High definition PHP structures with JSON-schema based validation.
Supported schemas:
composer require swaggest/json-schema
Structure definition can be done either with json-schema
or with
PHP
class extending Swaggest\JsonSchema\Structure\ClassStructure
Define your json-schema
$schemaJson = <<<'JSON'
{
"type": "object",
"properties": {
"id": {
"type": "integer"
},
"name": {
"type": "string"
},
"orders": {
"type": "array",
"items": {
"$ref": "#/definitions/order"
}
}
},
"required":["id"],
"definitions": {
"order": {
"type": "object",
"properties": {
"id": {
"type": "integer"
},
"price": {
"type": "number"
},
"updated": {
"type": "string",
"format": "date-time"
}
},
"required":["id"]
}
}
}
JSON;
Load it
$schema = Schema::import(json_decode($schemaJson));
Validate data
$schema->in(json_decode(<<<'JSON'
{
"id": 1,
"name":"John Doe",
"orders":[
{
"id":1
},
{
"price":1.0
}
]
}
JSON
)); // Exception: Required property missing: id at #->properties:orders->items[1]->#/definitions/order
You can also call Schema::import
on string uri
to schema json data.
$schema = Schema::import('http://localhost:1234/my_schema.json');
Or with boolean argument.
$schema = Schema::import(true); // permissive schema, always validates
$schema = Schema::import(false); // restrictive schema, always invalidates
With complex schemas it may be hard to find out what's wrong with your data. Exception message can look like:
No valid results for oneOf {
0: Enum failed, enum: ["a"], data: "f" at #->properties:root->patternProperties[^[a-zA-Z0-9_]+$]:zoo->oneOf[0]
1: Enum failed, enum: ["b"], data: "f" at #->properties:root->patternProperties[^[a-zA-Z0-9_]+$]:zoo->oneOf[1]
2: No valid results for anyOf {
0: Enum failed, enum: ["c"], data: "f" at #->properties:root->patternProperties[^[a-zA-Z0-9_]+$]:zoo->oneOf[2]->$ref[#/cde]->anyOf[0]
1: Enum failed, enum: ["d"], data: "f" at #->properties:root->patternProperties[^[a-zA-Z0-9_]+$]:zoo->oneOf[2]->$ref[#/cde]->anyOf[1]
2: Enum failed, enum: ["e"], data: "f" at #->properties:root->patternProperties[^[a-zA-Z0-9_]+$]:zoo->oneOf[2]->$ref[#/cde]->anyOf[2]
} at #->properties:root->patternProperties[^[a-zA-Z0-9_]+$]:zoo->oneOf[2]->$ref[#/cde]
} at #->properties:root->patternProperties[^[a-zA-Z0-9_]+$]:zoo
For ambiguous schemas defined with oneOf
/anyOf
message is indented multi-line string.
Processing path is a combination of schema and data pointers. You can use InvalidValue->getSchemaPointer()
and InvalidValue->getDataPointer()
to extract schema/data pointer.
You can build error tree using InvalidValue->inspect()
.
/**
* @property int $quantity PHPDoc defined dynamic properties will be validated on every set
*/
class User extends ClassStructure
{
/* Native (public) properties will be validated only on import and export of structure data */
/** @var int */
public $id;
public $name;
/** @var Order[] */
public $orders;
/** @var UserInfo */
public $info;
/**
* @param Properties|static $properties
* @param Schema $ownerSchema
*/
public static function setUpProperties($properties, Schema $ownerSchema)
{
// You can add custom meta to your schema
$dbTable = new DbTable;
$dbTable->tableName = 'users';
$ownerSchema->addMeta($dbTable);
// Setup property schemas
$properties->id = Schema::integer();
$properties->id->addMeta(new DbId($dbTable)); // You can add meta to property.
$properties->name = Schema::string();
// You can embed structures to main level with nested schemas
$properties->info = UserInfo::schema()->nested();
// You can set default value for property
$defaultOptions = new UserOptions();
$defaultOptions->autoLogin = true;
$defaultOptions->groupName = 'guest';
// UserOptions::schema() is safe to change as it is protected with lazy cloning
$properties->options = UserOptions::schema()->setDefault(UserOptions::export($defaultOptions));
// Dynamic (phpdoc-defined) properties can be used as well
$properties->quantity = Schema::integer();
$properties->quantity->minimum = 0;
// Property can be any complex structure
$properties->orders = Schema::create();
$properties->orders->items = Order::schema();
$ownerSchema->required = array(self::names()->id);
}
}
class UserInfo extends ClassStructure {
public $firstName;
public $lastName;
public $birthDay;
/**
* @param Properties|static $properties
* @param Schema $ownerSchema
*/
public static function setUpProperties($properties, Schema $ownerSchema)
{
$properties->firstName = Schema::string();
$properties->lastName = Schema::string();
$properties->birthDay = Schema::string();
}
}
class UserOptions extends ClassStructure
{
public $autoLogin;
public $groupName;
/**
* @param Properties|static $properties
* @param Schema $ownerSchema
*/
public static function setUpProperties($properties, Schema $ownerSchema)
{
$properties->autoLogin = Schema::boolean();
$properties->groupName = Schema::string();
}
}
class Order implements ClassStructureContract
{
use ClassStructureTrait; // You can use trait if you can't/don't want to extend ClassStructure
const FANCY_MAPPING = 'fAnCy'; // You can create additional mapping namespace
public $id;
public $userId;
public $dateTime;
public $price;
/**
* @param Properties|static $properties
* @param Schema $ownerSchema
*/
public static function setUpProperties($properties, Schema $ownerSchema)
{
// Add some meta data to your schema
$dbMeta = new DbTable();
$dbMeta->tableName = 'orders';
$ownerSchema->addMeta($dbMeta);
// Define properties
$properties->id = Schema::integer();
$properties->userId = User::properties()->id; // referencing property of another schema keeps meta
$properties->dateTime = Schema::string();
$properties->dateTime->format = Format::DATE_TIME;
$properties->price = Schema::number();
$ownerSchema->required[] = self::names()->id;
// Define default mapping if any
$ownerSchema->addPropertyMapping('date_time', Order::names()->dateTime);
// Define additional mapping
$ownerSchema->addPropertyMapping('DaTe_TiMe', Order::names()->dateTime, self::FANCY_MAPPING);
$ownerSchema->addPropertyMapping('Id', Order::names()->id, self::FANCY_MAPPING);
$ownerSchema->addPropertyMapping('PrIcE', Order::names()->price, self::FANCY_MAPPING);
}
}
Validation of dynamic properties is performed on set, this can help to find source of invalid data at cost of some performance drop
$user = new User();
$user->quantity = -1; // Exception: Value more than 0 expected, -1 received
Validation of native properties is performed only on import/export
$user = new User();
$user->quantity = 10;
User::export($user); // Exception: Required property missing: id
Error messages provide a path to invalid data
$user = new User();
$user->id = 1;
$user->name = 'John Doe';
$order = new Order();
$order->dateTime = (new \DateTime())->format(DATE_RFC3339);
$user->orders[] = $order;
User::export($user); // Exception: Required property missing: id at #->properties:orders->items[0]
Nested structures allow you to make composition: flatten several objects in one and separate back.
$user = new User();
$user->id = 1;
$info = new UserInfo();
$info->firstName = 'John';
$info->lastName = 'Doe';
$info->birthDay = '1970-01-01';
$user->info = $info;
$json = <<<JSON
{
"id": 1,
"firstName": "John",
"lastName": "Doe",
"birthDay": "1970-01-01"
}
JSON;
$exported = User::export($user);
$this->assertSame($json, json_encode($exported, JSON_PRETTY_PRINT));
$imported = User::import(json_decode($json));
$this->assertSame('John', $imported->info->firstName);
$this->assertSame('Doe', $imported->info->lastName);
You can also use \Swaggest\JsonSchema\Structure\Composition
to dynamically create schema compositions.
This can be helpful to deal with results of database query on joined data.
$schema = new Composition(UserInfo::schema(), Order::schema());
$json = <<<JSON
{
"id": 1,
"firstName": "John",
"lastName": "Doe",
"price": 2.66
}
JSON;
$object = $schema->import(json_decode($json));
// Get particular object with `pick` accessor
$info = UserInfo::pick($object);
$order = Order::pick($object);
// Data is imported objects of according classes
$this->assertTrue($order instanceof Order);
$this->assertTrue($info instanceof UserInfo);
$this->assertSame(1, $order->id);
$this->assertSame('John', $info->firstName);
$this->assertSame('Doe', $info->lastName);
$this->assertSame(2.66, $order->price);
If property names of PHP objects should be different from raw data you
can call ->addPropertyMapping
on owner schema.
// Define default mapping if any
$ownerSchema->addPropertyMapping('date_time', Order::names()->dateTime);
// Define additional mapping
$ownerSchema->addPropertyMapping('DaTe_TiMe', Order::names()->dateTime, self::FANCY_MAPPING);
$ownerSchema->addPropertyMapping('Id', Order::names()->id, self::FANCY_MAPPING);
$ownerSchema->addPropertyMapping('PrIcE', Order::names()->price, self::FANCY_MAPPING);
It will affect data mapping:
$order = new Order();
$order->id = 1;
$order->dateTime = '2015-10-28T07:28:00Z';
$order->price = 2.2;
$exported = Order::export($order);
$json = <<<JSON
{
"id": 1,
"date_time": "2015-10-28T07:28:00Z",
"price": 2.2
}
JSON;
$this->assertSame($json, json_encode($exported, JSON_PRETTY_PRINT));
$imported = Order::import(json_decode($json));
$this->assertSame('2015-10-28T07:28:00Z', $imported->dateTime);
You can have multiple mapping namespaces, controlling with mapping
property of Context
$options = new Context();
$options->mapping = Order::FANCY_MAPPING;
$exported = Order::export($order, $options);
$json = <<<JSON
{
"Id": 1,
"DaTe_TiMe": "2015-10-28T07:28:00Z",
"PrIcE": 2.2
}
JSON;
$this->assertSame($json, json_encode($exported, JSON_PRETTY_PRINT));
$imported = Order::import(json_decode($json), $options);
$this->assertSame('2015-10-28T07:28:00Z', $imported->dateTime);
You can create your own pre-processor implementing Swaggest\JsonSchema\DataPreProcessor
.
Meta
is a way to complement Schema
with your own data. You can keep and retrieve it.
You can store it.
$dbMeta = new DbTable();
$dbMeta->tableName = 'orders';
$ownerSchema->addMeta($dbMeta);
And get back.
// Retrieving meta
$dbTable = DbTable::get(Order::schema());
$this->assertSame('orders', $dbTable->tableName);
If you want to tolerate invalid data or improve mapping performance you can specify skipValidation
flag in processing Context
$schema = Schema::object();
$schema->setProperty('one', Schema::integer());
$schema->properties->one->minimum = 5;
$options = new Context();
$options->skipValidation = true;
$res = $schema->in(json_decode('{"one":4}'), $options);
$this->assertSame(4, $res->one);
If you want to map data to a different class you can register mapping at top level of your importer structure.
class CustomSwaggerSchema extends SwaggerSchema
{
public static function import($data, Context $options = null)
{
if ($options === null) {
$options = new Context();
}
$options->objectItemClassMapping[Schema::className()] = CustomSchema::className();
return parent::import($data, $options);
}
}
Or specify it in processing context
$context = new Context();
$context->objectItemClassMapping[Schema::className()] = CustomSchema::className();
$schema = SwaggerSchema::schema()->in(json_decode(
file_get_contents(__DIR__ . '/../../../../spec/petstore-swagger.json')
), $context);
$this->assertInstanceOf(CustomSchema::className(), $schema->definitions['User']);
Some code quality best practices are deliberately violated here (see ) to allow best performance at maintenance cost.
Those violations are secured by comprehensive test coverage:
- draft-04, draft-06, draft-07 of JSON-Schema-Test-Suite
- test cases (excluding
$data
and few tests) of epoberezkin/ajv (a mature js implementation)