Revision: Wed, 27 Nov 2024 11:25:26 GMT
v3.12 – outdated
This version of the documentation is outdated. Consider upgrading your project to Spiral Framework 3.14
Edit this page

Testing — Database Seeder

When you build apps that use databases, it's really important to make sure your database works right. This means checking if it stores, changes, and gives back data the way it's supposed to. But sometimes, testing databases can be tricky and a bit boring. You might have to write a lot of complicated commands and be very careful about how data is added or removed.

This is where Spiral's database testing tools come in handy. They make this job much easier and faster. Spiral has a special set of tools in a package called spiral-packages/database-seeder. It's designed to help you test databases without too much hassle.

What Database Seeder tools offer

  1. Easy Testing: With Spiral, you don't need to deal with complex commands. The tools are simple to use, which means your tests are easier to write and understand.

  2. Different Ways to Reset Your Database: After you test something, you need to make your database clean again for the next test. Spiral has different ways to do this, like the Transaction, Migration, Refresh, and SqlFile methods. Each one has its own way of working, so you can choose what fits best for your test.

  3. Seeders and Factories: These are like shortcuts to fill your database with test data. This data looks like the real data you would use in your app. You can quickly set up the data you need for testing with these tools.

  4. Checking Your Database: After you do something in your database, you want to make sure it worked right. Spiral's tools let you check if the data is there or not, and if your database structure is correct.

Spiral's database testing tools are great for any developer, no matter how much experience you have. They help make sure your database is doing what it should, which is really important for your app to work well.

Installation

To install the package, run the following command:

composer require spiral-packages/database-seeder --dev

Note
It's important to note that the --dev flag is used because the package is intended for use in development and testing environments.

After package install you need to register bootloader from the package.

php
app/src/Application/Kernel.php
public function defineBootloaders(): array
{
    return [
        // ...
        \Spiral\DatabaseSeeder\Bootloader\DatabaseSeederBootloader::class,
        // ...
    ];
}

Read more about bootloaders in the Framework — Bootloaders section.

We also recommend you to add the following to your composer.json file. This allows you to separate your factories, seeders from the rest of your application code.

json
{
  // ...
  "autoload": {
    "psr-4": {
      "Database\\": "app/database/"
    }
  }
}

After these steps, you have installed the package and registered required bootloader, so you can now use the package in your application.

Setting up database testing environment

Environment variables

When running unit tests in a Spiral environment, a crucial step to improve efficiency is the proper configuration of environment variables. Specifically, setting CYCLE_SCHEMA_CACHE and TOKENIZER_CACHE_TARGETS to true in your phpunit.xml file can significantly enhance the speed of your tests. These settings prevent unnecessary repeated operations, such as directory scanning and rebuilding of the Cycle ORM cache for each test run.

To apply these optimizations, modify your phpunit.xml file accordingly. Here’s an example of how you might set it up:

xml
phpunit.xml
<?xml version="1.0" encoding="UTF-8"?>
<phpunit>
    <php>
        <env name="CYCLE_SCHEMA_CACHE" value="true"/>
        <env name="TOKENIZER_CACHE_TARGETS" value="true"/>
        <!-- ... other configurations ... -->
    </php>
</phpunit>

When you're testing your database, one key step is setting up your environment correctly. This means getting your database ready for each test and then putting things back the way they were afterward. There are several strategies to help you do this in Spiral. Let's look at different ways to set up and reset your database for testing, using simple test cases as examples.

1. Transaction Strategy

It sets up your database for testing when you run your first test. Then, for each test, it does everything inside a transaction – a kind of protected space. After the test is done, it undoes everything (rolls back) that happened in the transaction. Imagine you're testing adding a new user to your database. The Transaction Strategy would add this user inside a transaction and then, after the test, remove any trace of this new user, as if they were never added.

This is the fastest way and is great for most tests. It's especially good when you have lots of tests that don't change the database structure but just add, change, or remove data.

Here is an example of a test case that uses it:

php
declare(strict_types=1);

namespace Tests;

use Spiral\DatabaseSeeder\Database\Traits\{
    DatabaseAsserts, Helper, ShowQueries, Transactions
};

abstract class DatabaseTestCase extends TestCase
{
    use Transactions,
        Helper,
        DatabaseAsserts,
        ShowQueries;
}

Let's examine each trait used:

  • Transactions trait: It manages the wrapping of each test case in a database transaction. When a test begins, it starts a transaction. Once the test is completed, whether it passes or fails, the transaction is rolled back. This ensures that each test is isolated and the database is left unchanged after the test, making the tests independent of each other. It also handles running database migrations before the tests, setting up the necessary database schema.
  • Helper trait: It provides a set of helper methods for interacting with the database. It includes utility functions that simplify various database operations, making it easier to perform common tasks in your tests without writing repetitive code.
  • DatabaseAsserts trait: It offers a range of assertions you can use to validate the state of your database, such as checking if a record exists, verifying the count of records, etc. It's crucial for ensuring that your database operations produce the expected results.
  • ShowQueries trait: This is particularly useful for debugging. When you enable this trait, it logs all the SQL queries executed during the test to the terminal. This can help you understand what's happening in your database during tests and troubleshoot any issues that arise.

2. Migration Strategy

This method sets up your database using migrations (steps to prepare your database) before each test. After the test, it rollbacks these migrations. t's slower than the Transaction Strategy but useful if you need to test changes in the database structure itself.

Example: If you're testing creating a new table in your database, the Migration Strategy would create this table for the test and then remove it afterward.

Here is an example of a test case that uses it:

php
declare(strict_types=1);

namespace Tests;

use Spiral\DatabaseSeeder\Database\Traits\{
    DatabaseAsserts, DatabaseMigrations, Helper, ShowQueries
};

abstract class DatabaseTestCase extends TestCase
{
    use DatabaseMigrations,
        Helper,
        DatabaseAsserts,
        ShowQueries;
}
  • DatabaseMigrations trait: It manages the running of database migrations before each test and rolling them back afterward. This ensures that each test is isolated and the database is left unchanged after the test, making the tests independent of each other. It also handles running database migrations before the tests, setting up the necessary database schema.

3. Refresh Strategy

This strategy cleans your database after each test using Spiral\DatabaseSeeder\Database\Cleaner. It's useful if your database already has the structure you need, and you're just testing data changes. It's slower than the Transaction Strategy.

Here is an example of a test case that uses it:

php
declare(strict_types=1);

namespace Tests;

use Spiral\DatabaseSeeder\Database\Traits\{
    DatabaseAsserts, Helper, RefreshDatabase, ShowQueries
};

abstract class DatabaseTestCase extends TestCase
{
    use RefreshDatabase,
        Helper,
        DatabaseAsserts,
        ShowQueries;
}
  • RefreshDatabase trait: It manages the cleaning of the database tables (except migrations table) after each test.

4. SqlFile Strategy

This approach uses an SQL file to set up your database with the necessary structure and data for testing. It's handy when you have a complex setup that can be easily loaded from a file, but it's slower than the Transaction Strategy.

Example: If you need a specific setup with many tables and data, the SqlFile Strategy lets you load all this from a pre-prepared SQL file.

Here is an example of a test case that uses it:

php
declare(strict_types=1);

namespace Tests;

use Spiral\DatabaseSeeder\Database\Traits\{
    DatabaseAsserts, Helper, DatabaseFromSQL, ShowQueries
};

abstract class DatabaseTestCase extends TestCase
{
    use DatabaseFromSQL,
        Helper,
        DatabaseAsserts,
        ShowQueries;
   
   protected function getPrepareSQLFilePath(): string
   {
       return __DIR__ . '/database/prepare.sql';
   }

    protected function getDropSQLFilePath(): string
    {
        return __DIR__ . '/database/prepare.sql';
    }
}

As you can see, there are lots of ways to set up and reset your database for testing. You can choose the one that fits your needs best. After you set up your database, you can start defining your factories and seeders.

Defining entity factories

To define a seed factory, you should create a class that extends the Spiral\DatabaseSeeder\Factory\AbstractFactory class.

Here is an example of a seed factory that defines a User entity:

php
namespace Database\Factory;

use App\Entity\User;
use Spiral\DatabaseSeeder\Factory\AbstractFactory;
use Spiral\DatabaseSeeder\Factory\FactoryInterface;

/**
 * @implements FactoryInterface<User>
 */
final class UserFactory extends AbstractFactory
{
    public function entity(): string
    {
        return User::class;
    }
    
    public function makeEntity(array $definition): User
    {
        return new User(
            username: $definition['username']
        );
    }

    public function definition(): array
    {
        return [
            'firstName' => $this->faker->firstName(),
            'lastName' => $this->faker->lastName(),
            'birthday' => \DateTimeImmutable::createFromMutable($this->faker->dateTime()),
            'comments' => CommentFactory::new()->times(3)->make(), // Can use other factories.
            // Be careful, circular dependencies are not allowed!
        ];
    }
}

Note
Since v2.4.0 you can define the return type for a factory @implements FactoryInterface<...> annotation as in the example above. With this annotation in place, your IDE will now provide suggestive autocomplete options, making the code interaction more intuitive and error-prone. image

The entity method should return the fully qualified class name of the target entity that the factory is responsible for creating. In some cases, the factory may use the class name returned by the entity method to create a new instance of the entity without calling its constructor. Instead, it may use reflection to directly set the properties of the entity using data from the definition method.

The makeEntity method allows you to control the process of creating an entity through its constructor. The method takes an array of definitions as an argument, which is generated by the definition method.

The definition method is where you can define all of the properties of the entity. The method should return an array where the keys are the property names and the values are the values that should be set for those properties. These values can be hard-coded or generated using the Faker library, which provides a wide range of fake data generation methods such as names, addresses, phone numbers, etc. This method is responsible for generating the definition array that will be passed to the makeEntity method to construct the entity object or used to set properties directly.

Using entity factory

A factory can be created by calling the new method on the factory class:

php
use Database\Factory\UserFactory;

$factory = UserFactory::new();

This will create a new instance of the factory. It provides several useful methods for generating entities.

You can also pass an array of definition to the new method of the factory class.

php
use Database\Factory\UserFactory;

$factory = UserFactory::new(['admin' => true]);

This is a useful feature when you have a common set of definitions that you want to use across multiple factories or when you want to set a default value for a property that will be overridden in specific cases.

Create entities

The create method creates an array of entities, stores them in the database and returns them for further use in the code.

php
$users = $factory->times(10)->create();

The createOne method creates a single entity, stores it in the database and returns it for further use in the code.

php
$user = $factory->createOne();

Make entities

The make method creates an array of entities and returns them for further use in code, but does not store them in the database.

php
$users = $factory->times(10)->make();

The makeOne method creates a single entity and returns it for further use in code, but does not store it in the database.

php
$user = $factory->makeOne();

Entity states

The state method allows developers to easily define specific states for entities. It can be used to set a specific set of properties on an entity when it is created. The returned values from the closure will replace the corresponding values in the definition array, allowing developers to easily change the state of an entity in a specific way.

Note
It is non-destructive, it will only update the properties passed in the returned array and will not remove any properties from the definition array.

php
$factory->state(fn(\Faker\Generator $faker, array $definition) => [
    'admin' => $faker->boolean(),
])->times(10)->create();

In addition to the state method, there also the entityState method. This method allows developers to change the state of an entity object using the available methods of that entity. It takes a closure as an argument, which should accept the entity as an argument and should return the modified entity. This allows developers to take full advantage of the object-oriented nature of their entities and use the methods that are already defined on the entity to change its state.

php
$factory->entityState(static function(User $user) {
    return $user->markAsDeleted();
})->times(10)->create();

The state and entityState methods can be used inside a factory class to create additional methods for creating entities with specific states. By creating these additional methods, developers can create a more expressive and readable code, and make it easier to understand the intent of the code.

php
final class UserFactory extends AbstractFactory
{
    // ....

    public function admin(): self
    {
        return $this->state(fn(\Faker\Generator $faker, array $definition) => [
            'admin' => true,
        ]);
    }
    
    public function fromCity(string $city): self
    {
        return $this->state(fn(\Faker\Generator $faker, array $definition) => [
            'city' => $city,
        ]);
    }
    
    public function deleted(): self
    {
        return $this->entityState(static function (User $user) {
            return $user->markAsDeleted();
        });
    }
    
    public function withBirthday(\DateTimeImmutable $date): self
    {
        return $this->entityState(static function (User $user) use ($date) {
            $user->birthday = $date;
            return $user;
        });
    }
}

And example of using these methods:

php
$factory->admin()
    ->fromCity('New York')
    ->deleted()
    ->withBirthday(new \DateTimeImmutable('2010-01-01 00:00:00'))
    ->times(10)
    ->create();

Factories can be used in your feature test cases to create entities in the database without seeding. This can be useful in situations where you want to create a specific set of test data for a feature test, or when you want to test the behavior of your application with a specific set of data.

Here is an example of using a factory in a feature test case:

php
final class UserServiceTest extends DatabaseTestCase
{
    // ...

    public function testDeleteUser(): void
    {
        $user = UserFactory::new()
            ->fromCity('New York')
            ->createOne();

        $this->userService->delete(
            uuid: $user->uuid,
        );

        $this->assertEntity(User::class)->where([
            'uuid' => (string)$user->uuid,
            'deleted_at' => ['!=' => null],
        ])->assertExists();
    }
}

Seeding

The package provides the ability to seed the database with test data. To do this, developers can create a Seeder class and extend it from the Spiral\DatabaseSeeder\Seeder\AbstractSeeder class. The Seeder class should implement the run method which returns a generator with entities to store in the database.

php
namespace Database\Seeder;

use Spiral\DatabaseSeeder\Seeder\AbstractSeeder;
use Spiral\DatabaseSeeder\Attribute\Seeder;

#[Seeder]
final class UserTableSeeder extends AbstractSeeder
{
    public function run(): \Generator
    {
        foreach (UserFactory::new()->times(100)->make() as $user) {
            yield $user;
        }
        
        foreach (UserFactory::new()->admin()->times(10)->make() as $user) {
            yield $user;
        }
        
        foreach (UserFactory::new()->admin()->deleted()->times(1)->make() as $user) {
            yield $user;
        }
    }
}

Note
Don't forget to add the #[Seeder] attribute to your seeder class. This attribute is used to register the seeder class with the package. If you don't add this attribute, the seeder class will not be registered and the seeder will not be executed.

Seeders are primarily used to fill the database with test data for your stage server, providing a consistent set of data for the developers and stakeholders to test and use.

They are especially useful when testing complex applications that have many different entities and relationships between them. By using seeders to populate the database with test data, you can ensure that your tests are run against a consistent and well-defined set of data, which can help to make your tests more reliable and less prone to flakiness.

Running seeders

Use the following command to run the seeders:

php app.php db:seed

This will execute all of the seeder classes that are registered and insert the test data into the database.

Assertions

The package provides several additional assertion methods with Spiral\DatabaseSeeder\Database\Traits\DatabaseAsserts trait. These methods can be used to test the state of the database after performing some operation.

Entity assertions

The assertEntity method can be used to assert the state of an entity in the database. It takes the entity class name and provides methods for testing database entities:

assertExists()

Checks if the entity exists in the database.

php
$this->assertEntity(User::class)->where([
    'uuid' => (string)$user->uuid,
    'deleted_at' => ['!=' => null],
])->assertExists();

assertMissing()

Checks that the entity is not in the database.

php
$this->assertEntity(User::class)->where([
    'uuid' => (string)$user->uuid,
])->assertMissing();

assertCount(int $total)

Checks that there is a specified number of entity items in the database.

php
$this->assertEntity(User::class)->where([
    'uuid' => (string)$user->uuid,
])->assertCount(1);

assertEmpty()

Checks that there are no entity items in the database.

php
$this->assertEntity(User::class)->assertEmpty();

where(array $where)

Allows you to specify the conditions for the entity in the database.

php
$this->assertEntity(User::class)->where([
    'uuid' => (string)$user->uuid,
])->...;

count()

Returns the number of entity items in the database.

php
$total = $this->assertEntity(User::class)->where([
    'deleted_at' => ['!=' => null],
])->count();

select(\Closure $closure)

Allows you to make complex queries to the database.

php
$this->assertEntity(User::class)->select(function(\Cycle\ORM\Select $select) {
    $select->where('deleted_at', '!=', null);
})->...;

withoutScope

Allows you to disable the global scope for the entity.

php
$this->assertEntity(User::class)->withoutScope()->...;

withScope(ScopeInterface $scope)

Allows you to enable the global scope for the entity.

php
$this->withScope(new SoftDeletedScope())->...;

Table assertions

select(\Closure $closure)

Allows you to make complex queries to the database.

php
$this->assertTable('users')->select(function(\Cycle\Database\Query\SelectQuery $query) {
    $query->where('deleted_at', '!=', null);
})->...;

where(array $where)

Allows you to specify the conditions for the table in the database.

php
$this->assertTable('users')->where([
    'id' => 5,
])->assertCount(1);

assertRecordExists()

Allows you to check if record exists in the database.

php
$this->assertTable('users')->where([
    'id' => 5,
])->assertRecordExists();

assertCountRecords(int $total)

Allows you to check if there is a specified number of records in the database.

php
$this->assertTable('users')->where([
    'id' => 5,
])->assertCountRecords(2);

countRecords()

Returns the number of records in the database.

php
$total = $this->assertTable('users')->where([
    'deleted_at' => ['!=' => null],
])->countRecords();

assertEmpty()

Allows you to check if there are no records in the database.

php
$this->assertTable('users')->where([
    'id' => 5,
])->assertEmpty();

assertExists()

Allows you to check if table exists in the database.

php
$this->assertTable('users')->assertExists();

assertColumnExists(string $column)

Allows you to check if column exists in the table.

php
$this->assertTable('users')->assertColumnExists('id');

assertColumnMissing(string $column)

Allows you to check if column missing in the table.

php
$this->assertTable('users')->assertColumnMissing('id');

assertColumnSame(...)

Asserts that a column in the table matches specified characteristics, including type, size, precision, scale, nullability, and default value. This method is particularly useful for validating the detailed schema of a table.

php
$assertion = $this->assertTable('users');

$assertion->assertColumnSame(
     column: 'id', 
     type: 'uuid', 
     size: 64,
     ...
);

assertIndexPresent(array $columns)

Asserts the presence of an index on specified columns in the table. It's vital for ensuring that necessary indexes, which can affect performance and uniqueness constraints, are in place.

php
$this->assertTable('users')->assertIndexPresent(['email']);

assertForeignKeyPresent(array $columns)

Checks for the existence of a foreign key related to specified table columns. It's important for validating relational integrity constraints in the database schema.

php
$this->assertTable('users')->assertForeignKeyPresent(['org_id']);

assertPrimaryKeyExists(string ...$columns)

Asserts that specified columns are part of the primary key for the table. It's crucial for ensuring the correct configuration of primary keys, which are fundamental to the identification and integrity of records in a table.

php
$this->assertTable('users')->assertPrimaryKeyExists('id');

Console commands

The package provides several console commands to quickly create a factory, seeder. These commands make it easy to quickly create the necessary classes for seeding and testing your application, without having to manually create the files or ensure that the necessary boilerplate code is included.

Create a factory

The Spiral\DatabaseSeeder\Console\Command\FactoryCommand console command is used to create a factory. The name of the factory is passed as an argument.

For example, to create a factory named UserFactory, the command would be:

php app.php create:factory UserFactory

Create a seeder

The Spiral\DatabaseSeeder\Console\Command\SeederCommand console command is used to create a seeder. The name of the seeder is passed as an argument.

For example, to create a seeder named UserSeeder, the command would be:

php app.php create:seeder UserSeeder