NAME

Catalyst::Model::Adaptor - use a plain class as a Catalyst model

SYNOPSIS

Given a good old perl class like:

    package NotMyApp::SomeClass;
    use Moose; # to provide "new"
    sub method { 'yay' }

Wrap it with a Catalyst model:

    package MyApp::Model::SomeClass;
    use base 'Catalyst::Model::Adaptor';
    __PACKAGE__->config( class => 'NotMyApp::SomeClass' );

Then you can use NotMyApp::SomeClass from your Catalyst app:

    sub action :Whatever {
        my ($self, $c) = @_;
        my $someclass = $c->model('SomeClass');
        $someclass->method; # yay
    }

Note that NotMyApp::SomeClass is instantiated at application startup time. If you want the adapted class to be created for call to $c->model, see Catalyst::Model::Factory instead. If you want the adapted class to be created once per request, see Catalyst::Model::Factory::PerRequest.

DESCRIPTION

The idea is that you don't want your Catalyst model to be anything other than a line or two of glue. Using this module ensures that your Model classes are separate from your application and therefore are well-abstracted, reusable, and easily testable.

Right now there are too many modules on CPAN that are Catalyst-specific. Most of the models would be better written as a class that handles most of the functionality with just a bit of glue to make it work nicely with Catalyst. This module aims to make integrating your class with Catalyst trivial, so you won't have to do any extra work to make your model generic.

For a good example of a Model that takes the right design approach, take a look at Catalyst::Model::DBIC::Schema. All it does is glues an existing DBIx::Class::Schema to Catalyst. It provides a bit of sugar, but no actual functionality. Everything important happens in the DBIx::Class::Schema object.

The end result of that is that you can use your app's DBIC schema without ever thinking about Catalyst. This is a Good Thing.

Catalyst is glue, not a way of life!

CONFIGURATION

Subclasses of this model accept the following configuration keys, which can be hard-coded like:

   package MyApp::Model::SomeClass;
   use base 'Catalyst::Model::Adaptor';
   __PACKAGE__->config( class => 'NotMyApp::SomeClass' );

Or be specified as application config:

   package MyApp;
   MyApp->config->{'Model::SomeClass'} = { class => 'NotMyApp::SomeClass' };

Or in your ConfigLoader-loaded config file:

   ---
   Model::SomeClass:
     class: NotMyApp::SomeClass
     args:
       foo: ...
       bar: ...

This is exactly like every other Catalyst component, so you should already know this.

Anyway, here are the options:

class

This is the name of the class you're adapting to Catalyst. It MUST be specified.

Your application will die horribly if it can't require this package.

constructor

This is the name of the class method in class that will create an instance of the class. It defaults to new.

Your application will die horribly if it can't call this method.

args

This is a hashref of arguments to pass to the constructor of class. It is optional, of course. If you omit it, nothing is passed to the constructor (as opposed to {}, an empty hashref).

METHODS

There are no methods that you call directly. When you call $c->model on a model that subclasses this, you'll get back an instance of the class being adapted, not this model.

These methods are called by Catalyst:

COMPONENT

Setup this component.

CUSTOMIZING THE PROCESS

By default, the instance of your adapted class is instantiated like this:

    my $args = $self->prepare_arguments($app); # $app sometimes called $c
    $adapted_class->$constructor($self->mangle_arguments($args));

Since a static hashref of arguments may not be what $class needs, you can override the following methods to change what $args is.

NOTE: If you need to pass some args at instance time, you can do something like:

    my $model = $c->model('MyFoo', { foo => 'myfoo' });

or

    my $model = $c->model('MyFoo', foo => 'myfoo');

prepare_arguments

This method is passed the entire configuration for the class and the Catalyst application, and returns the hashref of arguments to be passed to the constructor. If you need to get dynamic data out of your application to pass to the consturctor, do it here.

By default, this method returns the args configuration key.

Example:

    sub prepare_arguments {
        my ($self, $app) = @_; # $app sometimes written as $c
        return { foobar => $app->config->{foobar}, baz => $self->{baz} };
    }

mangle_arguments

This method is passed the hashref from prepare_arguments, mangles them into a form that your constructor will like, and returns the mangled form. If your constuctor wants a list instead of a hashref, this is your opportunity to do the conversion.

Example:

    sub mangle_arguments {
        my ($self, $args) = @_;
        return %$args; # now the args are a plain list
    }

If you need to do more than this, you might as well just write the whole class yourself. This module is designed to make the common case work with 1 line of code. For special needs, it's easier to just write the model yourself.

SEE ALSO

If you need a new instance returned each time $c->model is called, use Catalyst::Model::Factory instead.

If you need to have exactly one instance created per request, use Catalyst::Model::Factory::PerRequest instead.

AUTHOR

Jonathan Rockway <jrockway@cpan.org>

CONTRIBUTORS

Wallace Reis <wreis@cpan.org>

LICENSE

This module is Copyright (c) 2007 Jonathan Rockway. You may use, modify, and redistribute it under the same terms as Perl itself.