Ai Mongo Db Message Store Laravel Package

Getting Started

Minimal Setup

1. Install Dependencies:

   composer require symfony/ai-chat symfony/ai-mongo-db-message-store mongodb/mongodb
   

2. Configure MongoDB Connection: Add to config/services.yaml (Symfony) or config/mongodb.php (Laravel):

   # config/packages/ai.yaml (Symfony)
   framework:
       ai:
           message_store: mongodb
           mongodb:
               uri: 'mongodb://user:pass@host:port/db'
               collection: 'ai_chat_messages'
   

3. Initialize Collection (run once):

   use MongoDB\Client;
   $client = new Client('mongodb://user:pass@host:port/db');
   $client->selectDatabase('db')->createCollection('ai_chat_messages');
   

4. First Use Case:

   use Symfony\Component\AI\Chat\MessageStore\MongoDbMessageStore;
   use Symfony\Component\AI\Chat\Message;
   
   $store = new MongoDbMessageStore($client->selectDatabase('db'));
   $store->save(new Message('user', 'Hello!'));
   $messages = $store->findByConversationId('conv_123');
   

Where to Look First

• Symfony AI Chat Docs: symfony/ai-chat for message store interfaces.
• MongoDB PHP Library: Indexing Guide for performance tuning.
• Laravel-Symfony Bridge: Check symfony/dependency-injection for container integration.

Implementation Patterns

Core Workflows

1. Chat Session Management:

   // Laravel Service Example
   public function handleChatRequest(string $conversationId, string $userInput): array
   {
       $store = app(MongoDbMessageStore::class);
       $messages = $store->findByConversationId($conversationId);
   
       $response = $this->callSymfonyAIChat($userInput, $messages);
       $store->save(new Message('ai', $response));
   
       return ['response' => $response, 'history' => $messages];
   }
   

2. Conversation Pruning:

   // TTL Index Setup (via MongoDB CLI or PHP)
   $collection->createIndex(['timestamp' => 1], ['expireAfterSeconds' => 2592000]); // 30 days
   

3. Hybrid Laravel-Symfony Integration:
   • Option A: API Layer

     // Laravel Route
     Route::post('/chat', function (Request $request) {
         $response = Http::post('http://symfony-ai-service/api/chat', $request->all());
         return $response->json();
     });
     

   • Option B: Shared Container

     // Laravel Service Provider
     public function register()
     {
         $this->app->singleton(MongoDbMessageStore::class, function ($app) {
             $client = new Client(config('mongodb.connection'));
             return new MongoDbMessageStore($client->selectDatabase(config('mongodb.database')));
         });
     }
     

Integration Tips

• Laravel Event Listeners: Bind to Symfony\Component\AI\Chat\Event\MessageSentEvent for post-processing:

  public function handle(MessageSentEvent $event)
  {
      // Log, analyze, or sync to CRM
  }
  

• Query Optimization: Use MongoDB’s $text search for full-text queries:

  $messages = $store->findBy([
      'conversationId' => 'conv_123',
      'text' => ['$text' => ['$search' => 'urgent']]
  ]);
  

• Fallback Mechanism: Implement a decorator pattern for resilience:

  class ResilientMongoDbMessageStore implements MessageStoreInterface
  {
      public function __construct(
          private MongoDbMessageStore $primary,
          private ?MessageStoreInterface $fallback = null
      ) {}
  
      public function save(Message $message)
      {
          try {
              $this->primary->save($message);
          } catch (\Exception $e) {
              $this->fallback?->save($message);
          }
      }
  }
  

Gotchas and Tips

Pitfalls

1. Driver Conflicts:

   • Issue: mongodb/mongodb (v2.0+) vs. jenssegers/mongodb (Laravel’s legacy driver).
   • Fix: Standardize on mongodb/mongodb and update Laravel’s MongoDB config:

     // config/mongodb.php
     'connections' => [
         'default' => [
             'driver' => 'mongodb',
             'host' => env('DB_HOST', '127.0.0.1'),
             'port' => env('DB_PORT', 27017),
             'database' => env('DB_DATABASE', 'forge'),
             'username' => env('DB_USERNAME', 'forge'),
             'password' => env('DB_PASSWORD', ''),
             'options' => [
                 'typeMap' => ['root' => 'array', 'document' => 'array'], // Critical for Symfony serialization
             ],
         ],
     ],
     

2. Serialization Mismatches:

   • Issue: Symfony’s serializer may conflict with Laravel’s JSON handling.
   • Fix: Configure Symfony’s serializer to avoid overlaps:

     # config/packages/serializer.yaml
     framework:
         serializer:
             mapping:
                 paths: ['%kernel.project_dir%/config/serializer']
             ignore_annotations: true
     

3. Missing Indexes:

   • Issue: Slow queries on conversationId or timestamp.
   • Fix: Pre-create indexes:

     $collection->createIndex(['conversationId' => 1, 'timestamp' => -1]);
     

4. No Schema Validation:

   • Issue: MongoDB accepts any document structure, risking inconsistent data.
   • Fix: Use a validator or schema library (e.g., spatie/laravel-mongodb-schema):

     use Spatie\LaravelMongoDbSchema\Schema;
     Schema::create('ai_chat_messages', function (Blueprint $collection) {
         $collection->index('conversationId');
         $collection->index('timestamp');
         $collection->required('userId');
         $collection->required('role'); // 'user' or 'ai'
         $collection->required('content');
     });
     

5. Symfony-Specific Assumptions:

   • Issue: Assumes Symfony’s Message class; Laravel may need adapters.
   • Fix: Create a Laravel-compatible message class:

     class LaravelMessage implements \Symfony\Component\AI\Chat\MessageInterface
     {
         public function __construct(
             public string $role,
             public string $content,
             public ?string $id = null,
             public ?string $conversationId = null,
             public ?\DateTimeInterface $timestamp = null
         ) {}
     }
     

Debugging Tips

• Enable MongoDB Logging:

  $client->selectDatabase('db')->manage()->getLogger()->setCallback(function ($level, $message) {
      error_log("[MongoDB] [$level] $message");
  });
  

• Query Profiling: Use MongoDB’s explain() to analyze slow queries:

  $explanation = $collection->find(['conversationId' => 'conv_123'])->explain();
  

• Common Errors:
  • InvalidArgumentException: Check typeMap in MongoDB options (must match Symfony’s expectations).
  • ClassNotFoundException: Ensure Symfony\Component\AI\Chat\Message is autoloaded or use the Laravel adapter.

Extension Points

1. Custom Message Fields: Extend the message store to include metadata:

   class ExtendedMongoDbMessageStore extends MongoDbMessageStore
   {
       public function saveWithMetadata(Message $message, array $metadata)
       {
           $document = $this->convertMessageToDocument($message);
           $document['metadata'] = $metadata;
           $this->collection->insertOne($document);
       }
   }
   

2. Event Dispatching: Trigger Laravel events on message operations:

   class EventfulMongoDbMessageStore extends MongoDbMessageStore
   {
       public function save(Message $message)
       {
           $result = parent::save($message);
           event(new MessageStored($message, $result->getInsertedId()));
           return $result;
       }
   }
   

3. Bulk Operations: Optimize batch inserts for high-throughput scenarios:

   public function bulkSave(iterable $messages)
   {
       $documents = array_map([$this, 'convertMessageToDocument'], $messages);
       $this->collection->insertMany($documents);
   }
   

4