Qdrant PHP Client

To connect to a Qdrant instance, initialize a Config object with the host, optionally set an API key, build a transport layer using the Builder, and instantiate the Qdrant client.

use Qdrant\Qdrant;
use Qdrant\Config;
use Qdrant\Http\Builder;

$config = new Config(QDRANT_HOST);
$config->setApiKey(QDRANT_API_KEY);

$transport = (new Builder())->build($config);
$client = new Qdrant($transport);

Install the client in your PHP project using Composer.

composer require hkulekci/qdrant

The Points endpoint provides methods to interact with individual points or groups of points within a specific Qdrant collection. You can use it to upsert, delete, retrieve, count, and search for points.

Key operations include:

• Upserting: Use upsert() to insert or update points.
• Batching: Use batch() for high-performance batch operations.
• Retrieving: Use id() to get a single point or ids() to get multiple points by their IDs.
• Deleting: Use delete() for specific IDs or deleteByFilter() to remove points matching a Filter.
• Counting: Use count() to get the number of points, optionally filtered.

To perform a semantic search, first generate an embedding for your query string using an OpenAI client, then pass that embedding into a SearchRequest via a VectorStruct.

$openai = OpenAI::client(OPENAI_API_KEY);

$query = 'lorem ipsum dolor sit amed';
$response = $openai->embeddings()->create([
    'model' => 'text-embedding-ada-002',
    'input' => $query,
]);
$embedding = array_values($response->embeddings[0]->embedding);

$searchRequest = (new SearchRequest(new VectorStruct($embedding, 'content')))
    ->setLimit(10)
    ->setParams([
        'hnsw_ef' => 128,
        'exact' => false,
    ])
    ->setWithPayload(true);

$response = $client->collections('contents')->points()->search($searchRequest);

foreach ($response['result'] as $item) {
    echo $item['score'] . ';' . $item['payload']['id'] . ';' . $item['payload']['meta_data'] . PHP_EOL;
}

When performing an upsert operation, you can pass an options array to ensure the operation is acknowledged before the request returns. Use ['wait' => 'true'] to wait for the data to be processed.

$client->collections('contents')->points()->upsert($points, ['wait' => 'true']);

To insert data, wrap your points in a PointsStruct. Each point is represented by a PointStruct containing an integer ID, a VectorStruct (mapping the embedding to a specific vector name), and an optional payload array. Use the upsert() method on the collection's points endpoint to save them.

use Qdrant\Models\PointsStruct;
use Qdrant\Models\PointStruct;
use Qdrant\Models\VectorStruct;

// Assuming $embedding is an array of floats from an embedding model
$points = new PointsStruct();
$points->addPoint(
    new PointStruct(
        (int) $imageId,
        new VectorStruct($embedding, 'content'),
        [
            'id' => 1,
            'meta' => 'Meta data'
        ]
    )
);

$client->collections('contents')->points()->upsert($points);

Use the collections() method on the client to create a new collection. You must define VectorParams specifying the vector size and the distance metric (e.g., VectorParams::DISTANCE_COSINE).

use Qdrant\Endpoints\Collections;
use Qdrant\Models\Request\CreateCollection;
use Qdrant\Models\Request\VectorParams;

$createCollection = new CreateCollection();
$createCollection->addVector(new VectorParams(1536, VectorParams::DISTANCE_COSINE), 'content');
$response = $client->collections('contents')->create($createCollection);

You can remove points from a collection in two ways:

1. By specific IDs: Use delete(array $points) passing an array of IDs.
2. By Filter: Use deleteByFilter(Filter $filter) to remove all points that match a specific criteria.

// Delete by IDs
$collection->points()->delete([1, 2, 3]);

// Delete by Filter
$filter = new Filter([...]);
$collection->points()->deleteByFilter($filter);

Create a new collection by passing a CreateCollection model instance to the create() method. You can optionally provide query parameters as an array.

use Qdrant\Models\Request\CreateCollection;

$params = new CreateCollection([
    // collection configuration parameters
]);

$response = $collections->create($params);

Use the ids() method to retrieve multiple points at once. You can control whether to include the payload or the vector in the response.

Parameters:

• array $ids: An array of point IDs.
• bool $withPayload: Whether to include the payload (default false).
• bool $withVector: Whether to include the vector (default true).

$ids = [1, "uuid-string", 3];
$response = $collection->points()->ids($ids, true, true);