fakemock: A PHP repository from 1on

Fakemock

Faker is an amazing tool for mocking things but has a one drawback — you have to do much of work in order to map all things you need. Especially when you are working with DTOs/Entities and they already have some assertions configured — the dev has to create very own rules from scratch.

This library solves that problem. I have introduced a FakeMock library that takes care of filling up as many objects as you need.

ToC

Install
Quick example
Configuration
Populating multiple objects
Groups
phpDoc
Supported DTOs
Asserts
Supported asserts
Internal architecture
Advanced concepts

Install

Install the library:

composer require er1z/fakemock

Quick example

We assume all the autoloaders stuff is configured so create (or re-use) your DTO:

use Er1z\FakeMock\Annotations\FakeMock as FakeMock;
use Er1z\FakeMock\Annotations\FakeMockField as FakeMockField;

/**
 * @FakeMock()
 */
class MyDto {
    
    /**
     * @FakeMockField()
     */
    public $username;
    
}

Now — fill up above with some random data:

$fakemock = new Er1z\FakeMock\FakeMock();

$dto = new MyDto();
$fakemock->fill($dto);

echo $dto->username;   // mr.handsome

What's happened — name guesser is used here so it assumed that $username may contain your user's login. But guessing not always would fit your needs. It's possible to specify any Faker's method to fill it with random data:

/**
 * @FakeMockField("name")
 */
public $username;

and we end up with generated some random first and last name.

Configuration

Most part of behavior is controlled via annotations. We can specify two types of configuration: global (object-scope) and local (property-scope). All available properties for global scope:

type	name	default value	description
`bool`	`satisfyAssertsConditions`	`true`	enables/disables asserts decorator (see: supported asserts)
`bool`	`useAsserts`	`true`	should FakeMock use assertions to generate data?
`array`	`classMappings`	`[]`	specifies a dictionary of `SomeClass=>FQCN` for mapping interfaces within this field
`string`\|`null`	`locale`	`null`	locale to use with Faker

Local scope:

type	name	default value	description
`null`\|`array`	`arguments`	`null`	an array of arguments for Faker method
`null`\|`string`	`faker`	`null`	specify desired faker method. Set to null if you want to generator chain do it's best on guessing
`null`\|`array\|string`	`groups`	`null`	validation groups this rule for this rule is being processed.
`null`\|`string`	`regex`	`null`	a regular expression to generate random data against
`null`\|`bool`	`satisfyAssertConditions`	`null`	turns off or on assertion decorator — `null` inherits value from global configuration
`null`\|`bool`	`useAsserts`	`null`	should FakeMock use validation rules to generate? If `null`, value is inherited from global configuration
`null`\|`mixed`	`value`	`null`	literal value on field. Stops guessing
`null`\|`string`	`mapToClass`	`null`	FQCN of class the object should be instantiated as
`string`\|`null`	`locale`	`null`	locale to use with Faker

Local scope configuration constructor has a possibility to create an annotation from string-argument which is populated to faker key.

Populating multiple objects

Developers are lazy so am I — you have to take care of things you really need to. So let's populate a few objects:

$fakemock = new FakeMock();

$results = [];

for( $a=0; $a<1000; $a++ )
{
    $results[] = $fakemock->fill(MyDto::class);
}

That's all. They all are fresh instances so don't be concerned any references.

Groups

Sometimes it's needed to populate objects conditionally. Let's try with populating every 3rd generated object. First, declare a group of field:

use Er1z\FakeMock\Annotations\FakeMock as FakeMock;
use Er1z\FakeMock\Annotations\FakeMockField as FakeMockField;

/**
 * @FakeMock()
 */
class GroupedDto {

    /**
     * @FakeMockField(groups={"first"})
     */
    public $field;
    
}

Generate:

$fakemock = new FakeMock();

$results = [];

for( $a=0; $a<1000; $a++ )
{
    $results[] = $fakemock->fill(MyDto::class, $a%3==0 ?? 'first');
}

Now, check your results. This behavior is similar to Symfony's validation groups.

phpDoc

If no guess is possible and you haven't mapped any particular Faker's type, FakeMock tries to guess type according to phpDoc variable type:

use Er1z\FakeMock\Annotations\FakeMock as FakeMock;
use Er1z\FakeMock\Annotations\FakeMockField as FakeMockField;

/**
 * @FakeMock()
 */
class DocDTO {
    
    /**
     * @var float
     * @FakeMockField()
     */
    public $field;
    
}

$f = new FakeMock();
$obj = new DocDTO();

$data = $f->fill($obj);

var_dump($data->field); // eg. 1.24422

Supported DTOs

FakeMock relies on PropertyAccess component so different kinds of DTOs are supported, even Doctrine entities. You can leave an object with exposed public fields but also encapsulate data via setters and getters:

use Er1z\FakeMock\Annotations\FakeMock as FakeMock;
use Er1z\FakeMock\Annotations\FakeMockField as FakeMockField;

/**
 * @FakeMock()
 */
class EncapsulatedDTO
{
    /**
     * @FakeMockField();
     */
    protected $field;
    
    public function getField()
    {
        return $this->field;
    }
    
    public function setField($field)
    {
        $this->field = $field;
    }
    
}

And this will just work.

Asserts

If your project is using Symfony's validate component, it's possible to utilize validation rules to tell the FakeMock how generate fields contents. For example:

use Er1z\FakeMock\Annotations\FakeMock as FakeMock;
use Er1z\FakeMock\Annotations\FakeMockField as FakeMockField;
use Symfony\Component\Validator\Constraints as Assert;

/**
 * @FakeMock()
 */
class ValidatedDTO {
    
    /**
     * @FakeMockField()
     * @Assert\Email()
     */
    public $email;
    
}

and calling fill method against this object will produce fake e-mail address on $email field.

Supported asserts

Generators:

Assert	Support status	Description
All	unsupported
Bic	supported	Swift BIC code
Blank	supported	nullifies field
Callback	unsupported
CardScheme	partial-support	generates a valid credit card number — currently only `visa`, `mastercard` and `amex` are supported
Choice	supported	returns a value based on choices — supported are both single and multiple
Collection	unsupported
Count	unsupported
Country	supported	generates a country code
Currency	supported	currency code
Date	partial-supported	generates a date, currently according to `DATE_ATOM` format
DateTime	partial-supported	generates a date with time, currently: `ISO` format
Email	supported	generates a valid e-mail address
Expression	unsupported
File	unsupported
Iban	supported	generates an IBAN account number
Image	supported	generates a `SplFileObject` with image of specified dimensions
Ip	supported	generates an IPv4/v6 address
Isbn	supported	generates an ISBN
IsFalse	supported	returns false to field
IsNull	supported	nullifies field
Issn	partial-support	only single hard-coded ISSN number
IsTrue	supported	returns true to field
Language	supported	generates a string with language code
Locale	supported	generates a locale string
Luhn	partial-support	generates a credit card number for LUHN algorithm validation
NotBlank	supported	checks whether field is not empty
NotNull	supported	field is not null
Regex	supported	generates a string that matches pattern
Time	partial-supported	generates a time with `H:i:s` format
Type	unsupported
UniqueEntity	unsupported
Url	supported	generates a valid URL
UserPassword	unsupported
Uuid	supported	generates an UUID
Valid	unsupported

Decorators/conditionals:

Assert	support	description
EqualTo	supported	equalizes value to specified literal or `propertyPath` field value
GreaterThan	supported	checks if generated value matches lower boundary (not inclusive)
GreaterThanOrEqual	supported	checks if generated value matches lower boundary (inclusive)
IdenticalTo	partial-support	similar to `EqualTo` but also checks type, currently: alias to `EqualTo`
Length	supported	generates a string between min-max length boundary
LessThan	supported	checks if generated value matches upper boundary (not inclusive)
LessThanOrEqual	supported	checks if generated value matches upper boundary (inclusive)
NotEqualTo	supported	makes sure that field's value is different from literal or `propertyPath` field value
NotIdenticalTo	partial-support	similar to `NotEqualTo` but also checks type, currently: alias to `NotEqualTo`
Range	supported	generates a number within specified boundaries

FakeMock is smart enough to guess what you want to get — asserts are also decorated against specified phpDoc type, for example if you specify LessThan constraint and @var float, you get float value and so on. This feature is useful when you need a DateTimeInterface in string:

/**
 * @FakeMock()
 * @Assert\DateTime()
 * @var string
 */

Internal architecture

FakeMock is a library with fair amount of tests so you don't need to bother if you want to make a contribution and concerned your code will mess up something.

Modular architecture allows to enhance and extend functionality. The main entrypoint is a FakeMock class which needs three elements:

Metadata\FactoryInterface — builds some information on fields metadata, eg. if it should be processed, what rules are specified and so on,
GeneratorChainInterface — maintains a list of field-data generators,
DecoratorChainInterface — holds a list of decorators which can be used to modify generated value according to various rules, eg. convert DateTimeInterface to string.

Almost all modules could be overrided thanks to passing the dependencies mostly by interfaces or constructor arguments. So feel free to play with all components.

Generating of data step-by-step:

Create a FakeMock instance with specified object/FQCN to generate (if FQCN, instantiate silently),
Pass object to Metadata\FactoryInterface in order to get main object configuration,
If object is configured, iterate over object properties, create FieldMetadata merging some configuration variables with object configuration and check if group is specified and if it should be processed,
Tell GeneratorChainInterface to getValueForField. Available adapters are executed one-by-one until one of them returns non-null value,
Run DecoratorChainInterface with getDecoratedValue — mostly, they are all ran one-by-one except currently processed returns false which breaks the chain,
Set generated value by accessor.

Default generator chain:

TypedGenerator — handles two cases: value or regex. Nothing less, nothing more,
RecursiveGenerator — if variable class has FQCN specified in phpDoc, it's processed unless recursive field flag is set to false,
If package symfony/validator is installed and available, AssertGenerator is being checked against,
FakerGenerator — provides methods for generating specified Faker's generator or guess field content by NameGuesser,
PhpDocGenerator — generates data according to the property type,
LastResortGenerator — if everything above fails, generates simple name as string.

Default decorators chain:

AssertDecorator — restrict values to validation rules — its behavior is controlled by satisfyAssertsConditions field configuration,
PhpDocDecorator — converts values types.

Advanced Concepts

This is a „skeleton” of the steps required to do something more complicated within this library. For example, we want to use mapped interfaces/abstract on DTOs. Assume structure:

interface SomeDTOInterface {
    
}

use Er1z\FakeMock\Annotations\FakeMock as FakeMock;
use Er1z\FakeMock\Annotations\FakeMockField as FakeMockField;

/**
 * @FakeMock()
 */
class InnerDTO implements SomeDTOInterface {
    
    /**
     * @FakeMockField()
     */
    public $field;
    
}

use Er1z\FakeMock\Annotations\FakeMock as FakeMock;
use Er1z\FakeMock\Annotations\FakeMockField as FakeMockField;

/**
 * @FakeMock()
 */
class MainDTO {
    
    /**
     * @FakeMockField()
     * @var SomeDTOInterface
     */
    public $nested;
    
}

Running basic FakeMock scenario will produce nothing — $nested is null. We have to tell the library, what object should we map to desired interface.

$generators = GeneratorChain::getDefaultGeneratorsSet();

foreach($generators as $g)
{
    if( $g instanceof RecursiveGenerator::class )
    {
        $g->addClassMapping(SomeDTOInterface::class, InnerDTO::class);
    }
}

$generatorChain = new GeneratorChain($generators);

$fakemock = new FakeMock(null, $generatorChain);

$mainDto = new MainDto();
$result = $fakemock->fill($mainDto);

Of course, you also can map interfaces using annotations on the global and/or local scope:

use Er1z\FakeMock\Annotations\FakeMock as FakeMock;
use Er1z\FakeMock\Annotations\FakeMockField as FakeMockField;

/**
 * @FakeMock(classMappings={"Namespace\SomeDTOInterface"=>"Some\Other\Class"})
 */
class MainDTO {
    
    /**
     * @FakeMockField()
     * @var SomeDTOInterface
     */
    public $nested;
    
    /**
     * @FakeMockField("mapClass"="Some\Other\Class")
     * @var SomeDTOInterface
     */
    public $secondNested;
    
}

Changelog

0.3.1

Fixed: value retrieval to PropertyInfo-based method

0.3

New: option to specify Faker's locale and Faker Generator Registry

0.2.1

Fixed: correct asserts group processing

0.2

New: recursive fields processing

0.1

First public version

1on/fakemock