Parse and generate data schema formats

v2.6.2 2020-03-09 20:04 UTC



This library helps to generate PHP classes from JsonSchema definitions and vice versa. You can then use those POPOs to import and validate json data. It was inspired by JAXB where it is possible generate POJOs from XSD files. We have also created an online tool to test the JsonSchema to PHP conversion.


This example shows how to transform data into an object based on a jsonschema. We use the example schema from the json schema website.

  "title": "Example Schema",
  "type": "object",
  "properties": {
    "firstName": {
      "type": "string"
    "lastName": {
      "type": "string"
    "age": {
      "description": "Age in years",
      "type": "integer",
      "minimum": 0
  "required": [

Based on the schema we can generate the PHP classes with the following command:
vendor/bin/schema schema:parse --format=php schema.json


namespace PSX\Generation;

 * @Title("Example Schema")
 * @Required({"firstName", "lastName"})
class Example_Schema
     * @Key("firstName")
     * @Type("string")
    public $firstName;
     * @Key("lastName")
     * @Type("string")
    public $lastName;
     * @Key("age")
     * @Description("Age in years")
     * @Type("integer")
     * @Minimum(0)
    public $age;
    public function setFirstName($firstName)
        $this->firstName = $firstName;
    public function getFirstName()
        return $this->firstName;
    public function setLastName($lastName)
        $this->lastName = $lastName;
    public function getLastName()
        return $this->lastName;
    public function setAge($age)
        $this->age = $age;
    public function getAge()
        return $this->age;

Now we can use this class to read json data which fits to the schema

// the data which we want to import
$data = json_decode({"firstName": "foo", "lastName": "bar"});

$schemaManager = new \PSX\Schema\SchemaManager();

// we read the schema annotations from the class
$schema = $schemaManager->getSchema(Example_Schema::class);

try {
    $traverser = new SchemaTraverser();
    $example   = $traverser->traverse($data, $schema, new TypeVisitor());
    // $example contains now an instance of the Example_Schema class containing 
    // the firstName and lastName property
    echo $example->getFirstName();

} catch (\PSX\Schema\ValidationException $e) {
    // the validation failed
    echo $e->getMessage();

You can use the dumper to create a JSON representation based on your POPO objects:

$schema = new Example_Schema();

$dumper = new \PSX\Schema\Parser\Popo\Dumper();
$data   = $dumper->dump($schema);

echo json_encode($data);

// would result in
// {"firstName": "foo", "lastName": "bar", "age": 12}

Please note that we use the doctrine/annotations library to parse the annotations. Because of this you need to setup a fitting annotation loader which can load the classes. To use the registered autoloader you can simply use: AnnotationRegistry::registerLoader('class_exists'). More information how to configure the loader at the documentation.


The following annotations are available:

Annotation Target Example
@AdditionalItems Property @AdditionalItems(true)
@AdditionalProperties Class @AdditionalProperties(true)
@AllOf Property @AllOf(@Ref("FooClass"), @Ref("BarClass"))
@AnyOf Property @AnyOf(@Schema(type="integer", minimum=0), @Schema(type="string", maxLength=64))
@Dependencies Class @Dependencies(property="name", value={"age", "gender"})
@Description Class/Property @Description("content")
@Enum Property @Enum({"foo", "bar"})
@Exclude Property @Exclude
@ExclusiveMaximum Property @ExclusiveMaximum(true)
@ExclusiveMinimum Property @ExclusiveMinimum(true)
@Format Property @Format("uri")
@Items Property @Items(@Ref("FooClass"))
@Key Property @Key("$ref")
@Maximum Property @Maximum(16)
@MaxItems Property @MaxItems(16)
@MaxLength Property @MaxLength(16)
@MaxProperties Class @MaxProperties(16)
@Minimum Property @Minimum(4)
@MinItems Property @MinItems(4)
@MinLength Property @MinLength(4)
@MinProperties Property @MinProperties(4)
@MultipleOf Property @MultipleOf(2)
@Not Property @Not(@Schema(type="string"))
@OneOf Property @OneOf(@Ref("FooClass"), @Ref("BarClass"))
@Pattern Property @Pattern("A-z+")
@PatternProperties Class @PatternProperties(pattern="^foo", type="string")
@Ref Property @Ref("FooClass")
@Required Class @Required({"name", "title"})
@Title Class @Title("foo")
@Type Property @Type("string")
@UniqueItems Property @UniqueItems(true)



The following parser classes are available which produce a schema object:

  • JsonSchema (Parses a JsonSchema file)
  • Popo (Parses the annotations of a class)


The following generator classes are available which generate a representation of a schema object:

  • HTML (Generates a HTML representation of the schema)
  • JsonSchema (Generates a JsonSchema specification)
  • Markdown (Generates a markdown representation of the schema)
  • PHP (Generates PHP classes representing the schema using annotations)
  • Protobuf (Generates a proto3 definition)
  • Typescript (Generates a TypeScript interface)

Code generation

This library tries to generate useful PHP classes based on a defined JsonSchema. That being said the main goal of JsonSchema is to validate JSON data and not to model a JSON structure. Because of this nature it is difficult to generate a proper class hierarchy using inheritance based on the validation keywords (@OneOf, @AllOf, etc). To avoid these complications we use the following simple logic to generate a class:

If a JsonSchema has a properties keyword we generate a class which contains the properties which are defined. Each property contains annotations which describe the defined validation keywords. If the property reference another schema we use the @Ref annotation otherwise we use the inline @Schema annotation.

So developers should see the generated classes only as a starting point. If inheritance is needed it must be manually implemented. Despite that this approach is pretty robust and allows us to also generate classes for recursive schemas like i.e. the JsonSchema spec itself.

JSON Schema code

While developing this library we have experienced many pitfalls with the JsonSchema specification. In order to avoid those and to help to write better JsonSchema schemas we have developed a JsonSchema code specification which restricts the JsonSchema keywords. This specification contains also a JsonSchema meta schema to validate a schema against those rules.