refactor: extract logic from blog embed command

src/Blog/Embedder.php

@@ -0,0 +1,60 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\Blog;
+
+use Codewithkyrian\ChromaDB\Client;
+use PhpLlm\LlmChain\Bridge\OpenAI\Embeddings;
+use PhpLlm\LlmChain\Document\Vector;
+use PhpLlm\LlmChain\Model\Response\AsyncResponse;
+use PhpLlm\LlmChain\Model\Response\VectorResponse;
+use PhpLlm\LlmChain\PlatformInterface;
+
+final readonly class Embedder
+{
+    public function __construct(
+        private Loader $loader,
+        private PlatformInterface $platform,
+        private Client $chromaClient,
+    ) {
+    }
+
+    public function embedBlog(): void
+    {
+        $posts = $this->loader->load();
+        $vectors = $this->createEmbeddings($posts);
+        $this->pushToChromaDB($posts, $vectors);
+    }
+
+    /**
+     * @param Post[] $posts
+     *
+     * @return Vector[]
+     */
+    private function createEmbeddings(array $posts): array
+    {
+        $texts = array_map(fn (Post $post) => $post->toString(), $posts);
+        $response = $this->platform->request(new Embeddings(), $texts);
+
+        assert($response instanceof AsyncResponse);
+        $response = $response->unwrap();
+        assert($response instanceof VectorResponse);
+
+        return $response->getContent();
+    }
+
+    /**
+     * @param Post[]   $posts
+     * @param Vector[] $vectors
+     */
+    private function pushToChromaDB(array $posts, array $vectors): void
+    {
+        $collection = $this->chromaClient->getOrCreateCollection('symfony_blog');
+
+        $ids = array_map(fn (Post $post) => $post->id, $posts);
+        $vectors = array_map(fn (Vector $vector) => $vector->getData(), $vectors);
+
+        $collection->upsert($ids, $vectors, $posts);
+    }
+}

src/Blog/Loader.php

@@ -0,0 +1,42 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\Blog;
+
+use Symfony\Component\DomCrawler\Crawler;
+use Symfony\Component\Uid\Uuid;
+use Symfony\Contracts\HttpClient\HttpClientInterface;
+
+class Loader
+{
+    public function __construct(
+        private HttpClientInterface $httpClient,
+    ) {
+    }
+
+    /**
+     * @return Post[]
+     */
+    public function load(): array
+    {
+        $response = $this->httpClient->request('GET', 'https://feeds.feedburner.com/symfony/blog');
+
+        $posts = [];
+        $crawler = new Crawler($response->getContent());
+        $crawler->filter('item')->each(function (Crawler $node) use (&$posts) {
+            $title = $node->filter('title')->text();
+            $posts[] = new Post(
+                Uuid::v5(Uuid::fromString('6ba7b810-9dad-11d1-80b4-00c04fd430c8'), $title),
+                $title,
+                $node->filter('link')->text(),
+                $node->filter('description')->text(),
+                (new Crawler($node->filter('content\:encoded')->text()))->text(),
+                $node->filter('dc\:creator')->text(),
+                new \DateTimeImmutable($node->filter('pubDate')->text()),
+            );
+        });
+
+        return $posts;
+    }
+}

src/Blog/Post.php

@@ -0,0 +1,55 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\Blog;
+
+use Symfony\Component\Uid\Uuid;
+
+final readonly class Post
+{
+    public function __construct(
+        public Uuid $id,
+        public string $title,
+        public string $link,
+        public string $description,
+        public string $content,
+        public string $author,
+        public \DateTimeImmutable $date,
+    ) {
+    }
+
+    public function toString(): string
+    {
+        return <<<TEXT
+            Title: {$this->title}
+            From: {$this->author} on {$this->date->format('Y-m-d')}
+            Description: {$this->description}
+            {$this->content}
+            TEXT;
+    }
+
+    /**
+     * @return array{
+     *     id: string,
+     *     title: string,
+     *     link: string,
+     *     description: string,
+     *     content: string,
+     *     author: string,
+     *     date: string,
+     * }
+     */
+    public function toArray(): array
+    {
+        return [
+            'id' => $this->id->toRfc4122(),
+            'title' => $this->title,
+            'link' => $this->link,
+            'description' => $this->description,
+            'content' => $this->content,
+            'author' => $this->author,
+            'date' => $this->date->format('Y-m-d'),
+        ];
+    }
+}

src/Command/BlogEmbedCommand.php

@@ -4,39 +4,18 @@
 
 namespace App\Command;
 
-use Codewithkyrian\ChromaDB\Client;
-use PhpLlm\LlmChain\Bridge\OpenAI\Embeddings;
-use PhpLlm\LlmChain\Document\Vector;
-use PhpLlm\LlmChain\Model\Response\AsyncResponse;
-use PhpLlm\LlmChain\Model\Response\VectorResponse;
-use PhpLlm\LlmChain\PlatformInterface;
+use App\Blog\Embedder;
 use Symfony\Component\Console\Attribute\AsCommand;
 use Symfony\Component\Console\Command\Command;
 use Symfony\Component\Console\Input\InputInterface;
 use Symfony\Component\Console\Output\OutputInterface;
 use Symfony\Component\Console\Style\SymfonyStyle;
-use Symfony\Component\DomCrawler\Crawler;
-use Symfony\Component\Uid\Uuid;
-use Symfony\Contracts\HttpClient\HttpClientInterface;
 
-/**
- * @phpstan-type Post array{
- *     id: Uuid,
- *     title: string,
- *     link: string,
- *     description: string,
- *     content: string,
- *     author: string,
- *     date: \DateTimeImmutable,
- * }
- */
 #[AsCommand('app:blog:embed', description: 'Create embeddings for Symfony blog and push to ChromaDB.')]
 final class BlogEmbedCommand extends Command
 {
     public function __construct(
-        private readonly HttpClientInterface $httpClient,
-        private readonly PlatformInterface $platform,
-        private readonly Client $chromaClient,
+        private readonly Embedder $embedder,
     ) {
         parent::__construct();
     }
@@ -46,85 +25,10 @@ protected function execute(InputInterface $input, OutputInterface $output): int
         $io = new SymfonyStyle($input, $output);
         $io->title('Loading RSS of Symfony blog as embeddings into ChromaDB');
 
-        $posts = $this->loadBlogPosts();
-        $vectors = $this->createEmbeddings($posts);
-        $this->pushToChromaDB($posts, $vectors);
+        $this->embedder->embedBlog();
 
         $io->success('Symfony Blog Successfully Embedded!');
 
         return Command::SUCCESS;
     }
-
-    /**
-     * @return list<array{
-     *     id: Uuid,
-     *     title: string,
-     *     link: string,
-     *     description: string,
-     *     content: string,
-     *     author: string,
-     *     date: \DateTimeImmutable,
-     * }>
-     */
-    private function loadBlogPosts(): array
-    {
-        $response = $this->httpClient->request('GET', 'https://feeds.feedburner.com/symfony/blog');
-
-        $posts = [];
-        $crawler = new Crawler($response->getContent());
-        $crawler->filter('item')->each(function (Crawler $node) use (&$posts) {
-            $title = $node->filter('title')->text();
-            $posts[] = [
-                'id' => Uuid::v5(Uuid::fromString('6ba7b810-9dad-11d1-80b4-00c04fd430c8'), $title),
-                'title' => $title,
-                'link' => $node->filter('link')->text(),
-                'description' => $node->filter('description')->text(),
-                'content' => (new Crawler($node->filter('content\:encoded')->text()))->text(),
-                'author' => $node->filter('dc\:creator')->text(),
-                'date' => new \DateTimeImmutable($node->filter('pubDate')->text()),
-            ];
-        });
-
-        return $posts;
-    }
-
-    /**
-     * @param Post[] $posts
-     *
-     * @return Vector[]
-     */
-    private function createEmbeddings(array $posts): array
-    {
-        $texts = [];
-        foreach ($posts as $post) {
-            $texts[] = <<<TEXT
-                Title: {$post['title']}
-                From: {$post['author']} on {$post['date']->format('Y-m-d')}
-                Description: {$post['description']}
-                {$post['content']}
-                TEXT;
-        }
-
-        $response = $this->platform->request(new Embeddings(), $texts);
-
-        assert($response instanceof AsyncResponse);
-        $response = $response->unwrap();
-        assert($response instanceof VectorResponse);
-
-        return $response->getContent();
-    }
-
-    /**
-     * @param Post[]   $posts
-     * @param Vector[] $vectors
-     */
-    private function pushToChromaDB(array $posts, array $vectors): void
-    {
-        $collection = $this->chromaClient->getOrCreateCollection('symfony_blog');
-
-        $ids = array_column($posts, 'id');
-        $vectors = array_map(fn (Vector $vector) => $vector->getData(), $vectors);
-
-        $collection->upsert($ids, $vectors, $posts);
-    }
 }

tests/Blog/EmbedderTest.php

@@ -0,0 +1,101 @@
+<?php
+
+declare(strict_types=1);
+
+namespace App\Tests\Blog;
+
+use App\Blog\Embedder;
+use App\Blog\Loader;
+use App\Blog\Post;
+use Codewithkyrian\ChromaDB\Client;
+use Codewithkyrian\ChromaDB\Resources\CollectionResource;
+use PhpLlm\LlmChain\Document\Vector;
+use PhpLlm\LlmChain\Model\Model;
+use PhpLlm\LlmChain\Model\Response\AsyncResponse;
+use PhpLlm\LlmChain\Model\Response\ResponseInterface as LlmResponse;
+use PhpLlm\LlmChain\Model\Response\VectorResponse;
+use PhpLlm\LlmChain\Platform\ResponseConverter;
+use PhpLlm\LlmChain\PlatformInterface;
+use PHPUnit\Framework\Attributes\CoversClass;
+use PHPUnit\Framework\Attributes\UsesClass;
+use PHPUnit\Framework\TestCase;
+use Symfony\Component\HttpClient\MockHttpClient;
+use Symfony\Component\HttpClient\Response\MockResponse;
+use Symfony\Contracts\HttpClient\ResponseInterface as HttpResponse;
+
+#[CoversClass(Embedder::class)]
+#[UsesClass(Loader::class)]
+#[UsesClass(Post::class)]
+final class EmbedderTest extends TestCase
+{
+    public function testEmbedBlog(): void
+    {
+        $response = MockResponse::fromFile(__DIR__.'/fixtures/blog.rss');
+        $client = new MockHttpClient([$response, $response]);
+        $loader = new Loader($client);
+        $platform = $this->createMock(PlatformInterface::class);
+        $chromaClient = $this->createMock(Client::class);
+        $posts = $loader->load();
+        $vectors = [
+            new Vector([0.1, 0.2, 0.3]),
+            new Vector([0.4, 0.5, 0.6]),
+            new Vector([0.7, 0.8, 0.9]),
+            new Vector([1.0, 1.1, 1.2]),
+            new Vector([1.3, 1.4, 1.5]),
+            new Vector([1.6, 1.7, 1.8]),
+            new Vector([1.9, 2.0, 2.1]),
+            new Vector([2.2, 2.3, 2.4]),
+            new Vector([2.5, 2.6, 2.7]),
+            new Vector([2.8, 2.9, 3.0]),
+        ];
+        $platform
+            ->method('request')
+            ->willReturn($this->createAsyncResponse($vectors));
+
+        $collection = $this->createMock(CollectionResource::class);
+        $chromaClient
+            ->expects($this->once())
+            ->method('getOrCreateCollection')
+            ->with('symfony_blog')
+            ->willReturn($collection);
+
+        $collection
+            ->expects($this->once())
+            ->method('upsert')
+            ->with(
+                array_map(fn (Post $post) => $post->id, $posts),
+                array_map(fn (Vector $vector) => $vector->getData(), $vectors),
+                $posts,
+            );
+
+        $embedder = new Embedder($loader, $platform, $chromaClient);
+        $embedder->embedBlog();
+    }
+
+    /**
+     * @param Vector[] $vectors
+     */
+    private function createAsyncResponse(array $vectors): AsyncResponse
+    {
+        $converter = new class($vectors) implements ResponseConverter {
+            /**
+             * @param Vector[] $vectors
+             */
+            public function __construct(private readonly array $vectors)
+            {
+            }
+
+            public function supports(Model $model, object|array|string $input): bool
+            {
+                return true;
+            }
+
+            public function convert(HttpResponse $response, array $options = []): LlmResponse
+            {
+                return new VectorResponse(...$this->vectors);
+            }
+        };
+
+        return new AsyncResponse($converter, new MockResponse());
+    }
+}