Love Mike G

For all the former Cake folks in the audience considering a Lithium migration, you’ll appreciate this gist to enable “admin_” prefixed controller actions.

The first step is to create a rule in your action bootstrap. Don’t worry, it sounds more complicated than it really is. Just open /config/bootstrap/action.php and add the following code to the bottom.

Dispatcher::config(array(
    'rules' => array('admin' => array('action' => 'admin_{:action}'))
));

After that, enable a persistent route parameter (another incredibly useful feature) for all admin routes. Open /config/routes.php and add the following route definition:

Router::connect(
    '/admin/{:controller}/{:action}/{:args}',
    array('admin' => true),
    array('persist' => array('admin', 'controller'))
);

I highly recommend reading the docs for the Dispatcher class — there’s a lot you can do with this setup. Likewise, there’s a good introduction to persistent parameters on the Union of Rad blog.

Check it: gist: 628541 – How to save a file using MongoDB’s GridFS in Lithium- GitHub.

I just saved my first GridFS file using Lithium. I’m sure there’s a much better way of doing this outside of my controller using filters, but I’m not quite sure how those work yet.

Basically, I have two models: Job and File. The key here is to create a GridFS-aware model. In my case, it’s the File model. This is done by setting the $_meta property’s source field.

protected $_meta = array('source' => 'fs.files');

For this app’s purpose, keeping Job data separate from File data is important so I needed two save operations. First, I save my Job, then my File.

if ($this->request->data) {
    if ($job->save($this->request->data)) {
        $file = File::create();
        $fileData = array(
            'file' => $this->request->data['file'],
            'job_id' => (string) $job->_id
        );
        if ($file->save($fileData)) {
            $this->redirect('Books::index');
        }
    }
}

Lithium handles all the GridFS operations (as expected) from there. It knows where to look by looking specifically at the file field in your data.

What’s the result you ask? It’s a complete win, that’s what:

> db.jobs.find();
{ "_id" : ObjectId("4cb88674ad7bef7559010000"), "file" : "DeathValley02.jpeg", "first" : "Mike", "last" : "Girouard", "email" : "mikeg@migimedia.com", "key" : "blue-full", "submit_x" : "0", "submit_y" : "0", "submit" : "Upload & Preview" }
> db.fs.files.find();
{ "_id" : ObjectId("4cb88674ad7bef7559030000"), "job_id" : "4cb88674ad7bef7559010000", "chunkSize" : 262144, "length" : 18, "md5" : "c99041c6d57b58d9ae07983aec79d54d" }

Hopefully this saves someone some time. It took me all morning to get here. :)