apioo/psx-data

Data processing library to read and write POPOs in different formats

Data

Data processing library which helps to read and write data to and from POPOs in different formats.

Usage

The following example showcases how you could read/write a complex model.

// create processor
$processor = new Processor(Configuration::createDefault());

// example json data which we want to parse in our model
$in = <<<JSON
{
    "id": 1,
    "title": "Lorem ipsum",
    "author": {
        "id": 1,
        "name": "Foo",
        "email": "[email protected]",
    },
    "comments": [{
        "id": 1,
        "author": {
            "id": 1,
            "name": "Foo",
            "email": "[email protected]",
        },
        "text": "Lorem ipsum"
    },{
        "id": 2,
        "author": {
            "id": 1,
            "name": "Foo",
            "email": "[email protected]",
        },
        "text": "Lorem ipsum"
    }],
    "date": "2016-03-28T22:40:00Z"
}
JSON;

// reads the json data into a custom model class
$model = $processor->read(News::class, Payload::json($in));

// the model can be used to get or set data
$model->getAuthor()->getName();
$model->getComments()[0]->getText();

// writes the model back to json
$out = $processor->write(Payload::json($model));

// model classes
class News
{
    private ?int $id = null;
    private ?string $title = null;
    protected ?Author $author = null;
    /**
     * @var array<Comment>|null
     */
    private ?array $comments = null;
     #[Format('date-time')]
    private ?string $date;

    // getter/setter implementations removed for readability
}

class Author
{
    private ?int $id = null;
    private ?string $name = null;
    private ?string $email = null;

    // getter/setter implementations removed for readability
}

class Comment
{
    private ?int $id = null;
    private ?Author $author = null;
    private ?string $text = null;

    // getter/setter implementations removed for readability
}

Formats

The library supports different reader and writer classes to produce different data formats. If you want to read a specific format you can provide the content type of the data. I.e. if you want read XML you could use the following payload:

$payload = Payload::create($in, 'application/xml');

The processor uses a reader factory to obtain the fitting reader for a specific content type. In this case it would use the XML reader. The reader factory can be easily extended with different reader classes to support other data formats.

$configuration->getReaderFactory()->addReader(new Acme\Reader(), 32);

In order to produce a payload from an incoming HTTP request you simply have to set the body as data and the content type from the header. How you access this data depends on the HTTP interface. For an PSR-7 request you could use:

$payload = Payload::create(
    (string) $request->getBody(),
    $request->getHeaderLine('Content-Type')
);

On the other hand if you want to write data as response you would use:

$payload = Payload::create(
    $model,
    $request->getHeaderLine('Accept')
);

The writer factory can also be extended with custom writer implementations.

$configuration->getWriterFactory()->addWriter(new Acme\Writer(), 64);

Constraints

It is also possible to add specific constraints to your model class. In the following some examples:

#[Required(['title'])]
class News
{
    #[Pattern('[A-z]')]
    private ?string $title = null;

     #[MinLength(3)]
     #[MaxLength(255)]
    private ?string $text = null;

    #[Enum(['active', 'deleted'])]
    private ?string $status = null;

    #[Minimum(0)]
    #[Maximum(5)]
    private ?int $rating = null;
}

All available attributes are located at the psx/schema project.

Transformations

Each reader class returns the data in a form which can be easily processed. I.e. the json reader returns a stdClass produced by json_decode and the xml reader returns a DOMDocument. To unify the output we use transformation classes which take the output of a reader and return a normalized format. I.e. for xml content we apply by default the XmlArray transformer which transforms the DOMDocument. So you can use a transformer if you directly want to work with the output of the reader.

In case you want to validate incoming XML data after a XSD schema you could use the XmlValidator transformer:

$payload = Payload::xml($data);
$payload->setTransformer(new XmlValidator('path/to/schema.xsd'));

$model = $processor->read(News::class, $payload);

Exporter

If you write data you can set as payload an arbitrary object. We use an exporter class to return the actual data representation of that object. By default the exporter reads also the psx/schema attributes so you can use the same model for incoming and outgoing data. But it is also possible to use different classes. I.e. you could create model classes using the psx/schema attributes only for incoming data and for outgoing data you could use the JMS exporter in case you have already objects which have these annotations.

If you have another way how to extract data of an object (i.e. a toArray method which returns the available fields of the object) you can easily write a custom exporter.

psx-data's Issues

Bug: Jsonp callbackName can be longer than 32 chars

The jsonp writer accepts only a callback name which is shorter than 33 chars.
But with jQuery 2.1.3 the callback can be longer (ex. jQuery213040548181975297726_1486482323181).

Why is there the limit set to 32?

        if (preg_match('/^([A-Za-z0-9._]{3,32})$/', $callbackName)) {
            $this->callbackName = $callbackName;
        }

Please see in this file:

psx-data/src/Writer/Jsonp.php

Line 66 in f3b3ead

if (preg_match('/^([A-Za-z0-9._]{3,32})$/', $callbackName)) {