feat: add voyage embeddings model

.env

@@ -6,6 +6,9 @@ OPENAI_API_KEY=
 # For using Claude on Anthropic
 ANTHROPIC_API_KEY=
 
+# For using Voyage
+VOYAGE_API_KEY=
+
 # For using GPT on Azure
 AZURE_OPENAI_BASEURL=
 AZURE_OPENAI_DEPLOYMENT=

README.md

@@ -29,15 +29,17 @@ Currently supported models and platforms:
 | Vendor         | Model                  | Platform             |
 |----------------|------------------------|----------------------|
 | **OpenAI**     | - GPT<br/>- Embeddings | - OpenAI<br/>- Azure |
-| **Anthropic**  | - Claude | - Anthropic          |
+| **Anthropic**  | - Claude               | - Anthropic          |
+| **Voyage**     | - Voyage               | - Voyage             |
 
 Planned Models & Platforms (not implemented yet):
 
-| Vendor         | Model                  | Platform                         |
-|----------------|------------------------|----------------------------------|
-| **Anthropic**  | - Voyage | - GPC<br/>- AWS                  |
-| **Google**     | - Gemini<br/>- Gemma | - GPC                            |
-| **Meta**       | - Llama | - Meta AI<br/>- GPC<br/>- Ollama |
+| Vendor         | Model                     | Platform                         |
+|----------------|---------------------------|----------------------------------|
+| **Anthropic**  | - Claude                  | - GPC<br/>- AWS                  |
+| **Voyage**     | - Voyage                  | - AWS                            |
+| **Google**     | - Gemini<br/>- Gemma      | - GPC                            |
+| **Meta**       | - Llama                   | - Meta AI<br/>- GPC<br/>- Ollama |
 | **Mistral AI** | - Mistral<br/>- Codestral | - Mistral<br/>- GPT<br/>- Ollama |
 
 Supported Stores
@@ -72,13 +74,13 @@ To run all examples, just use `make run-all-examples`.
 
 ### Chat Examples
 
-1. Chat Example: OpenAI's GPT
+1. OpenAI's GPT
    ```bash
    export OPENAI_API_KEY=sk-...
    php examples/chat-gpt-openai.php
    ```
 
-1. Chat Example: OpenAI's GPT With Azure
+1. OpenAI's GPT with Azure
    ```bash
    export AZURE_OPENAI_BASEURL=... // e.g. your-resource.openai.azure.com
    export AZURE_OPENAI_DEPLOYMENT=...
@@ -87,12 +89,26 @@ To run all examples, just use `make run-all-examples`.
    php examples/chat-gpt-azure.php
    ```
 
-1. Chat Example: Anthropic's Claude
+1. Anthropic's Claude
    ```bash
    export ANTHROPIC_API_KEY=sk-...
    php examples/chat-claude-anthropic.php
    ```
 
+### Embeddings Examples
+
+1. OpenAI's Emebddings
+   ```bash
+   export OPENAI_API_KEY=sk-...
+   php examples/embeddings-openai.php
+   ```
+
+1. Voyage's Embeddings
+   ```bash
+   export VOYAGE_API_KEY=sk-...
+   php examples/embeddings-voyage.php
+    ```
+
 ### Tool Examples
 
 1. Simple Clock Tool

examples/embeddings-openai.php

@@ -0,0 +1,21 @@
+<?php
+
+use PhpLlm\LlmChain\OpenAI\Model\Embeddings;
+use PhpLlm\LlmChain\OpenAI\Model\Embeddings\Version;
+use PhpLlm\LlmChain\OpenAI\Platform\OpenAI;
+use Symfony\Component\Dotenv\Dotenv;
+use Symfony\Component\HttpClient\HttpClient;
+
+require_once dirname(__DIR__).'/vendor/autoload.php';
+(new Dotenv())->loadEnv(dirname(__DIR__).'/.env');
+
+$platform = new OpenAI(HttpClient::create(), $_ENV['OPENAI_API_KEY']);
+$embeddings = new Embeddings($platform, Version::textEmbedding3Small());
+
+$vector = $embeddings->create(<<<TEXT
+    Once upon a time, there was a country called Japan. It was a beautiful country with a lot of mountains and rivers.
+    The people of Japan were very kind and hardworking. They loved their country very much and took care of it. The
+    country was very peaceful and prosperous. The people lived happily ever after.
+    TEXT);
+
+echo $vector->getDimensions().PHP_EOL;

examples/embeddings-voyage.php

@@ -0,0 +1,20 @@
+<?php
+
+use PhpLlm\LlmChain\Voyage\Model\Voyage;
+use PhpLlm\LlmChain\Voyage\Platform\Voyage as VoyagePlatform;
+use Symfony\Component\Dotenv\Dotenv;
+use Symfony\Component\HttpClient\HttpClient;
+
+require_once dirname(__DIR__).'/vendor/autoload.php';
+(new Dotenv())->loadEnv(dirname(__DIR__).'/.env');
+
+$platform = new VoyagePlatform(HttpClient::create(), $_ENV['VOYAGE_API_KEY']);
+$embeddings = new Voyage($platform);
+
+$vector = $embeddings->create(<<<TEXT
+    Once upon a time, there was a country called Japan. It was a beautiful country with a lot of mountains and rivers.
+    The people of Japan were very kind and hardworking. They loved their country very much and took care of it. The
+    country was very peaceful and prosperous. The people lived happily ever after.
+    TEXT);
+
+echo $vector->getDimensions().PHP_EOL;

src/Anthropic/Model/Claude.php

@@ -4,8 +4,8 @@
 
 namespace PhpLlm\LlmChain\Anthropic\Model;
 
-use PhpLlm\LlmChain\Anthropic\ClaudePlatform;
 use PhpLlm\LlmChain\Anthropic\Model\Claude\Version;
+use PhpLlm\LlmChain\Anthropic\Platform;
 use PhpLlm\LlmChain\LanguageModel;
 use PhpLlm\LlmChain\Message\MessageBag;
 use PhpLlm\LlmChain\Response\Choice;
@@ -14,19 +14,19 @@
 final class Claude implements LanguageModel
 {
     /**
-     * @param array<mixed> $options The default options for the model usage
+     * @param array<string, mixed> $options The default options for the model usage
      */
     public function __construct(
-        private readonly ClaudePlatform $platform,
+        private readonly Platform $platform,
         private ?Version $version = null,
         private readonly array $options = ['temperature' => 1.0, 'max_tokens' => 1000],
     ) {
         $this->version ??= Version::sonnet35();
     }
 
     /**
-     * @param array<mixed> $options The options to be used for this specific call.
-     *                              Can overwrite default options.
+     * @param array<string, mixed> $options The options to be used for this specific call.
+     *                                      Can overwrite default options.
      */
     public function call(MessageBag $messages, array $options = []): Response
     {

src/Anthropic/ClaudePlatform.php → src/Anthropic/Platform.php

@@ -4,7 +4,7 @@
 
 namespace PhpLlm\LlmChain\Anthropic;
 
-interface ClaudePlatform
+interface Platform
 {
     /**
      * @param array<string, mixed> $body

src/Anthropic/Platform/Anthropic.php

@@ -4,10 +4,10 @@
 
 namespace PhpLlm\LlmChain\Anthropic\Platform;
 
-use PhpLlm\LlmChain\Anthropic\ClaudePlatform;
+use PhpLlm\LlmChain\Anthropic\Platform;
 use Symfony\Contracts\HttpClient\HttpClientInterface;
 
-final readonly class Anthropic implements ClaudePlatform
+final readonly class Anthropic implements Platform
 {
     public function __construct(
         private HttpClientInterface $httpClient,

src/DocumentEmbedder.php

@@ -10,7 +10,7 @@
 final readonly class DocumentEmbedder
 {
     public function __construct(
-        private EmbeddingModel $embeddings,
+        private EmbeddingsModel $embeddings,
         private StoreInterface $store,
     ) {
     }

src/EmbeddingsModel.php

@@ -0,0 +1,23 @@
+<?php
+
+declare(strict_types=1);
+
+namespace PhpLlm\LlmChain;
+
+use PhpLlm\LlmChain\Document\Vector;
+
+interface EmbeddingsModel
+{
+    /**
+     * @param array<string, mixed> $options
+     */
+    public function create(string $text, array $options = []): Vector;
+
+    /**
+     * @param list<string>         $texts
+     * @param array<string, mixed> $options
+     *
+     * @return Vector[]
+     */
+    public function multiCreate(array $texts, array $options = []): array;
+}

src/OpenAI/Model/Embeddings.php

@@ -5,11 +5,11 @@
 namespace PhpLlm\LlmChain\OpenAI\Model;
 
 use PhpLlm\LlmChain\Document\Vector;
-use PhpLlm\LlmChain\EmbeddingModel;
+use PhpLlm\LlmChain\EmbeddingsModel;
 use PhpLlm\LlmChain\OpenAI\Model\Embeddings\Version;
 use PhpLlm\LlmChain\OpenAI\Platform;
 
-final class Embeddings implements EmbeddingModel
+final class Embeddings implements EmbeddingsModel
 {
     public function __construct(
         private readonly Platform $platform,
@@ -18,14 +18,14 @@ public function __construct(
         $this->version ??= Version::textEmbedding3Small();
     }
 
-    public function create(string $text): Vector
+    public function create(string $text, array $options = []): Vector
     {
         $response = $this->platform->request('embeddings', $this->createBody($text));
 
         return $this->extractVector($response);
     }
 
-    public function multiCreate(array $texts): array
+    public function multiCreate(array $texts, array $options = []): array
     {
         $bodies = array_map([$this, 'createBody'], $texts);
 

src/ToolBox/Tool/SimilaritySearch.php

@@ -5,7 +5,7 @@
 namespace PhpLlm\LlmChain\ToolBox\Tool;
 
 use PhpLlm\LlmChain\Document\Document;
-use PhpLlm\LlmChain\EmbeddingModel;
+use PhpLlm\LlmChain\EmbeddingsModel;
 use PhpLlm\LlmChain\Store\VectorStoreInterface;
 use PhpLlm\LlmChain\ToolBox\AsTool;
 
@@ -18,7 +18,7 @@ final class SimilaritySearch
     public array $usedDocuments = [];
 
     public function __construct(
-        private readonly EmbeddingModel $embeddings,
+        private readonly EmbeddingsModel $embeddings,
         private readonly VectorStoreInterface $vectorStore,
     ) {
     }

src/Voyage/Model/Voyage.php

@@ -0,0 +1,37 @@
+<?php
+
+declare(strict_types=1);
+
+namespace PhpLlm\LlmChain\Voyage\Model;
+
+use PhpLlm\LlmChain\Document\Vector;
+use PhpLlm\LlmChain\EmbeddingsModel;
+use PhpLlm\LlmChain\Voyage\Model\Voyage\Version;
+use PhpLlm\LlmChain\Voyage\Platform;
+
+final class Voyage implements EmbeddingsModel
+{
+    public function __construct(
+        private readonly Platform $platform,
+        private ?Version $version = null,
+    ) {
+        $this->version ??= Version::v3();
+    }
+
+    public function create(string $text, array $options = []): Vector
+    {
+        $vectors = $this->multiCreate([$text], $options);
+
+        return $vectors[0];
+    }
+
+    public function multiCreate(array $texts, array $options = []): array
+    {
+        $response = $this->platform->request(array_merge($options, [
+            'model' => $this->version->name,
+            'input' => $texts,
+        ]));
+
+        return array_map(fn (array $data) => new Vector($data['embedding']), $response['data']);
+    }
+}

src/Voyage/Model/Voyage/Version.php

@@ -0,0 +1,46 @@
+<?php
+
+declare(strict_types=1);
+
+namespace PhpLlm\LlmChain\Voyage\Model\Voyage;
+
+use Webmozart\Assert\Assert;
+
+final readonly class Version
+{
+    public function __construct(
+        public string $name,
+    ) {
+        Assert::stringNotEmpty($name);
+    }
+
+    public static function v3(): self
+    {
+        return new self('voyage-3');
+    }
+
+    public static function v3lite(): self
+    {
+        return new self('voyage-3-lite');
+    }
+
+    public static function finance2(): self
+    {
+        return new self('voyage-finance-2');
+    }
+
+    public static function multilingual2(): self
+    {
+        return new self('voyage-multilingual-2');
+    }
+
+    public static function law2(): self
+    {
+        return new self('voyage-law-2');
+    }
+
+    public static function code2(): self
+    {
+        return new self('voyage-code-2');
+    }
+}

src/Voyage/Platform.php

@@ -0,0 +1,15 @@
+<?php
+
+declare(strict_types=1);
+
+namespace PhpLlm\LlmChain\Voyage;
+
+interface Platform
+{
+    /**
+     * @param array<string, mixed> $body
+     *
+     * @return array<string, mixed>
+     */
+    public function request(array $body): array;
+}