franzose / kontrolio
Simple and standalone data validation library inspired by Laravel and Symfony
Requires
- php: >=8.1
- ext-ctype: *
- ext-mbstring: *
Requires (Dev)
- phpunit/phpunit: ^9.5
README
Kontrolio: keep your data under control.
Kontrolio is a simple standalone data validation library inspired by Laravel and Symfony and compatible with PHP 8.1+. [ Read on Medium ]
Setting up validator
The best way to set up validator:
// In container unaware environments $valid = Factory::getInstance()->make($data, $rules, $messages)->validate(); // Using a service container implementation $container->singleton('validation', static fn() => new Factory()); $container->get('validation')->make($data, $rules, $messages)->validate();
Of course, you can use Kontrolio\Validator
class directly but then you'll need to provide available validation rules by yourself:
$validator = new Validator($data, $rules, $messages)->extend($custom)->validate();
Data here is supposed to be the key-value pairs—attributes and their values:
$data = [ 'foo' => 'bar', 'bar' => 'baz', 'baz' => 'taz' ];
Validation rules can be set in three different formats:
- Laravel-like string
- Instances of the class based rules
- Callables (closures or callbacks)
You can also mix instances and callables when you set multiple validation rules to a single attribute. Here's a simple example:
$rules = [ 'one' => 'not_empty|length:5,15', 'two' => new Email, 'three' => static fn ($value) => $value === 'taz', 'four' => [ static fn ($value) => is_numeric($value), new GreaterThan(5), ] ];
When you set validation rules as string, validator will parse it to an ordinary array of rules before applying them to the attribute, so when you write 'some' => 'not_empty|length:5,15'
, it becomes
'some' => [ new NotEmpty, new Length(5, 15) ]
It's pretty straightforward but remember that all arguments you pass after semicolon (separating them by commas) become the arguments of the validation rule constructor.
When you set validation rule as callback, internally it is wrapped by special object called Kontrolio\Rules\CallbackRuleWrapper
to keep consistency defined by Kontrolio\Rules\RuleInterface
interface.
Rule options
Allowing empty value
A single rule validation can be skipped when the validated attribute value is empty. If you feel that you need this option, you instantiate a rule using named constructor allowingEmptyValue()
or by calling allowEmptyValue()
method on the already existing instance:
'some' => [ MyRule::allowingEmptyValue(), // (new MyRule())->allowEmptyValue() ]
Skipping rule validation
When creating a new custom class based rule you might need an option to skip its validation based on some conditions. You can define the behavior using canSkipValidation()
method:
class MyRule extends AbstractRule { public function canSkipValidation($input = null) { return $input === 'bar'; } // ... }
Callable rules
Callable rule is nothing more than a closure or function that takes an attribute value and returns either boolean result of the validation or an options array equivalent to options provided by a class based validation rule:
'foo' => static fn ($value) => is_string($value), 'bar' => static function ($value) { return [ // required when array 'valid' => $value === 'taz', // optionals 'name' => 'baz', // rule identifier 'empty_allowed' => true, // allowing empty value 'skip' => false // don't allow skipping current rule validation, 'violations' => [] // rule violations ]; }
Custom rules
Of course you can create your custom rules. Just remember that each rule must be an instance of Kontrolio\Rules\RuleInterface
, implement isValid()
method and have an identifier. By default, identifiers are resolved by Kontrolio\Rules\AbstractRule
and are based on the rule class name without namespace. However, you can override this behavior if you wish overriding getName()
method.
Custom rules can be added eighter via factory or validator itself:
$factory = (new Factory())->extend([CustomRule::class]); // with a custom identifier $factory = (new Factory())->extend(['some_custom' => CustomRule::class]); $validator = $factory->make([], [], []); // if you don't use factory $validator = new Validator([], [], []); $validator->extend([CustomRule::class]); // with a custom identifier $validator->extend(['custom' => CustomRule::class]); $validator->validate();
Bypassing attribute's value validation completely
It's not the same as using allowEmptyValue()
or canSkipValidation()
on a rule. With those you can skip only a rule. But you can also bypass a whole attribute by using Kontrolio\Rules\Core\Sometimes
rule. Sometimes
tells validator to bypass validation when the value of the attribute is null or empty. That's all. You can prepend Sometimes
to the attribute's rules array or use its identifier in a ruleset string:
$rules = [ 'one' => 'sometimes|length:5,15', // 'one' => [ // new Sometimes(), // new Length(5, 15) // ] ];
Stopping validation on the first failure
You can tell the validator to stop validation if any validation error occurs:
$validator->shouldStopOnFailure()->validate();
Stopping certain attribute's validation on the first failure
Using UntilFirstFailure
validation rule, you can also stop validation of a single attribute while keeping validation as a whole:
$data = [ 'attr' => '', 'attr2' => 'value2' ]; $rules = [ 'attr' => [ new UntilFirstFailure(), new NotBlank(), new NotFooBar() ] ]; $messages = [<...>]; $validator = new Validator($data, $rules, $messages)->validate();
Now when attr
fail with NotBlank
rule, its validation will be stopped and validator will proceed to the attr2
.
Error messages and rule violations
Error messages has a single simple format you'll love:
$messages = [ 'foo' => 'Foo cannot be null', 'foo.length' => 'Wrong length of foo', 'foo.length.min' => 'Foo is less than 3' // '[attribute].[rule].[violation] ];
Every message key can have three segments separated by dot. They are:
- Attribute name
- Validation rule identifier
- Validation rule violation
With these options you can customize messages from the most general to the most specific. Each violation is set by the rule validating the attribute value. So when writing your own validation rule you may provide your own violations to provide customizability of the validation result and error messages.