//
// Backbone.PagedCollection
//
// A smarter way to consume list resources over HTTP. Call it Hypermedia, REST,
// whatever. There are other problems that we may wish to solve with the
// current implementation of PagedCollection:
//
// 1. Jumping between specific pages;
// 2. Web Linking works, however `per_page` paramter must be implemented;
// 3. Call to `fetch` is required to retreive Link field, which means bootstrapping is a problem; and
// 4. `parser.js` which is used to parse Link field is too big. May have to move off PEG.js version.
//
// @depends on parser.js which was generated via PEG.js to parse Link-field in the HTTP response header.
//

Backbone.PagedCollection = (function (parser) {

	var find = function (link, type) {
		return _.find(link || [], function (value) {
			return value.rel == type;
		});
	};

	var PagedCollection = Backbone.Collection.extend({
		parse: function (response, xhr) {
			var link = xhr.getResponseHeader('Link');

			if (link) {
				this.link = parser.parse(link);
			}

			return Backbone.Collection.prototype.parse.apply(this, arguments);
		},

// Jump to a given related link as specified by the `rel` attribute.

		jump: function (type, options) {
			var link = find(this.link, type);
			if (link && link.href) {
				this.url = link.href;
				return this.fetch(options);
			}
		}
	});

// As a collection, the only type of related links that we care about are the
// the following 4 methods. You can use `jump` directly, however types such as
// `document`, `external` won't make sense in this context. We can probably
// make `Backbone.Model` smarter as well using similar Web Linking techniques
// however, that is not part of this exercise.

	_.each(['first', 'prev', 'next', 'last'], function  (method) {
		PagedCollection.prototype[method] = function (options) {
			return this.jump(method, options);
		};
	});

	return PagedCollection;

}(parser));