Tower Resource

Data models that can be stored in any database or remote service using adapters.

Installation

node:

$ npm install tower-resource

browser:

$ component install tower/resource

Examples

var resource = require('tower-resource');

Attributes

resource('post')
  .attr('title') // defaults to 'string'
  .attr('body', 'text')
  .attr('published', 'boolean', false);

Validations

resource('user')
  .attr('email')
    .validate('presence')
    .validate('isEmail')
    .validate('emailProvider', { in: [ 'gmail.com' ] }) // some hypothetical one
  .attr('username')
    .validator(function(val){
      return !!val.match(/[a-zA-Z]/);
    });

There are two DSL methods for validation.

1. validate: for using predefined validations (see tower-validator), purely to clean up the API.
2. validator: for defining custom validator functions right inline. If you want to reuse your custom validator function across resources, just move the function into tower-validator.

Queries

resource('post')
  .where('published').eq(true)
  .all(function(err, posts){

  });

See tower-query for the entire syntax. The where method just delegates to a Query instance. You can also access the query object directly (it just adds .select(resourceName) for you):

resource('post').query().sort('title', -1).all();

Actions

There are 4 main actions for resources (which are just delegated to query().action(name):

• create
• all
• update
• remove

resource('post').create();
resource('post').all();
resource('post').update({ published: true });
resource('post').remove();

Under the hood, when you execute one of these actions, they get handled by a database-/service-specific adapter (mongodb, cassandra, facebook, etc.). Those adapters can perform optimizations such as streaming query results back.

License

MIT