Skip to content

Latest commit

 

History

History
559 lines (415 loc) · 15.4 KB

README.md

File metadata and controls

559 lines (415 loc) · 15.4 KB

Strongly typed hydrator for PHP 7.4+ with extensibility

Build Status Code Coverage Scrutinizer Code Quality Total Downloads Latest Stable Version License

php, dto, hydrator, mapper, populator, data-mapper


Installation

composer require sunrise/hydrator

How to use

Let's consider a typical DTO set:

enum Status: int
{
    case Disabled = 0;
    case Enabled = 1;
}
final readonly class CategoryDto
{
    public function __construct(
        public string $name,
    ) {
    }
}
final readonly class TagDto
{
    public function __construct(
        public string $name,
    ) {
    }
}
final class TagDtoCollection extends ArrayObject
{
    public function __construct(TagDto ...$tags)
    {
        parent::__construct($tags);
    }
}
final readonly class PublicationDto
{
    public function __construct(
        public string $name,
        public CategoryDto $category,
        public TagDtoCollection $tags,
        public Status $status = Status::Disabled,
        #[\Sunrise\Hydrator\Annotation\Format(DateTimeInterface::RFC3339)]
        public DateTimeImmutable $createdAt = new DateTimeImmutable(),
    ) {
    }
}

Now, let's populate them all from an array:

$data = [
    'name' => 'Some product',
    'category' => [
        'name' => 'Some category',
    ],
    'tags' => [
        [ 'name' => 'foo' ],
        [ 'name' => 'bar' ],
    ],
    'status' => 0,
];

$product = (new \Sunrise\Hydrator\Hydrator)->hydrate(PublicationDto::class, $data);

Or, you can populate them using JSON:

$json = <<<JSON
{
    "name": "Some product",
    "category": {
        "name": "Some category"
    },
    "tags": [
        { "name": "foo" },
        { "name": "bar" }
    ],
    "status": 0
}
JSON;

$product = (new \Sunrise\Hydrator\Hydrator)->hydrateWithJson(PublicationDto::class, $json);

Allowed property types

Required

public readonly string $value;

If a property has no a default value, then the property is required.

Optional

public string $value = 'foo';

If a property has a default value, then the property is optional.

There are situations when it is necessary to assign a value to a readonly property. This cannot be done outside the constructor; however, you can use a special annotation, as shown in the example below:

#[\Sunrise\Hydrator\Annotation\DefaultValue('foo')]
public readonly string $value;

Null

public readonly ?string $value;

If a property is nullable, then the property can accept null.

Also, please note that empty strings or strings consisting only of whitespace will be handled as null for certain properties.

Boolean

public readonly bool $value;

A boolean value in a dataset can be represented as:

true false type
true false bool
"1" "0" string
"true" "false" string
"yes" "no" string
"on" "off" string

Also, please note that if a value in a dataset for this property is represented as an empty string or a string consisting only of whitespace, then the value will be handled as null.

Integer

public readonly int $value;

An integer value in a dataset can also be represented as a string, e.g., "42".

Also, please note that if a value in a dataset for this property is represented as an empty string or a string consisting only of whitespace, then the value will be handled as null.

Number

public readonly float $value;

A numerical value in a dataset can be represented not only as an integer or a floating-point number but also as a string containing a number. However, regardless of how such a value is represented in the dataset, it will always be stored in this property as a floating-point number.

Also, please note that if a value in a dataset for this property is represented as an empty string or a string consisting only of whitespace, then the value will be handled as null.

String

public readonly string $value;

This property has no any additional behavior and only accepts strings.

Array

public readonly array $value;

By default, this property accepts an array with any data. However, it can also be used to store relationships by using a special annotation, as shown in the example below:

#[\Sunrise\Hydrator\Annotation\Subtype(SomeDto::class)]
public readonly array $value;

In other words, the Subtype annotation can contain the same types as the types of class properties...

Having an unlimited number of relationships in an array is a potentially bad idea as it can lead to memory leaks. To avoid this, it is recommended to limit such an array, as shown in the example below:

#[\Sunrise\Hydrator\Annotation\Subtype(SomeDto::class, limit: 100)]
public readonly array $value;

In addition to arrays, you can also use collections, in other words, classes implementing the ArrayAccess interface, for example:

final class TagDto
{
}
final class TagDtoCollection implements \ArrayAccess
{
}
final class CreateProductDto
{
    public function __construct(
        #[\Sunrise\Hydrator\Annotation\Subtype(TagDto::class, limit: 10)]
        public readonly TagDtoCollection $tags,
    ) {
    }
}

Note that for collections, instead of the Subtype annotation, you can use typing through its constructor. It is important that there is only one variadic parameter in it. Please refer to the example below:

Please note that in this case, you take on the responsibility of limiting the collection. To ensure that the hydrator understands when the collection is full, the offsetSet method should throw an OverflowException.

final class TagDtoCollection implements \ArrayAccess
{
    public function __construct(public TagDto ...$tags)
    {
    }
}
final class CreateProductDto
{
    public function __construct(
        public readonly TagDtoCollection $tags,
    ) {
    }
}

In general, remember that regardless of whether arrays or collections are used, their elements can be typed. For example, if you need an array that should consist only of dates, your code should look something like this:

#[\Sunrise\Hydrator\Annotation\Subtype(\DateTimeImmutable::class, limit: 100)]
#[\Sunrise\Hydrator\Annotation\Format('Y-m-d H:i:s')]
public readonly array $value;

Sometimes, there is a need to map a typed list to a collection. It's a relatively simple task; just use the example below:

use Sunrise\Hydrator\Type;

$data = [...];

$collection = $hydrator->castValue($data, Type::fromName(SomeDtoCollection::class));

This property has no any additional behavior and only accepts arrays.

Timestamp

Only the DateTimeImmutable type is supported.

#[\Sunrise\Hydrator\Annotation\Format('Y-m-d H:i:s')]
public readonly DateTimeImmutable $value;

This property accepts a date as a string in the specified format, but it can also accept a Unix timestamp as an integer or a string. To specify the Unix timestamp format, it should be indicated as follows:

#[\Sunrise\Hydrator\Annotation\Format('U')]
public readonly DateTimeImmutable $value;

Also, please note that if a value in a dataset for this property is represented as an empty string or a string consisting only of whitespace, then the value will be handled as null.

Default timestamp format

use Sunrise\Hydrator\Dictionary\ContextKey;
use Sunrise\Hydrator\Hydrator;

$hydrator = new Hydrator([
    ContextKey::TIMESTAMP_FORMAT => 'Y-m-d H:i:s',
]);

Timezone

Only the DateTimeZone type is supported.

public readonly DateTimeZone $value;

Also, please note that if a value in a dataset for this property is represented as an empty string or a string consisting only of whitespace, then the value will be handled as null.

Default timezone

use Sunrise\Hydrator\Dictionary\ContextKey;
use Sunrise\Hydrator\Hydrator;

$hydrator = new Hydrator([
    ContextKey::TIMEZONE => 'Europe/Kyiv',
]);

UUID

Using the ramsey/uuid package

composer require ramsey/uuid
public readonly \Ramsey\Uuid\UuidInterface $value;

Using the symfony/uid package

composer require symfony/uid
public readonly \Symfony\Component\Uid\UuidV4 $value;

Also, please note that if a value in a dataset for this property is represented as an empty string or a string consisting only of whitespace, then the value will be handled as null.

Enumeration

PHP 8.1 built-in enumerations

public readonly SomeEnum $value;

This property should be typed only with typed enumerations. Therefore, for integer enumerations, a value in a dataset can be either an integer or an integer represented as a string. For string enumerations, a value in a dataset can only be a string.

MyCLabs enumerations

The popular alternative for PHP less than 8.1...

composer require myclabs/php-enum
public readonly SomeEnum $value;

Also, please note that if a value in a dataset for this property is represented as an empty string or a string consisting only of whitespace, then the value will be handled as null.

Relationship

public readonly SomeDto $value;

A value in a dataset can only be an array. However, please note that if you need a one-to-many relationship, you should refer to the array section for further information.

Support for custom types

If you need support for a custom type, it is a relatively simple task. Let's write such support for PSR-7 URI from the sunrise/http-message package:

composer require sunrise/http-message
use Sunrise\Hydrator\Exception\InvalidValueException;
use Sunrise\Hydrator\Type;
use Sunrise\Hydrator\TypeConverterInterface;
use Psr\Message\UriInterface;
use Sunrise\Http\Message\Uri;

final class UriTypeConverter implements TypeConverterInterface
{
    public function castValue($value, Type $type, array $path): Generator
    {
        if ($type->getName() <> UriInterface::class) {
            return;
        }

        if (!\is_string($value)) {
            throw InvalidValueException::mustBeString($path, $value);
        }

        try {
            yield new Uri($value);
        } catch (\InvalidArgumentException $e) {
            throw new InvalidValueException(
                'This value is not a valid URI.',
                'c66741c6-e3c0-4522-a8e3-97528d7712a3',
                $path,
            );
        }
    }

    public function getWeight(): int
    {
        return 0;
    }
}

Now, let's inform the hydrator about the new type:

$hydrator->addTypeConverter(new UriTypeConverter());

Ignored property

If you need a property to be ignored and not populated during the object hydration process, use a special annotation like the example below:

#[\Sunrise\Hydrator\Annotation\Ignore]
public string $value;

Property alias

If you need to handle an unnormalized key in a dataset or have other reasons to associate such a key with a property that has a different name, you can use a special annotation for this purpose:

#[\Sunrise\Hydrator\Annotation\Alias('error-codes')]
public array $errorCodes = [];

Error handling

try {
    $hydrator->hydrate(...);
} catch (\Sunrise\Hydrator\Exception\InvalidDataException $e) {
    // It's runtime error
} catch (\Sunrise\Hydrator\Exception\InvalidObjectException $e) {
    // It's logic error
}

The InvalidDataException exception contains errors related to an input dataset and is designed to display errors directly on the client side.

If you are using the symfony/validator package, you may find it useful to present the errors as a \Symfony\Component\Validator\ConstraintViolationListInterface. To achieve this, you can call the following method on this exception:

try {
    $hydrator->hydrate(...);
} catch (\Sunrise\Hydrator\Exception\InvalidDataException $e) {
    $violations = $e->getViolations();
}

Or you can retrieve the list of errors in the standard way, as shown in the example below:

try {
    $hydrator->hydrate(...);
} catch (\Sunrise\Hydrator\Exception\InvalidDataException $e) {
    $errors = $e->getExceptions();
    foreach ($errors as $error) {
        echo $error->getMessage(), PHP_EOL;
        echo $error->getPropertyPath(), PHP_EOL;
        echo $error->getErrorCode(), PHP_EOL;
    }
}

Localization

$localizedMessages = [
    // using the error code dictionary...
    \Sunrise\Hydrator\Dictionary\ErrorCode::INVALID_TIMESTAMP => 'Это значение не является валидной временной меткой; ожидаемый формат: {{ expected_format }}.',

    // or using the error message dictionary...
    \Sunrise\Hydrator\Dictionary\ErrorMessage::INVALID_TIMESTAMP => 'Это значение не является валидной временной меткой; ожидаемый формат: {{ expected_format }}.',
];

try {
    $hydrator->hydrate(...);
} catch (\Sunrise\Hydrator\Exception\InvalidDataException $e) {
    foreach ($e->getExceptions() as $error) {
        // original message...
        $message = $error->getMessage();

        // using the error code dictionary...
        if (isset($localizedMessages[$error->getErrorCode()])) {
            // localized message
            $message = \strtr($localizedMessages[$error->getErrorCode()], $error->getMessagePlaceholders()), PHP_EOL;
        }

        // or using the error message dictionary...
        if (isset($localizedMessages[$error->getMessageTemplate()])) {
            // localized message
            $message = \strtr($localizedMessages[$error->getMessageTemplate()], $error->getMessagePlaceholders()), PHP_EOL;
        }
    }
}

Doctrine annotations

To use annotations, you need to install the doctrine/annotations package:

composer require doctrine/annotations

To use annotations in PHP 7 or explicitly in PHP 8, you can use the following method of the hydrator:

$hydrator->useDefaultAnnotationReader();

If you need to provide your instance of an annotation reader, you can use the following method:

$hydrator->setAnnotationReader(...);

Test run

composer test