PHPORM-13 Feature Queryable Encryption

.github/workflows/continuous-integration.yml

@@ -163,3 +163,4 @@ jobs:
         env:
           DOCTRINE_MONGODB_SERVER: ${{ steps.setup-mongodb.outputs.cluster-uri }}
           USE_LAZY_GHOST_OBJECTS: ${{ matrix.proxy == 'lazy-ghost' && '1' || '0' }}"
+          CRYPT_SHARED_LIB_PATH: ${{ steps.setup-mongodb.outputs.crypt-shared-lib-path }}

composer.json

@@ -30,7 +30,7 @@
         "doctrine/persistence": "^3.2 || ^4",
         "friendsofphp/proxy-manager-lts": "^1.0",
         "jean85/pretty-package-versions": "^1.3.0 || ^2.0.1",
-        "mongodb/mongodb": "^1.21 || ^2.0@dev",
+        "mongodb/mongodb": "^1.21.2 || ^2.1.1",
         "psr/cache": "^1.0 || ^2.0 || ^3.0",
         "symfony/console": "^5.4 || ^6.0 || ^7.0",
         "symfony/deprecation-contracts": "^2.2 || ^3.0",

docs/en/cookbook/queryable-encryption.rst

@@ -0,0 +1,270 @@
+Queryable Encryption
+====================
+
+This cookbook provides a tutorial on setting up and using Queryable Encryption
+(QE) with Doctrine MongoDB ODM to protect sensitive data in your documents.
+
+Introduction
+------------
+
+In many applications, you need to store sensitive information like social
+security numbers, financial data, or personal details. MongoDB's Queryable
+Encryption allows you to encrypt this data on the client-side, store it as
+fully randomized encrypted data, and still run expressive queries on it. This
+ensures that sensitive data is never exposed in an unencrypted state on the
+server, in system logs, or in backups.
+
+This tutorial will guide you through the process of securing a document's
+fields using queryable encryption, from defining the document and configuring
+the connection to storing and querying the encrypted data.
+
+.. note::
+
+    Queryable Encryption is only available on MongoDB Enterprise 7.0+ or
+    MongoDB Atlas.
+
+The Scenario
+------------
+
+We will model a ``Patient`` document that has an embedded ``PatientRecord``.
+This record contains sensitive information:
+
+- A Social Security Number (``ssn``), which we need to query for exact
+  matches.
+- A ``billingAmount``, which should support range queries.
+- A ``billing`` object, which should be encrypted but not directly queryable.
+
+Defining the Documents
+----------------------
+
+First, let's define our ``Patient``, ``PatientRecord``, and ``Billing``
+classes. We use the :ref:`#[Encrypt] <encrypt_attribute>` attribute to mark
+fields that require encryption.
+
+.. code-block:: php
+
+    <?php
+
+    namespace Documents;
+
+    use Doctrine\ODM\MongoDB\Mapping\Annotations as ODM;
+    use Doctrine\ODM\MongoDB\Mapping\Annotations\Encrypt;
+    use Doctrine\ODM\MongoDB\Mapping\Annotations\EncryptQuery;
+
+    #[ODM\Document]
+    class Patient
+    {
+        #[ODM\Id]
+        public string $id;
+
+        #[ODM\EmbedOne(targetDocument: PatientRecord::class)]
+        public PatientRecord $patientRecord;
+    }
+
+    #[ODM\EmbeddedDocument]
+    class PatientRecord
+    {
+        /**
+         * Encrypted with equality queries.
+         * This allows us to find a patient by their exact SSN.
+         */
+        #[ODM\Field(type: 'string')]
+        #[Encrypt(queryType: EncryptQuery::Equality)]
+        public string $ssn;
+
+        /**
+         * The entire embedded document is encrypted as an object.
+         * By not specifying a queryType, we make it non-queryable.
+         */
+        #[ODM\EmbedOne(targetDocument: Billing::class)]
+        #[Encrypt]
+        public Billing $billing;
+
+        /**
+         * Encrypted with range queries.
+         * This allows us to query for billing amounts within a certain range.
+         */
+        #[ODM\Field(type: 'int')]
+        #[Encrypt(queryType: EncryptQuery::Range, min: 0, max: 5000, sparsity: 1)]
+        public int $billingAmount;
+    }
+
+    #[ODM\EmbeddedDocument]
+    class Billing
+    {
+        #[ODM\Field(type: 'string')]
+        public string $creditCardNumber;
+    }
+
+Configuration and Usage
+-----------------------
+
+The following example demonstrates how to configure the ``DocumentManager`` for
+encryption and how to work with encrypted documents.
+
+Step 1: Configure the DocumentManager
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+First, we configure the ``DocumentManager`` with ``autoEncryption`` options.
+For more details on the available options, see the `MongoDB\Driver\Manager`_
+documentation. We'll use the ``local`` KMS provider for simplicity. For this
+provider, you need a 96-byte master key.
+The following code will look for the key in a local file (``master-key.bin``)
+and generate it if it doesn't exist. In a production environment, you should
+use a non-local key management service (KMS).
+For each field marked with ``#[Encrypt]``, the MongoDB driver will generate
+a Data Encryption Key (DEK), encrypt it with the master key, and store it in
+the key vault collection. In Doctrine ODM, the key vault collection is set
+to ``<database>.datakeys`` by default, but you can change it using the
+``keyVaultNamespace`` option.
+
+.. code-block:: php
+
+    <?php
+
+    use Doctrine\ODM\MongoDB\Configuration;
+    use Doctrine\ODM\MongoDB\Mapping\Driver\AttributeDriver;
+    use MongoDB\BSON\Binary;
+
+    // For the local KMS provider, we need a 96-byte master key.
+    // We'll store it in a local file. If the file doesn't exist, we generate
+    // one. In a production environment, ensure this key file is properly
+    // secured.
+    $keyFile = __DIR__ . '/master-key.bin';
+    if (!file_exists($keyFile)) {
+        file_put_contents($keyFile, random_bytes(96));
+    }
+    $masterKey = new Binary(file_get_contents($keyFile), Binary::TYPE_GENERIC);
+
+    $config = new Configuration();
+    // Enable auto encryption and set the KMS provider.
+    $config->setAutoEncryption([
+        'keyVaultNamespace' => 'encryption.datakeys'
+    ]);
+    $config->setKmsProvider([
+        'type' => 'local',
+        'key' => new Binary($masterKey),
+    ]);
+
+    // Other configuration
+    $config->setProxyDir(__DIR__ . '/Proxies');
+    $config->setProxyNamespace('Proxies');
+    $config->setHydratorDir(__DIR__ . '/Hydrators');
+    $config->setHydratorNamespace('Hydrators');
+    $config->setPersistentCollectionDir(__DIR__ . '/PersistentCollections');
+    $config->setPersistentCollectionNamespace('PersistentCollections');
+    $config->setDefaultDB('my_db');
+    $config->setMetadataDriverImpl(new AttributeDriver([__DIR__]));
+
+Step 2: Create the DocumentManager
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+The ``MongoDB\Client`` will be instantiated with the options from the
+configuration.
+
+.. code-block:: php
+
+    <?php
+
+    use Doctrine\ODM\MongoDB\DocumentManager;
+    use MongoDB\Client;
+
+    $client = new Client(
+        uri: 'mongodb://localhost:27017/',
+        uriOptions: [],
+        driverOptions: $config->getDriverOptions(),
+    );
+    $documentManager = DocumentManager::create($client, $config);
+
+The ``driverOptions`` passed to the client contain the ``autoEncryption`` option
+that was configured in the previous step.
+
+Step 3: Create the Encrypted Collection
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Next, we use the ``SchemaManager`` to create the collection with the necessary
+encryption metadata. To make the example re-runnable, we can drop the collection
+first.
+
+.. code-block:: php
+
+    <?php
+
+    $schemaManager = $documentManager->getSchemaManager();
+    $schemaManager->dropDocumentCollection(Patient::class);
+    $schemaManager->createDocumentCollection(Patient::class);
+
+Step 4: Persist and Query Documents
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Finally, we can persist and query documents as usual. The encryption and
+decryption will be handled automatically.
+
+.. code-block:: php
+
+    <?php
+
+    $patient = new Patient();
+    $patient->patientRecord = new PatientRecord();
+    $patient->patientRecord->ssn = '123-456-7890';
+    $patient->patientRecord->billingAmount = 1500;
+    $patient->patientRecord->billing = new Billing();
+    $patient->patientRecord->billing->creditCardNumber = '9876-5432-1098-7654';
+
+    $documentManager->persist($patient);
+    $documentManager->flush();
+    $documentManager->clear();
+
+    // Query the document using an encrypted field
+    $foundPatient = $documentManager->getRepository(Patient::class)->findOneBy([
+        'patientRecord.ssn' => '123-456-7890',
+    ]);
+
+    // The document is retrieved and its fields are automatically decrypted
+    assert($foundPatient instanceof Patient);
+    assert($foundPatient->patientRecord->billingAmount === 1500);
+
+What the Document Looks Like in the Database
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+When you inspect the document directly in the database (e.g., using ``mongosh``
+or `MongoDB Compass`_), you will see that the fields marked with ``#[Encrypt]``
+are stored as BSON binary data (subtype 6), not the original BSON type. The
+driver also adds a ``__safeContent__`` field to the document. For more details,
+see the `Queryable Encryption Fundamentals`_ in the MongoDB manual.
+
+.. code-block:: js
+
+    {
+      "_id": ObjectId("..."),
+      "patientRecord": {
+        "ssn": Binary("...", 6),
+        "billing": Binary("...", 6),
+        "billingAmount": Binary("...", 6)
+      },
+      "__safeContent__": [
+        Binary("...", 0)
+      ]
+    }
+
+Limitations
+-----------
+
+- The ODM simplifies configuration by supporting a single KMS provider per
+  ``DocumentManager`` through ``Configuration::setKmsProvider()``. If you need
+  to work with multiple KMS providers, you must manually configure the
+  ``kmsProviders`` array and pass it as a driver option, bypassing the ODM's
+  helper method.
+- Automatic generation of the ``encryptedFieldsMap`` is not compatible with
+  ``SINGLE_COLLECTION`` inheritance. Because all classes in the hierarchy
+  share a single collection, they must also share a single encryption schema.
+  To use QE with inheritance, you must manually define the complete
+  ``encryptedFieldsMap`` for the entire hierarchy and provide it directly in
+  the client options, bypassing the ODM's automatic generation.
+- For a complete list of hard limitations, please refer to the official
+  `Queryable Encryption Limitations`_ documentation.
+
+.. _MongoDB\Driver\Manager: https://www.php.net/manual/en/mongodb-driver-manager.construct.php#mongodb-driver-manager.construct-autoencryption
+.. _MongoDB Compass: https://www.mongodb.com/products/compass
+.. _Queryable Encryption Fundamentals: https://www.mongodb.com/docs/manual/core/queryable-encryption/fundamentals/#behavior
+.. _Queryable Encryption Limitations: https://www.mongodb.com/docs/manual/core/queryable-encryption/reference/limitations/

docs/en/reference/attributes-reference.rst

@@ -336,6 +336,94 @@ Unlike normal documents, embedded documents cannot specify their own database or
 collection. That said, a single embedded document class may be used with
 multiple document classes, and even other embedded documents!
 
+.. _encrypt_attribute:
+
+#[Encrypt]
+----------
+
+The ``#[Encrypt]`` attribute is used to define an encrypted field mapping for a
+document property. It allows you to configure fields for automatic and queryable
+encryption in MongoDB.
+
+Optional arguments:
+
+- ``queryType`` - Specifies the query type for the field. Possible values:
+  - ``null`` (default) - Field is not queryable.
+  - ``EncryptQuery::Equality`` - Enables equality queries.
+  - ``EncryptQuery::Range`` - Enables range queries.
+- ``min``, ``max`` - Specify minimum and maximum (inclusive) queryable values
+  for a field when possible, as smaller bounds improve query efficiency. If
+  querying values outside of these bounds, MongoDB returns an error.
+- ``sparsity``, ``precision``, ``trimFactor``, ``contention`` - For advanced
+  users only. The default values for these options are suitable for the majority
+  of use cases, and should only be modified if your use case requires it.
+
+.. note::
+
+    Queryable encryption is only supported in MongoDB version 7.0 and later.
+
+Example:
+
+.. code-block:: php
+
+    <?php
+
+    use Doctrine\ODM\MongoDB\Mapping\Annotations\Encrypt;
+    use Doctrine\ODM\MongoDB\Mapping\Annotations\EncryptQuery;
+
+    #[Document]
+    class Client
+    {
+        #[Field]
+        #[Encrypt(queryType: EncryptQuery::Equality)]
+        public string $name;
+    }
+
+The ``#[Encrypt]`` attribute can be added to a class with `#[EmbeddedDocument]`_.
+This will encrypt the entire embedded document, in the field that contains it.
+Fields within an encrypted embedded document cannot be individually encrypted.
+Queryable encryption is not supported for embedded documents, so the ``queryType``
+argument is not applicable. Encrypted embedded documents are stored as a binary
+value in the parent document.
+
+.. code-block:: php
+
+    <?php
+
+    use Doctrine\ODM\MongoDB\Mapping\Annotations\Encrypt;
+
+    #[Encrypt]
+    #[EmbeddedDocument]
+    class CreditCard
+    {
+        #[Field]
+        public string $number;
+
+        #[Field]
+        public string $expiryDate;
+    }
+
+    #[Document]
+    class User
+    {
+        #[EmbedOne(targetDocument: CreditCard::class)]
+        public CreditCard $creditCard;
+    }
+
+For more details, refer to the MongoDB documentation on
+`Queryable Encryption <https://www.mongodb.com/docs/manual/core/queryable-encryption/fundamentals/encrypt-and-query/>`_.
+
+
+.. note::
+
+    The encrypted collection must be created with the `Schema Manager`_ before
+    before inserting documents.
+
+.. note::
+
+    Due to the way the encrypted fields map is generated, the queryable encryption
+    is not compatible with ``SINGLE_COLLECTION`` inheritance.
+
 #[Field]
 --------
 
@@ -1399,5 +1487,6 @@ root class specified in the view mapping.
 .. _DBRef: https://docs.mongodb.com/manual/reference/database-references/#dbrefs
 .. _geoNear command: https://docs.mongodb.com/manual/reference/command/geoNear/
 .. _MongoDB\BSON\ObjectId: https://www.php.net/class.mongodb-bson-objectid
+.. _Schema Manager: ../reference/migrating-schemas
 .. |FQCN| raw:: html
   <abbr title="Fully-Qualified Class Name">FQCN</abbr>

docs/en/reference/basic-mapping.rst

@@ -145,6 +145,7 @@ Here is a quick overview of the built-in mapping types:
 -  ``hash``
 -  ``id``
 -  ``int``
+-  ``int64``
 -  ``key``
 -  ``object_id``
 -  ``raw``