Symfony AI - Store Component

Installation

1

$ composer require symfony/ai-store

Purpose

A typical use-case in agentic applications is a dynamic context-extension with similar and useful information, for so called Retrieval Augmented Generation (RAG). The Store component implements low-level interfaces, that can be implemented by different concrete and vendor-specific implementations, so called bridges. On top of those bridges, the Store component provides higher level features to populate and query those stores with and for documents.

Indexing

One higher level feature is the Indexer. The purpose of this service is to populate a store with documents. Therefore it accepts one or multiple TextDocument objects, converts them into embeddings and stores them in the used vector store:

1
2
3
4
5
6

use Symfony\AI\Store\Document\TextDocument;
use Symfony\AI\Store\Indexer;

$indexer = new Indexer($platform, $model, $store);
$document = new TextDocument('This is a sample document.');
$indexer->index($document);

You can find more advanced usage in combination with an Agent using the store for RAG in the examples folder.

Retrieving

The opposite of indexing is retrieving. The Retriever is a higher level feature that allows you to search for documents in a store based on a query string. It vectorizes the query and retrieves similar documents from the store:

1
2
3
4
5
6
7
8

use Symfony\AI\Store\Retriever;

$retriever = new Retriever($vectorizer, $store);
$documents = $retriever->retrieve('What is the capital of France?');

foreach ($documents as $document) {
    echo $document->metadata->get('source');
}

The retriever accepts optional parameters to customize the retrieval:

• $options: An array of options to pass to the underlying store query (e.g., limit, filters)

Supported Stores

• Azure AI Search
• Chroma (requires codewithkyrian/chromadb-php as additional dependency)
• Cloudflare
• InMemory
• Manticore Search
• MariaDB (requires ext-pdo)
• Meilisearch
• Milvus
• MongoDB Atlas (requires mongodb/mongodb as additional dependency)
• Neo4j
• OpenSearch
• Pinecone (requires probots-io/pinecone-php as additional dependency)
• Postgres (requires ext-pdo)
• Qdrant
• Supabase (requires manual database setup)
• SurrealDB
• Symfony Cache (requires symfony/cache as additional dependency)
• Typesense
• Weaviate

Commands

While using the Store component in your Symfony application along with the AiBundle, you can use the bin/console ai:store:setup command to initialize the store and bin/console ai:store:drop to clean up the store:

1
2
3
4
5
6
7
8

# config/packages/ai.yaml
ai:
    # ...

    store:
        chromadb:
            symfonycon:
                collection: 'symfony_blog'

1
2

$ php bin/console ai:store:setup symfonycon
$ php bin/console ai:store:drop symfonycon

Implementing a Bridge

The main extension points of the Store component is the StoreInterface, that defines the methods for adding vectorized documents to the store, and querying the store for documents with a vector.

This leads to a store implementing two methods:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

use Symfony\AI\Store\StoreInterface;
use Symfony\AI\Store\Vector;
use Symfony\AI\Store\VectorDocument;

class MyStore implements StoreInterface
{
    public function add(VectorDocument ...$documents): void
    {
        // Implementation to add a document to the store
    }

    public function query(Vector $vector, array $options = []): array
    {
        // Implementation to query the store for documents
        return $documents;
    }
}