Download the PHP package devster/alice without Composer

Alice - Expressive fixtures generator

Alice allows you to create a ton of fixtures/fake data for use while developing or testing your project. It gives you a few essential tools to make it very easy to generate complex data with constraints in a readable and easy to edit way, so that everyone on your team can tweak the fixtures if needed.

Installation

This is installable via Composer as nelmio/alice:

composer require nelmio/alice

To use it in Symfony2 you may want to use the hautelook/alice-bundle or h4cc/alice-fixtures-bundle package instead.

Table of Contents

• Usage
  • Basic Usage
  • Detailed Usage
• Reference
  • Creating Fixtures
  • Fixture Ranges
  • Faker Data
  • Calling Methods
  • Specifying Constructor Arguments
  • Optional Data
  • References
  • Multiple References
  • Handling Unique Constraints
  • Fixture Inheritance
  • Including files
  • Variables
  • Value Objects
  • Custom Faker Data Providers
  • Custom Setter
  • Complete Sample
  • Processors

Usage

Basic Usage

The easiest way to use this is to call the static Nelmio\Alice\Fixtures::load method. It will bootstrap everything for you and return you a set of persisted objects in the container you give it.

Examples:

Note: You can also pass an array of filenames if you have multiple files with references spanning more than one.

Options

Fixtures::load accepts a third $options argument that is an array with the following keys:

• locale: the default locale
• providers: an array of additional Faker providers
• seed: a seed to make sure Faker generates data consistently across runs, set to null to disable (defaults to 1)
• logger: a callable or Psr\Log\LoggerInterface object that will receive progress information during the loading of the fixtures
• persist_once: only persist objects once if multiple files are passed, by default objects are persisted after each file

Detailed Usage

If you want a bit more control you can instantiate the various object yourself and make it work just as easily:

Note: To load plain PHP files, you can use the \Nelmio\Alice\Loader\Base class instead. These PHP files must return an array containing the same structure as the yaml files have.

Reference

Creating Fixtures

The most basic functionality of this library is to turn flat yaml files into objects. You can define many objects of different classes in one file as such:

This works fine, but it is not very powerful and is completely static. You still have to do most of the work. Let's see how to make this more interesting.

Fixture Ranges

The first step is to let Alice create many copies of an object for you to remove duplication from the yaml file.

You can do that by defining a range in the fixture name:

Now it will generate ten users, with names user1 to user10. Pretty good but we only have 10 bobs with the same name, username and email, which is not so fancy yet.

You can also specify a list of values instead of a range:

To go further we can just randomize data.

Faker Data

Alice integrates with the Faker library. Using <foo()> you can call Faker data providers to generate random data. Check the list of Faker providers.

Let's turn our static bob user into a randomized entry:

As you see in the last line, you can also pass arguments to those just as if you were calling a function.

To pass Faker Data to another Faker provider, you can use the $fake() closure within faker calls. For example use $fake('firstName', 'de_DE') or $fake('numberBetween', null, 1, 200) to call Faker. Pass the provider to call followed by the locale (or null) and then the arguments to the provider.

In plain PHP fixtures the $fake closure is also available.

Localized Fake Data

Faker can create localized data for adresses, phone numbers and so on. You can set the default locale to use by passing a locale value in the $options array of Fixtures::load.

Additionally, you can mix locales by adding a locale prefix to the faker key, i.e. <fr_FR:phoneNumber()> or <de_DE:firstName()>.

Default Providers

Alice includes a default identity provider, <identity()>, that simply returns whatever is passed to it. This allows you among other things to use a PHP expression while still benefitting from variable replacement.

Some syntactic sugar is provided for this, <($whatever)> is an alias for <identity($whatever)>.

Calling Methods

Sometimes though you need to call a method to initialize some more data, you can do this just like with properties but instead using the method name and giving it an array of arguments. For example let's assume the user class has a setLocation method that requires a latitude and a longitude:

Specifying Constructor Arguments

When a constructor has mandatory arguments you must define it as explained above, for example if the User required a username in the constructor you could do the following:

If you want to call a static factory method instead of a constructor, you can specify a hash as the constructor:

If you specify false in place of constructor arguments, Alice will instantiate the object without executing the constructor:

Optional Data

Some fields do not have to be filled-in, like the favoriteNumber in this example might be personal data you don't want to share, to reflect this in our fixtures and be sure the site works and looks alright even when users don't enter a favorite number, we can make Alice fill it in sometimes using the 50%? value : empty value notation. It's a bit like the ternary operator, and you can omit the empty value if null is ok as such: 50%? value.

Let's update the user definition with this new information:

Now only half the user will have a number filled-in.

References

Let's get back to the Group. Ideally a group should have members, and Alice allows you to reference one object from another one. You can do that with the @name notation, where name is a fixture name from any class.

Let's add a fixed owner to the group:

Alice also allows you to directly reference objects' properties using the notation.

To be able to use this feature, your entities have to match some requirements :

• You can reference public properties
• You can reference properties reachable through a getter (i.e : will call if is not public)
• You can reference entities' ID but you will then have to split fixtures in multiple files (this is because objects are persisted at the end of each file processing) :

If you want to create ten users and ten groups and have each user own one group, you can use <current()> which is replaced with the current id of each iteration when using fixture ranges:

If you would like a random user instead of a fixed one, you can define a reference with a wildcard:

It will then pick any object whose name matches user* where * can be any string.

There is one limitation, you can only refer to objects that are defined above in the file. If you want to use an existing object that is already present in your database you can also provide the id of the object. For this to work however the setter method for that property must have a type hint.

It is also possible to create a relation to a random object by id:

Note: To create a string @foo that is not a reference you can escape it as \@foo

Multiple References

If we want to also add group members, there are two ways to do this. One is to define an array of references to have a fixed set of members:

The other, which is more interesting, is to define a reference with a wildcard, and also tell Alice how many object you want:

In this case it will pick 5 fixture objects which have a name matching user*.

You can also randomize the amount by combining it with faker data:

Note: You do not need to define multi-references inside an array, since they are automatically translated to an array of objects.

Self reference

The @self reference is assigned to the current fixture instance.

Passing references to providers

You can pass references to providers much like you can pass variables:

Handling Unique Constraints

Quite often some database fields have a unique constraint set on them, in which case having the fixtures randomly failing to generate because of bad luck is quite annoying. This is especially important if you generate large amounts of objects, as otherwise you will most likely never encounter this issue.

By declaring the key as unique using the (unique) flag at the end, Alice will make sure every element of this class that is created has a unique value for that property. For example:

Fixture inheritance

Base fixtures, to be extended from, can be created to be able to only need to define less additional values in a set of common fixture definitions.

By declaring a fixture as a template using the (template) flag, Alice will set the instance as a template for that file. Templates instances are not persisted.

Templates can also make use of inheritance themselves, by extending from other templates, allowing you to create, mix and match templates. For example:

Templates can be extended by other fixtures making use of the (extends) flag followed by the name of the template to extend.

Inheritance also allows to extend from several templates. The last declared extends will always override values from previous declared extends templates. However, extensions properties will never override values set explicitly in the fixture spec itself.

In the following example, the age from user_young will override the age from user in user1, while username will remain user1

Including files

You may include other files from your fixtures using the top-level include key:

In relative/path/to/file.yml:

In relative/path/to/another/file.yml:

All files are merged in one data set before generation, and the includer's content takes precedence over included files' fixtures in case of duplicate keys.

Variables

For some advanced use cases you sometimes need to reference one property from another, for example to generate the update date while making sure it is after the creation date. If you simply use two random dates it might be that they are reversed, but Alice let's you refer to other properties using the traditional PHP $variable notation.

Let's add created/modified dates to our group:

As you can see, we make sure that the update date is between the creation date and the current time, which ensure the data will look real enough.

Value Objects

Sometimes you require value objects that are not persisted by an ORM, but are just stored on other objects. You can use the (local) flag on the class or the instance name to mark them as non-persistable. They will be available as references to use in other objects, but will not be returned by the LoaderInterface::load call.

For example this avoids getting an error because Geopoint is not an Entity if you use the Doctrine persister.

Custom Faker Data Providers

Sometimes you need more than what Faker and Alice provide you natively, and there are two ways to solve the problem:

1. Embed PHP code in the yaml file. It is included by the loader so you can add arbitrary PHP as long as it outputs valid yaml. That said, this is like PHP templates, it quickly ends up very messy if you do too much logic, so it's best to extract logic out of the templates.

2. Add a custom Faker Provider class. These are just classes that expose public methods, all the public methods are available as <method()> in the Alice fixture files. For example if you want a custom group name generator and you use the standard Doctrine Fixtures package in a Symfony2 project, you could do the following:

   That way you can now use name: <groupName()> to generate specific group names.

Custom Setter

In case, you want to specify a custom function that will be used to set all the values, you can specify a __set value:

When the objects are populated, the customSetter function will be called, with the first parameter being the key, the second one being the value (so similar to the magic PHP setter). In the above example, the following call will be made on the instance when populating:

Complete Sample

In the end, using most of the tools above, we have this file creating a bunch of users and a group, all of it being linked together, and with little typing:

If you like to have a few specific users with specific data to write tests against of course you can define them above/below the ones using the randomized data. Combine it all as you see fit!

Processors

Processors allow you to process objects before and/or after they are persisted. Processors must implement the ProcessorInterface.

Here is an example where we may use this feature when using the FOSUserBundle:

You can add a list of processors in the load method, e.g.

Or, you can ad them to your loader using the addProcessor() method, e.g.

License

Released under the MIT License, see LICENSE.