Rekognition Php Laravel Package

Getting Started

Minimal Setup

1. Installation

   composer require clickandmortar/rekognition-php
   

   Ensure your AWS credentials are configured via:

   • Environment variables (AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, AWS_REGION)
   • Or via ~/.aws/credentials file.

2. First Use Case: Detect Labels in an Image

   use ClickAndMortar\Rekognition\Rekognition;
   
   $rekognition = new Rekognition();
   $result = $rekognition->detectLabels('path/to/image.jpg');
   print_r($result);
   

   • Key Output: Returns an array of detected labels with confidence scores.

3. Where to Look First

   • AWS Rekognition Documentation (for feature reference).
   • src/ClickAndMortar/Rekognition/Rekognition.php (core class methods).
   • tests/ directory (for usage examples and edge cases).

Implementation Patterns

Common Workflows

1. Image Analysis Pipeline

   $rekognition = new Rekognition();
   $labels = $rekognition->detectLabels($imagePath);
   $faces = $rekognition->detectFaces($imagePath);
   $text = $rekognition->detectText($imagePath);
   
   // Filter labels by confidence threshold
   $filteredLabels = array_filter($labels, fn($label) => $label['Confidence'] > 80);
   

2. Batch Processing

   $rekognition = new Rekognition();
   $images = ['image1.jpg', 'image2.jpg', 'image3.jpg'];
   $results = [];
   
   foreach ($images as $image) {
       $results[$image] = $rekognition->detectLabels($image);
   }
   

3. Integration with Laravel

   • Store AWS responses in a database:

     $rekognition = new Rekognition();
     $labels = $rekognition->detectLabels($request->file('image')->path());
     
     Analysis::create([
         'user_id' => auth()->id(),
         'labels' => json_encode($labels),
     ]);
     

   • Use middleware to validate AWS responses:

     public function handle($request, Closure $next) {
         $response = $rekognition->detectLabels($request->image);
         if (empty($response)) {
             abort(400, 'No labels detected.');
         }
         return $next($request);
     }
     

4. Async Processing with Queues

   // Dispatch a job
   DetectLabelsJob::dispatch($imagePath, $userId);
   
   // Job class
   public function handle() {
       $rekognition = new Rekognition();
       $labels = $rekognition->detectLabels(storage_path($this->imagePath));
       Analysis::create([...]);
   }
   

Gotchas and Tips

Pitfalls

1. AWS Credentials

   • If using IAM roles (e.g., in EC2), ensure the role has rekognition:Detect* permissions.
   • Error: AccessDeniedException → Verify IAM policies and credentials.

2. Image Format/Size

   • Rekognition supports JPEG/PNG (max 10MB for DetectLabels, 5MB for DetectText).
   • Error: InvalidS3ObjectException → Check file size/type before upload.

3. Rate Limits

   • AWS Rekognition has service quotas.
   • Tip: Implement exponential backoff for throttling (ThrottlingException).

4. Deprecated Methods

   • The package is unmaintained (last release: 2019). Some AWS API changes may break compatibility.
   • Workaround: Extend the class or use the official AWS SDK as a fallback.

Debugging Tips

1. Enable AWS SDK Debugging

   use Aws\Common\Aws;
   Aws::setDebug(true); // Logs requests/responses to `storage/logs/aws.log`
   

2. Validate Responses

   $result = $rekognition->detectLabels($imagePath);
   if (!isset($result['Labels'])) {
       throw new \RuntimeException('Invalid Rekognition response');
   }
   

3. Handle Exceptions

   try {
       $result = $rekognition->detectLabels($imagePath);
   } catch (\Aws\Rekognition\Exception\RekognitionException $e) {
       Log::error('Rekognition Error: ' . $e->getMessage());
       abort(500, 'Image analysis failed.');
   }
   

Extension Points

1. Custom Response Mapping

   $rekognition = new Rekognition();
   $rawResult = $rekognition->detectLabels($imagePath);
   
   // Transform raw AWS response
   $formatted = array_map(function($label) {
       return [
           'name' => $label['Name'],
           'confidence' => $label['Confidence'] / 100, // Convert to decimal
           'parents' => $label['Parents'] ?? [],
       ];
   }, $rawResult['Labels']);
   

2. Add New Features

   • Extend the Rekognition class to support unsupported APIs (e.g., CompareFaces):

     class ExtendedRekognition extends Rekognition {
         public function compareFaces($sourceImage, $targetImage) {
             return $this->client->compareFaces([
                 'SourceImage' => ['Bytes' => file_get_contents($sourceImage)],
                 'TargetImage' => ['Bytes' => file_get_contents($targetImage)],
             ]);
         }
     }
     

3. Mocking for Tests

   $rekognition = Mockery::mock(ClickAndMortar\Rekognition\Rekognition::class);
   $rekognition->shouldReceive('detectLabels')
               ->once()
               ->andReturn(['Labels' => [['Name' => 'Car', 'Confidence' => 99]]]);