A pageable, drop-in replacement for Backbone.Collection inspired by Backbone.Paginator, but much better.
Table of Contents
- Supports client-side and server-side operations
- You can initialize
Backbone.PageableCollection
to paginate and/or sort on the client-side, server-side or both. - Infinite paging
- Many public APIs like Github
or Facebook support
infinite paging,
Backbone.PageableCollection
can handle them easily. - Comes with reasonable defaults
- Server API parameters preconfigured to work with most Rails RESTful APIs by default.
- Works well with existing server-side APIs
- Query parameter mappings are all configurable, and you can use either 0-based or 1-based indices.
- Bi-directional event handling
- In client-mode, any changes done on one page is immediately reflected on the others with the appropriate events propagated.
- 100% compatible with existing code
Backbone.PageableCollection
is a strict superset ofBackbone.Collection
and passes its test suite.- Well tested
- Comes with 100s of tests in addition to the
Backbone.Collection
test suite. - Well documented
- Use cases and functionality are thoroughly documented.
- No surprising behavior
Backbone.PageableCollection
performs internal state sanity checks at appropriate times, so it is next to impossible to get into a weird state.- Light-weight
- The library is only 4.2KB minified and gzipped.
The following examples utilizes Backgrid.js to render the collections.
npm install backbone-pageable
bower install backbone-pageable
<script src="underscore.js"></script>
<script src="backbone.js"></script>
<script src="backbone-pageable.js"></script>
var PageableCollection = require("backbone-pageable");
var PageableCollection = Backbone.PageableCollection;
Like Backbone.Collection, you can provide a URL endpoint, configure your initial
pagination state and server API mapping by extending
Backbone.PageableCollection
:
var Book = Backbone.Model.extend({});
var Books = Backbone.PageableCollection.extend({
model: Book,
url: "api.mybookstore.com/books",
// Any `state` or `queryParam` you override in a subclass will be merged with
// the defaults in `Backbone.PageableCollection` 's prototype.
state: {
// You can use 0-based or 1-based indices, the default is 1-based.
// You can set to 0-based by setting ``firstPage`` to 0.
firstPage: 0,
// Set this to the initial page index if different from `firstPage`. Can
// also be 0-based or 1-based.
currentPage: 2,
// Required under server-mode
totalRecords: 200
},
// You can configure the mapping from a `Backbone.PageableCollection#state`
// key to the query string parameters accepted by your server API.
queryParams: {
// `Backbone.PageableCollection#queryParams` converts to ruby's
// will_paginate keys by default.
currentPage: "current_page",
pageSize: "page_size"
}
});
You can initialize state
and queryParams
from the constructor too:
var Books = Backbone.PageableCollection.extend({
model: Book,
url:"api.mybookstore.com/books"
});
var books = new Books([], {
// All the `state` and `queryParams` key value pairs are merged with
// the defaults too.
state: {
firstPage: 0,
currentPage: 0
},
queryParams: {
currentPage: "current_page",
pageSize: "page_size"
}
});
To adapt to an existing server API that do not use will_paginate
keys, you
can configure the queryParams
object hash to map state
keys to the query
parameters your server will accept. Those query parameters will be in the query
string of the URL used for fetching. You can also put extra items into
queryParams
and they will be in the query string as is. Setting null
as
the value of any mapping will remove it from the query string. Finally, the
values in the queryParams
can be either a literal value or a parameter-less
function that returns a value.
This is a listing of the default state
and queryParam
values.
state |
queryParams |
||
---|---|---|---|
Attribute | Value | Attribute | Value |
firstPage | 1 | ||
lastPage | null | ||
currentPage | null | currentPage | "page" |
pageSize | 25 | pageSize | "per_page" |
totalPages | null | totalPages | "total_pages" |
totalRecords | null | totalRecords | "total_entries" |
sortKey | null | sortKey | "sort_by" |
order | -1 | order | "order" |
directions | { "-1": "asc", "1": "desc" } |
You can consult the API documentation for a detailed explaination of these fields.
You can access the pageable collection's internal state by looking at the
state
object attached to it. This state object, however, is generally
read-only after initialization. There are various methods to help you manage
this state, you should use them instead of manually modifying it. For the
unusual circumstances where you need to modify the state
object directly, a
sanity check will be performed at the next time you perform any
pagination-specific operations to ensure internal state consistency.
Method | Use When |
---|---|
setPageSize |
Changing the page size |
setSorting |
Changing the sorting |
switchMode |
Switching between modes |
state |
Need to read the internal state |
get*Page |
Need to go to a different page |
hasPrevious, hasNext |
Check if paging backward or forward is possible |
In addition to the above methods, you can also synchronize the state with the
server during a fetch. Backbone.PageableCollection
overrides the default
Backbone.Collection#parse method to
support an additional response data structure that contains an object hash of
pagination state. The following is a table of the response data structure
formats Backbone.PageableCollection
accepts.
Without State | With State |
---|---|
[{}, {}, ...] |
[{ pagination state }, [{}, {} ...]] |
Most of the time, providing something like this in your response is sufficient for updating the pagination state.
[{"total_entries": 100}, [{}, {}, ...]]
Since 1.1.7, customizing parse
has been simplified and the default
implementation now delegates to two new methods - parseState
and
parseRecords
. You are encouraged to override them instead of parse
if it
is not clear how to do so.
See the API for details on
customizing parseState
and parseRecords
.
Backbone.PageableCollection
is 100% compatible with Backbone.Collection
's interface, so you can bootstrap the models and supply a comparator to the
constructor just like you are used to:
// Bootstrap with just 1 page of data for server-mode, or all the pages for
// client-mode.
var books = new Books([
{ name: "A Tale of Two Cities" },
{ name: "Lord of the Rings" },
// ...
], {
// Paginate and sort on the client side, default is `server`.
mode: "client",
// This will maintain the current page in the order the comparator defined
// on the client-side, regardless of modes.
comparator: function (model) { return model.get("name"); }
});
Backbone.Pagination
defaults to server-mode, which means it only holds one
page of data at a time. All of the get*page
operations are done by
delegating to fetch
. They return a jqXHR
in this mode.
books.getFirstPage();
books.getPreviousPage();
books.getNextPage();
books.getLastPage();
// All the `get*Page` methods under server-mode delegates to `fetch`, so you
// can attach a callback to the returned `jqXHR` objects' `done` event.
books.getPage(2).done(function () {
// do something ...
});
All of the get*Page
methods accept the same options
Backbone.Collection#fetch accepts
under server-mode.
Client-mode is a very convenient mode for paginating a handful of pages entirely on the client side without going through the network page-by-page. This mode is best suited if you only have a small number of pages so sending all of the data to the client is not too time-consuming.
var books = new Books([
// Bootstrap all the records for all the pages here
], { mode: "client" });
All of the get*Page
methods reset the pageable collection's data to the models
belonging to the current page and return the collection itself instead of a
jqXHR
.
// You can immediately operate on the collection without waiting for jQuery to
// call your `done` callback.
var json = JSON.stringify(books.getLastPage());
// You can force a fetch in client-mode to get the most updated data if the
// collection has gone stale.
books.getFirstPage({ fetch: true });
// Do something interesting with books...
Infinite paging mode is a hybrid of server mode and client mode. Once initialized and bootstrapped, paging backwards will be done on the client-side by default while paging forward will be done by fetching.
As before, you can make use of getFirstPage
, getPreviousPage
,
getNextPage
, and getLastPage
for navigation under infinite-mode. If a
page has been fetched, you can use getPage
directly with the page number, an
error will be thrown if the page has not been fetched yet.
By default, Backbone.PageableCollection
parses the response headers to find
out what the first
, last
, next
and prev
links are. The parsed
links are available in the links
field.
var Issues = Backbone.PageableCollection.extend({
url: "https://api.github.com/repos/documentclound/backbone/issues?state=closed",
mode: "infinite"
// Initial pagination states
state: {
pageSize: 15,
sortKey: "updated",
order: 1
},
// You can remap the query parameters from ``state`` keys from the default
// to those your server supports. Setting ``null`` on queryParams removed them
// from being appended to the request URLs.
queryParams: {
totalPages: null,
totalRecords: null,
sortKey: "sort",
order: "direction",
directions: {
"-1": "asc",
"1": "desc"
}
}
});
var issues = new Issues();
issues.getFirstPage().done(function () {
// do something interesting...
});
If your server API does not return the links using the Link
header like
Github does, you can subclass
Backbone.PageableCollection
to override the parseLinks
methods to
return a links object.
var FBComment = Backbone.Model.extend({});
var FBComments = Backbone.PageableCollection.extend({
model: FBComment,
url: "https://graph.facebook.com/A_REALLY_LONG_FACEBOOK_OBJECT_ID",
mode: "infinite",
// Set the indices to 0-based for Graph API.
state: {
firstPage: 0
},
queryParams: {
pageSize: "limit",
// Setting a parameter mapping value to null removes it from the query string
currentPage: null,
// Any extra query string parameters are sent as is, values can be functions,
// which will be bound to the pageable collection instance temporarily
// when called.
offset: function () { return this.state.currentPage * this.state.pageSize; }
},
// Return all the comments for this Facebook object
parseRecords: function (resp) {
return resp.comments.data;
},
// Facebook's `paging` object is in the exact format
// `Backbone.PageableCollection` accepts.
parseLinks: function (resp, xhr) {
return resp.comments.paging;
}
});
To act on the newly fetched models under infinite mode, you can listen to the
fullCollection
reference's add
event like you would under client mode,
and render the newly fetched models accordingly.
var ToiletPaper = Backbone.View.extend({
events: {
"scroll": "fetchSheets"
},
initialize: function (options) {
this.listenTo(this.collection.fullCollection, "add", this.addSheet);
},
addSheet: function () {
// ...
},
fetchSheets: function () {
this.collection.getNextPage();
},
// ...
});
var wordsOfTheDay = new Backbone.PageableCollection({
mode: "infinite",
// url, initial state, etc...
});
var toiletPaper = new ToiletPaper({collection: wordsOfTheDay});
$("#toilet-paper-dispenser").append(toiletPaper.render().el);
wordsOfTheDay.fetch();
Sorting has been drastically simplified in the 1.0 release while retaining the full power it had in older versions.
The main way to define a sorting for a pageable collection is to utilize the
setSorting
method. Given a sortKey
and an order
, setSorting
sets state.sortKey
and state.order
to the given values. If order
is
not given, state.order
is assumed. By default a comparator is applied to the
full collection under client mode. Calling sort
on the full collection will
then get the entire pageable collection sorted globally. When operating under
server or infinite mode, no comparator will be applied to the collection as
sorting is assumed to be done on the server by default. Set options.full
to
false
to apply a comparator to the current page under any mode. To sort a
pageable collection under infinite mode on the client side, set options.side
to "client"
will apply a comparator to the full collection.
Setting sortKey
to null
removes the comparator from both the current
page and the full collection.
var books = new Books([
...
], {
mode: "client"
});
// Sets a comparator on `#fullCollection` that sorts the title in ascending
// order
books.setSorting("title");
// Don't forget to call `sort` just like you would on a `Backbone.Collection`
books.fullCollection.sort();
// Clears the comparator
books.setSorting(null);
// Sets a comparator on the current page that sorts the title in descending
// order
books.setSorting("title", 1, {full: false})
books.sort();
books.switchMode("infinite");
// Sorts the books collection under infinite paging mode on the client side
books.setSorting("title", -1, {side: "client"});
books.fullCollection.sort();
books.switchMode("server");
// Sets a comparator on the current page under server mode
books.setSorting("title", {side: "client", full: false});
books.sort();
This is one of the areas where Backbone.PageableCollection
truely shines. A
Backbone.PageableCollection
instance not only can do everything a plain
Backbone.Collection
can for the current page, in client-mode, it can also
synchronize changes and events across all of the pages. For example, you can add
or remove a model from either a Backbone.PageableCollection
instance, which
is holding the current page, or the
Backbone.PageableCollection#fullCollection
collection, which is a plain
Backbone.Collection
holding the models for all of the pages, and the pages
will all update themselves to maintain within a page size. Any additions,
removals, resets, model attribute changes and synchronization actions are
communicated between all the pages throught the two collections.
// The books collection is initialized to start at the first page.
var books = new Books([
// bootstrap with all of the models for all of the pages here
], {
mode: "client"
});
// A book is added to the end of the current page, which will overflow to the
// next page and trigger an `add` event on `fullCollection`.
books.push({ name: "The Great Gatsby"});
books.fullCollection.at(books.state.currentPage - 1 * books.state.pageSize).get("name");
>>> "The Great Gatsby"
// Add a new book to the beginning of the first page.
books.fullCollection.unshift({ name: "Oliver Twist" });
books.at(0).get("name");
>>> "Oliver Twist"
See here.
Why another paginator?
This project was born out of the needs for a backing model for Backgrid.Extension.Paginator - an extension for the Backgrid.js project. The project needed a smart and intuitive model that is well-documented and well-tested to manage the paginator view. Upon examining the popular project Backbone.Paginator, the author has concluded that it does not satisfy the above requirements. Furthermore, the progress of the the project is too slow. The author hopes to reinvent a better wheel that is better suited and supported for Backgrid.js.
Which package managers does backbone-pageable support?
bower, CommonJS and AMD as of 0.9.0.
Why doesn't backbone-pageable support filtering?
Wheels should be reinvented only when they are crooked. backbone-pageable aims to do one thing only and does it well, which is pagination and sorting. Besides, since Backbone.PageableCollection is 100% compatible with Backbone.Collection, you can do filtering fairly easily with Backbone's built-in support for Underscore.js methods.
- 1.3.2
- Removed support for Backbone <= 1.0.0.
- Switched build system to Grunt
- Backbone.PageableCollection is now set to PageableCollection under all supported environments (CommonJS, AMD and browser).
- Fetching under infinite mode no longer triggers a`reset` event everytime. (Issue #96)
- Setting a page size via setPageSize no longer throws exception when the collection is empty. (Issue #93)
- Fixed broken demos (Issue #99)
- 1.3.1
- Exclude extra query parameters with null values from the query string. (Issue #90)
- Propagate options from
parse
toparseState
andparseRecords
. (Pull #91) - Custom value extractor (sortValue) for
setSorting
and_makeComparator
. (Issue #89) - Changing page size at boundary pages no longer throws errors. (Issue #92)
- 1.3.0
- Pass
from
andto
to theoptions
object sent to event handlers afterget*Page
. - Fetching new page under infinite mode no longer silences
add
and triggersreset
. It will now simple triggersadd
. - Slight code clean up.
- Pass
- 1.2.4
- Moved initialization from
initialize
code to the constructor. (Issue #83)
- Moved initialization from
- 1.2.3
- Support non-array elements in the constructor. (Thanks Gabriel Bédard Sicé) (Pull #76)
- Added
getPageByOffset
method. (Thanks Kee-Yip Chan) (Pull #77)
- 1.2.2
parseRecords
doesn't get called twice during client and infinite mode anymore. (Issue #71)- No
RangeError
should be thrown iffirstPage == 1
,currentPage == 1
andtotalPages == 0
. (Issues #74)
- 1.2.1
parseLinks
now returns an empty object instead of throwing an error if there's noLinks
header found. (Issue #69)
- 1.2.0
- Tested against Backbone 1.0 and Underscore 1.4.4. (Issue #56)
- 1.1.9
- Bugs Fixed
- Copy instance properties from pageable collection to full collection. (Issue #55)
- 1.1.8
- Bugs Fixed
- Server can now return partial state and 0 for totalRecords. (Issue #41), (Issue #52).
- 0-based infinite mode pageable collection now initializes correctly. (Issue #51).
- 1.1.7
- Changes
- Simplified
parse
.parse
now delegates to two new methods -parseState
andparseRecords
. (Issue #49).
- Simplified
- Bugs Fixed
fetch
now accepts function as itsoptions.url
value. (Issue #50).
- 1.1.6
- Bugs Fixed
- Fixed bug where a page of models disappeared after adding an array of models to the current page under client mode. (Issue #43).
- 1.1.5
- Bugs Fixed
- Add event handlers should be triggered before remove handlers during client mode. (Issue #42).
- 1.1.4
- Changes
- Dropped Backbone 0.9.2 support
- Bugs Fixed
- TypeError thrown when adding a model to an empty client mode collection. (Issue #38).
- Adding with an index inserts into the wrong page under client mode. (Issue #39).
- 1.1.3
- Bugs Fixed
- Updating after fetching under infinite mode should not create useless
add
,remove
andsort
events. (Issue #34). - RangeError when emptying fullCollection during reset (Issue #37).
- Updating after fetching under infinite mode should not create useless
- 1.1.2
- Bugs Fixed
- Fix off by 1 error with
hasNext
andhasPrevious
(Issue #32).
- Fix off by 1 error with
- 1.1.1
- Bugs Fixed
- Fix regression where
fetch
errors out ifurl
is a function (Issue #30). - Fix temperatory state inconsistency when accessing
state
in event handlers during client mode. (Issue #27).
- Fix regression where
- 1.1
- Bugs Fixed
- Lots of fixes for infinite paging.
- Fixed incompatibility with Zepto.
- Enhancements
- Introduced
hasPrevious
andhasNext
for checking if the pageable collection can be paged backward or forward. - Tested against Backbone 0.9.10, jQuery 1.9 and Zepto 1.0rc1.
- Introduced
- 1.0
- Bugs Fixed
- Regression from 0.9.9 where
mode
wasn't saved after calledswitchMode
.
- Regression from 0.9.9 where
- Changed
makeComparator
has been renamed to_makeComparator
and is now a protected method.
- Enhancements
- Improved infinite-mode. Infinite paging mode now runs in a hybrid mode. (Issue #17).
- Greatly simplified sorting. (Issue #19).
- 0.9.13
- Bugs Fixed
pageSize
cannot be larger thantotalRecords
.- Off by 1 problem when shifting a model to the current page after removing
from
fullCollection
. RangeError
when removing the last element from the last page whentotalPages
is > 1.
- 0.9.12
- Enhancements
- Switching modes now resets the states by default.
- Infinite mode now updates
currentPage
as well.
- Bugs Fixed
state
counters will now update automatically under client mode.
- 0.9.11
- Changed
links.first
andlinks.next
is now initialized tourl
for infinite mode. Callingfetch
in infinite mode is now equivalent to callinggetNextPage
.
- 0.9.10
- Bugs Fixed
- The initial call to
getFirstPage
will no longer fail under infinite-mode and will now default to fetch from the collection'surl
. - Function values in
queryParams
now hasthis
bound to the collection instance when called.
- The initial call to
- 0.9.9
- Changed
switchMode
now accepts amode
as the first parameter.state.isClientMode
is removed. There is now a newBackbone.PageableCollection#mode
attribute for this purpose.queryParams.totalRecords
now maps to"total_entries"
.queryParams.directions
now maps to{"-1": "asc", "1": "desc"}
.
- Enhancements
- Support extra
queryParam
parameters and function values. - Infinite paging.
- Support extra
- 0.9.2
This release is tested against Backbone.js 0.9.2 and 0.9.9.
- Enhancements
currentPage
defaults tofirstPage
.
- 0.9.1
- Bugs Fixed
- Instantiating a
PageableCollection
in client-mode without giving it any models no longer throws errors.
- Instantiating a
- Enhancements
- Overriding
state
andqueryParams
in a subclass's prototype now merge with the defaults inBackbone.PageableCollection.prototype
. - fullCollection now respect the parent's prototype.
- Overriding
- 0.9.0
- Initial release
Copyright (c) 2013 Jimmy Yuen Ho Wong
Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.