */
+        public Builder withServiceId(String serviceId) {
+            this.serviceId = serviceId;
+            return this;
+        }
+
+        /**
+         * Builds the VoyageAI contextualized embedding generation service.
+         *
+         * @return A new instance of VoyageAIContextualizedEmbeddingGenerationService
+         */
+        public VoyageAIContextualizedEmbeddingGenerationService build() {
+            return new VoyageAIContextualizedEmbeddingGenerationService(client, modelId, serviceId);
+        }
+    }
+}