[Store] Add Neo4j

Author: Guikingone

examples/.env

@@ -90,3 +90,9 @@ QDRANT_SERVICE_API_KEY=changeMe
 SURREALDB_HOST=http://127.0.0.1:8000
 SURREALDB_USER=symfony
 SURREALDB_PASS=symfony
+
+# Neo4J
+NEO4J_HOST=http://127.0.0.1:7474
+NEO4J_DATABASE=neo4j
+NEO4J_USERNAME=neo4j
+NEO4J_PASSWORD=symfonyai

examples/compose.yaml

@@ -37,3 +37,11 @@ services:
           SURREAL_HTTP_MAX_KEY_BODY_SIZE: 49152
       ports:
           - '8000:8000'
+
+  neo4j:
+      image: neo4j
+      environment:
+          NEO4J_AUTH: 'neo4j/${NEO4J_PASSWORD:-symfonyai}'
+      ports:
+          - '7474:7474'
+          - '7687:7687'

examples/rag/neo4j.php

@@ -0,0 +1,75 @@
+<?php
+
+/*
+ * This file is part of the Symfony package.
+ *
+ * (c) Fabien Potencier <[email protected]>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+use Symfony\AI\Agent\Agent;
+use Symfony\AI\Agent\Toolbox\AgentProcessor;
+use Symfony\AI\Agent\Toolbox\Tool\SimilaritySearch;
+use Symfony\AI\Agent\Toolbox\Toolbox;
+use Symfony\AI\Fixtures\Movies;
+use Symfony\AI\Platform\Bridge\OpenAI\Embeddings;
+use Symfony\AI\Platform\Bridge\OpenAI\GPT;
+use Symfony\AI\Platform\Bridge\OpenAI\PlatformFactory;
+use Symfony\AI\Platform\Message\Message;
+use Symfony\AI\Platform\Message\MessageBag;
+use Symfony\AI\Store\Bridge\Neo4j\Store;
+use Symfony\AI\Store\Document\Metadata;
+use Symfony\AI\Store\Document\TextDocument;
+use Symfony\AI\Store\Document\Vectorizer;
+use Symfony\AI\Store\Indexer;
+use Symfony\Component\HttpClient\HttpClient;
+use Symfony\Component\Uid\Uuid;
+
+require_once dirname(__DIR__).'/bootstrap.php';
+
+// initialize the store
+$store = new Store(
+    httpClient: HttpClient::create(),
+    endpointUrl: env('NEO4J_HOST'),
+    username: env('NEO4J_USERNAME'),
+    password: env('NEO4J_PASSWORD'),
+    databaseName: env('NEO4J_DATABASE'),
+    vectorIndexName: 'Movies',
+    nodeName: 'movies',
+);
+
+// initialize the table
+$store->initialize();
+
+// create embeddings and documents
+$documents = [];
+foreach (Movies::all() as $i => $movie) {
+    $documents[] = new TextDocument(
+        id: Uuid::v4(),
+        content: 'Title: '.$movie['title'].\PHP_EOL.'Director: '.$movie['director'].\PHP_EOL.'Description: '.$movie['description'],
+        metadata: new Metadata($movie),
+    );
+}
+
+// create embeddings for documents
+$platform = PlatformFactory::create($_SERVER['OPENAI_API_KEY']);
+$vectorizer = new Vectorizer($platform, $embeddings = new Embeddings());
+$indexer = new Indexer($vectorizer, $store);
+$indexer->index($documents);
+
+$model = new GPT(GPT::GPT_4O_MINI);
+
+$similaritySearch = new SimilaritySearch($platform, $embeddings, $store);
+$toolbox = new Toolbox([$similaritySearch], logger: logger());
+$processor = new AgentProcessor($toolbox);
+$agent = new Agent($platform, $model, [$processor], [$processor]);
+
+$messages = new MessageBag(
+    Message::forSystem('Please answer all user questions only using SimilaritySearch function.'),
+    Message::ofUser('Which movie fits the theme of technology?')
+);
+$response = $agent->call($messages);
+
+echo $response->getContent().\PHP_EOL;

src/store/CHANGELOG.md

@@ -44,6 +44,7 @@ CHANGELOG
    - Pinecone
    - Qdrant
    - SurrealDB
+   - Neo4j
  * Add Retrieval Augmented Generation (RAG) support:
    - Document embedding storage
    - Similarity search for relevant documents

src/store/composer.json

@@ -6,7 +6,13 @@
         "ai",
         "mongodb",
         "pinecone",
-        "chromadb"
+        "chromadb",
+        "mariadb",
+        "postgres",
+        "meilisearch",
+        "surrealdb",
+        "qdrant",
+        "neo4j"
     ],
     "license": "MIT",
     "authors": [

src/store/doc/index.rst

@@ -44,6 +44,7 @@ You can find more advanced usage in combination with an Agent using the store fo
 * `Similarity Search with Pinecone (RAG)`_
 * `Similarity Search with Qdrant (RAG)`_
 * `Similarity Search with SurrealDB (RAG)`_
+* `Similarity Search with Neo4j (RAG)`_
 
 Supported Stores
 ----------------
@@ -58,6 +59,7 @@ Supported Stores
 * `Postgres`_ (requires `ext-pdo`)
 * `Qdrant`_
 * `SurrealDB`_
+* `Neo4j`_
 
 .. note::
 
@@ -98,6 +100,7 @@ This leads to a store implementing two methods::
 .. _`Similarity Search with SurrealDB (RAG)`: https://github.com/symfony/ai/blob/main/examples/rag/surrealdb.php
 .. _`Similarity Search with memory storage (RAG)`: https://github.com/symfony/ai/blob/main/examples/rag/in-memory.php
 .. _`Similarity Search with Qdrant (RAG)`: https://github.com/symfony/ai/blob/main/examples/rag/qdrant.php
+.. _`Similarity Search with Neo4j (RAG)`: https://github.com/symfony/ai/blob/main/examples/rag/neo4j.php
 .. _`Azure AI Search`: https://azure.microsoft.com/products/ai-services/ai-search
 .. _`Chroma`: https://www.trychroma.com/
 .. _`MariaDB`: https://mariadb.org/projects/mariadb-vector/
@@ -108,4 +111,5 @@ This leads to a store implementing two methods::
 .. _`SurrealDB`: https://surrealdb.com/
 .. _`InMemory`: https://www.php.net/manual/en/language.types.array.php
 .. _`Qdrant`: https://qdrant.tech/
+.. _`Neo4j`: https://neo4j.com/
 .. _`GitHub`: https://github.com/symfony/ai/issues/16

src/store/src/Bridge/Neo4j/Store.php

@@ -0,0 +1,117 @@
+<?php
+
+/*
+ * This file is part of the Symfony package.
+ *
+ * (c) Fabien Potencier <[email protected]>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace Symfony\AI\Store\Bridge\Neo4j;
+
+use Symfony\AI\Platform\Vector\NullVector;
+use Symfony\AI\Platform\Vector\Vector;
+use Symfony\AI\Store\Document\Metadata;
+use Symfony\AI\Store\Document\VectorDocument;
+use Symfony\AI\Store\Exception\InvalidArgumentException;
+use Symfony\AI\Store\InitializableStoreInterface;
+use Symfony\AI\Store\VectorStoreInterface;
+use Symfony\Component\Uid\Uuid;
+use Symfony\Contracts\HttpClient\HttpClientInterface;
+
+/**
+ * @author Guillaume Loulier <[email protected]>
+ */
+final readonly class Store implements InitializableStoreInterface, VectorStoreInterface
+{
+    public function __construct(
+        private HttpClientInterface $httpClient,
+        private string $endpointUrl,
+        #[\SensitiveParameter] private string $username,
+        #[\SensitiveParameter] private string $password,
+        #[\SensitiveParameter] private string $databaseName,
+        private string $vectorIndexName,
+        private string $nodeName,
+        private string $embeddingsField = 'embeddings',
+        private int $embeddingsDimension = 1536,
+        private string $embeddingsDistance = 'cosine',
+        private bool $quantization = false,
+    ) {
+    }
+
+    public function add(VectorDocument ...$documents): void
+    {
+        foreach ($documents as $document) {
+            $this->request('POST', \sprintf('db/%s/query/v2', $this->databaseName), [
+                'statement' => \sprintf('CREATE (n:%s {id: $id, metadata: $metadata, %s: $embeddings}) RETURN n', $this->nodeName, $this->embeddingsField),
+                'parameters' => [
+                    'id' => $document->id->toRfc4122(),
+                    'metadata' => json_encode($document->metadata->getArrayCopy()),
+                    'embeddings' => $document->vector->getData(),
+                ],
+            ]);
+        }
+    }
+
+    public function query(Vector $vector, array $options = [], ?float $minScore = null): array
+    {
+        $response = $this->request('POST', \sprintf('db/%s/query/v2', $this->databaseName), [
+            'statement' => \sprintf('CALL db.index.vector.queryNodes("%s", 5, $vectors) YIELD node, score RETURN node, score', $this->vectorIndexName),
+            'parameters' => [
+                'vectors' => $vector->getData(),
+            ],
+        ]);
+
+        return array_map($this->convertToVectorDocument(...), $response['data']['values']);
+    }
+
+    public function initialize(array $options = []): void
+    {
+        $this->request('POST', \sprintf('db/%s/query/v2', $this->databaseName), [
+            'statement' => \sprintf(
+                'CREATE VECTOR INDEX %s IF NOT EXISTS FOR (n:%s) ON n.%s OPTIONS { indexConfig: {`vector.dimensions`: %d, `vector.similarity_function`: "%s", `vector.quantization.enabled`: %s}}',
+                $this->vectorIndexName, $this->nodeName, $this->embeddingsField, $this->embeddingsDimension, $this->embeddingsDistance, $this->quantization ? 'true' : 'false',
+            ),
+        ]);
+    }
+
+    /**
+     * @param array<string, mixed> $payload
+     *
+     * @return array<string, mixed>
+     */
+    private function request(string $method, string $endpoint, array $payload = []): array
+    {
+        $url = \sprintf('%s/%s', $this->endpointUrl, $endpoint);
+
+        $response = $this->httpClient->request($method, $url, [
+            'auth_basic' => \sprintf('%s:%s', $this->username, $this->password),
+            'json' => $payload,
+        ]);
+
+        return $response->toArray();
+    }
+
+    /**
+     * @param array<string|int, mixed> $data
+     */
+    private function convertToVectorDocument(array $data): VectorDocument
+    {
+        $payload = $data[0];
+
+        $id = $payload['properties']['id'] ?? throw new InvalidArgumentException('Missing "id" field in the document data');
+
+        $vector = !\array_key_exists($this->embeddingsField, $payload['properties']) || null === $payload['properties'][$this->embeddingsField]
+            ? new NullVector()
+            : new Vector($payload['properties'][$this->embeddingsField]);
+
+        return new VectorDocument(
+            id: Uuid::fromString($id),
+            vector: $vector,
+            metadata: new Metadata(json_decode($payload['properties']['metadata'], true)),
+            score: $data[1] ?? null
+        );
+    }
+}