[Store] Add logging to Vectorizer

Author: OskarStark

src/ai-bundle/src/AiBundle.php

@@ -1057,6 +1057,7 @@ private function processVectorizerConfig(string $name, array $config, ContainerB
         $vectorizerDefinition = new Definition(Vectorizer::class, [
             new Reference($config['platform']),
             new Reference('ai.vectorizer.'.$name.'.model'),
+            new Reference('logger', ContainerInterface::IGNORE_ON_INVALID_REFERENCE),
         ]);
         $vectorizerDefinition->addTag('ai.vectorizer', ['name' => $name]);
         $container->setDefinition('ai.vectorizer.'.$name, $vectorizerDefinition);

src/ai-bundle/tests/DependencyInjection/AiBundleTest.php

@@ -22,6 +22,7 @@
 use Symfony\AI\Store\Document\Vectorizer;
 use Symfony\Component\Config\Definition\Exception\InvalidConfigurationException;
 use Symfony\Component\DependencyInjection\ContainerBuilder;
+use Symfony\Component\DependencyInjection\ContainerInterface;
 use Symfony\Component\DependencyInjection\Definition;
 use Symfony\Component\DependencyInjection\Reference;
 
@@ -622,6 +623,41 @@ public function testVectorizerConfiguration()
         $this->assertTrue($modelDefinition->hasTag('ai.model.embeddings_model'));
     }
 
+    public function testVectorizerWithLoggerInjection()
+    {
+        $container = $this->buildContainer([
+            'ai' => [
+                'vectorizer' => [
+                    'my_vectorizer' => [
+                        'platform' => 'my_platform_service_id',
+                        'model' => [
+                            'class' => 'Symfony\AI\Platform\Bridge\OpenAi\Embeddings',
+                            'name' => 'text-embedding-3-small',
+                        ],
+                    ],
+                ],
+            ],
+        ]);
+
+        $vectorizerDefinition = $container->getDefinition('ai.vectorizer.my_vectorizer');
+        $arguments = $vectorizerDefinition->getArguments();
+
+        $this->assertCount(3, $arguments, 'Vectorizer should have 3 arguments: platform, model, and logger');
+
+        // First argument should be platform reference
+        $this->assertInstanceOf(Reference::class, $arguments[0]);
+        $this->assertSame('my_platform_service_id', (string) $arguments[0]);
+
+        // Second argument should be model reference
+        $this->assertInstanceOf(Reference::class, $arguments[1]);
+        $this->assertSame('ai.vectorizer.my_vectorizer.model', (string) $arguments[1]);
+
+        // Third argument should be logger reference with IGNORE_ON_INVALID_REFERENCE
+        $this->assertInstanceOf(Reference::class, $arguments[2]);
+        $this->assertSame('logger', (string) $arguments[2]);
+        $this->assertSame(ContainerInterface::IGNORE_ON_INVALID_REFERENCE, $arguments[2]->getInvalidBehavior());
+    }
+
     public function testIndexerWithConfiguredVectorizer()
     {
         $container = $this->buildContainer([

src/store/src/Document/Vectorizer.php

@@ -11,6 +11,8 @@
 
 namespace Symfony\AI\Store\Document;
 
+use Psr\Log\LoggerInterface;
+use Psr\Log\NullLogger;
 use Symfony\AI\Platform\Capability;
 use Symfony\AI\Platform\Model;
 use Symfony\AI\Platform\PlatformInterface;
@@ -20,32 +22,46 @@
     public function __construct(
         private PlatformInterface $platform,
         private Model $model,
+        private LoggerInterface $logger = new NullLogger(),
     ) {
     }
 
     public function __invoke(array $documents): array
     {
+        $documentCount = \count($documents);
+        $this->logger->info('Starting vectorization process', ['document_count' => $documentCount]);
+
         if ($this->model->supports(Capability::INPUT_MULTIPLE)) {
+            $this->logger->debug('Using batch vectorization with model that supports multiple inputs');
             $result = $this->platform->invoke($this->model, array_map(fn (TextDocument $document) => $document->content, $documents));
 
             $vectors = $result->asVectors();
+            $this->logger->debug('Batch vectorization completed', ['vector_count' => \count($vectors)]);
         } else {
+            $this->logger->debug('Using sequential vectorization for model without multiple input support');
             $results = [];
-            foreach ($documents as $document) {
+            foreach ($documents as $i => $document) {
+                $this->logger->debug('Vectorizing document', ['document_index' => $i, 'document_id' => $document->id]);
                 $results[] = $this->platform->invoke($this->model, $document->content);
             }
 
             $vectors = [];
             foreach ($results as $result) {
                 $vectors = array_merge($vectors, $result->asVectors());
             }
+            $this->logger->debug('Sequential vectorization completed', ['vector_count' => \count($vectors)]);
         }
 
         $vectorDocuments = [];
         foreach ($documents as $i => $document) {
             $vectorDocuments[] = new VectorDocument($document->id, $vectors[$i], $document->metadata);
         }
 
+        $this->logger->info('Vectorization process completed', [
+            'document_count' => $documentCount,
+            'vector_document_count' => \count($vectorDocuments),
+        ]);
+
         return $vectorDocuments;
     }
 }