A simple extensible in-memory data structure of resources.
var ripple = core()
ripple(name, body) // setter
ripple(name) // getterYou can also use the method-chained API:
ripple
.resource(name, body)
.resource(name, body)
...The resources it registers are accesible under ripple.resources.
A canonical resource is an object with the following shape and three properties:
{ name: 'foo'
, body: 'bar'
, headers: { 'content-type': 'text/plain' }
}That is, it can be uniquely identified (name), the resource itself (body) and some arbitrary metadata (headers). Core only deals with the content-type header, however other modules may add and interpret their own per-resource metadata.
Core only comes with one type (text/plain) out of the box, so will fail to register anything other than a string. This is to make it very extensible and future-proof, such that you could for example create other exotic types like application/jsx, text/jade or data/immutable.
Note that you do not have to register a canonical resource, you will most likely use shortcuts in your application code (see API for more).
ripple('foo', 'bar')
// will result in ripple.resources ===
{
foo: { name: 'foo'
, body: 'bar'
, headers: { 'content-type': 'text/plain' }
}
}When an content-type header is not explicitly given, core will loop through it's registered types and see if any of them understand this particular resource (by passing { name, body, headers } to each type check function). If any of them do:
- The
content-typeheader will be set - The type
parsefunction will be run on the resource - The resource will be stored internally
- A change event will be emitted
You'll need to extend ripple.types to tell it how to interpret other resources. Each type object should have the following:
{ header: 'the content type you are registering'
, check: function // necessary
, parse: function // optional
}The parse function is a chance to initialise the resource, set default headers, etc. Some examples:
application/data- proxies change events so you can do per-resource change listenersripple('data').on('change, fn)application/javascript- turns a function as a string into a real function (useful since streamed over WS).
Other modules can also extend existing parse functions. For example, sync extends every type parse function to add the ability to define server/client transformation functions.
See other existing vanilla types for more examples: Data, Versioned Data, Functions, HTML, CSS.
The core instance is emitterified. Whenever a resource is registered, a change event will be emitted.
ripple.on('change', doSomething)ripple('name') // - returns the resource body if it exists
ripple('name', body) // - creates & returns resource, with specified name and body
ripple('name', body, headers }) // - creates & returns resource, with specified name, body and headers
ripple({ name, body, headers }) // - creates & returns resource, with specified name, body and headers
ripple([ ... ]) // - calls ripple on each item - registers an array of resources
ripple.resources // - returns raw resources
ripple.resource // - alias for ripple, returns ripple instead of resource for method chaining
ripple.register // - alias for ripple
ripple.on // - event listener for changes - all resources