Table of Contents
- About Apostrophe
- Adding Editable Areas: A Simple Example
- Displaying Single Widgets ("Singletons")
- Detecting Empty Areas and Singletons
- More About Grouping Areas into Pages
- Enforcing Permissions
- Minifying CSS and JS in Production
- Passing Data and Calling Functions in the Browser from Server-Side Code
- Cross-Module Includes in Nunjucks
- Loading "logged-in" JavaScript and CSS Assets for Logged-Out Users
- Apostrophe Command-Line Tasks
- Roadmap
- Conclusion and Contact Information
Apostrophe is a content management system. This core module provides rich content editing as well as essential services to tie Apostrophe to your Express application.
In addition to rich text, Apostrophe allows you to add rich media to documents. Apostrophe also includes simple facilities for storing your rich content areas in MongoDB and fetching them back again.
You can try a live demo of the Apostrophe 2 sandbox app here. (Note: the demo site resets at the top of the hour.) See also the apostrophe-sandbox github project.
The sandbox provides a much simpler getting-started guide. We recommend you start reading there. The rest of this document is mostly concerned with features of the apostrophe module itself and not the system as a whole.
Apostrophe introduces "widgets," separate editors for rich media items like photos, videos, pullquotes and code samples. Apostrophe's widgets handle these items much better than a rich text editor on its own.
Apostrophe also supports floating content properly, including wrapping text around images and video. Unlike other rich text editors, Apostrophe addresses the usability problems that go with floating content. Apostrophe users can see exactly where to add text above the floated element and where to add text after it so that it wraps around. When editing, Apostrophe displays positioning arrows before and after rich media elements that make it clear where they connect to the text and ensure it is always possible to add content above and below them. Apostrophe users can also easily select, cut, copy and paste rich content widgets exactly as if they were part of the text, without breaking them. You can even copy a video widget from one page of a site to another.
In summary, Apostrophe's rich media widgets are independently edited, but they are also part of the flow of a rich text document, with robust support for floating them if desired and displaying them at various well-chosen sizes rather than arbitrary sizes that may not suit your design. This is the major advantage of Apostrophe over other rich text editors.
Apostrophe also provides server-side node.js code providing a back end for all of the above: storing uploaded files, validating rich content, and storing rich content areas in MongoDB.
Apostrophe is led by the team at P'unk Avenue. Previously we created Apostrophe 1.5, a well-regarded content management system based on the Symfony framework for PHP. This Node-based version of Apostrophe is a complete rewrite, benefitting on everything we've learned since.
Apostrophe wouldn't be nearly so awesome without nunjucks, Express and Rangy. Please don't go anywhere near HTML's contentEditable
attribute without Rangy. And a hip flask.
Apostrophe 2 is in production on some client sites. Overall code stability is beta at this point.
Apostrophe is intended to work in all major browsers from IE7 up, with best results in modern browsers such as recent versions of Firefox and Chrome. Of course the content you create with Apostrophe could work with any browser.
Apostrophe's server-side components are built in Node and require Express 3.0. Although in principle browser-side components of Apostrophe could talk to other languages, right now a close partnership with Node code on the server is driving the flow of development.
Apostrophe's server-side code uses uploadfs to store media files. uploadfs allows you to decide whether to keep them in a local filesystem, Amazon S3 or a custom backend.
Apostrophe does not require any external CSS framework. Apostrophe's internal templates are processed with Nunjucks, which is awesome, but your Node application does not have to use Nunjucks.
You must have the following on your development and production servers:
node, of course
mongodb, on your local machine (or edit app.js to point somewhere else)
imagemagick, to resize uploaded images (specifically the convert
command line tool)
Mac developers can install imagemagick via MacPorts. Your production server will need it too; it's probably a simple apt-get install
or yum
command away. Heroku includes imagemagick as standard equipment.
Use the apostrophe-site module to configure your project. It is possible to configure a site without it; if you really want to, check out the source of the apostrophe-site module. But you really don't want to.
The easiest way to add Apostrophe-powered editable rich content areas to your Node Express 3.0 project is to use Apostrophe's aposArea
function, which is made available to your Express templates when you configure Apostrophe. Here's a simple example. Here we are using the global
page, a convenient virtual "page" for storing areas that are used throughout the site, like a shared footer:
{{ aposArea(global, 'content1', {
edit: edit
}) }}
This is from a Nunjucks template. The Nunjucks template language, which is compatible with the popular Jinja and Twig languages, is used throughout Apostrophe.
Sometimes Apostrophe's default set of controls include features that don't make sense in a sidebar or otherwise don't suit a design. In these cases you can limit the list.
This aposArea
call turns on all of the controls. You can leave anything you like off the controls
list:
{{ aposArea(global, 'footer', {
controls: [ 'style', 'bold', 'italic', 'createLink', 'image', 'video', 'pullquote', 'code' ]
}) }}
You can also change the "styles" menu. However keep in mind that each "style" must be a legitimate HTML block element name and that not all browsers may support every block element in the rich text editor:
{{ aposArea(global, 'footer', {
controls: [ 'style', 'bold', 'italic', 'createLink', 'image', 'video', 'pullquote', 'code' ],
styles: [ { value: 'div', label: 'Normal' }, { value: 'h3', label: 'Heading' }]
}) }}
Please note: at this time you should always list div
as the first style. We intend to remove this requirement in the future.
Of course, sometimes you want to enforce a more specific design for an editable page. You might, for instance, want to require the user to pick a video for the upper right corner. You can do that with aposSingleton
:
{{ aposSingleton(global, 'my-video', 'video') }}
Note that you may pass additional options to the widget by passing an object as the fourth argument.
Singletons are stored as areas. The only difference is that the interface only displays and edits the first item of the specified type found in the area. There is no rich text editor "wrapped around" the widget, so clicking "edit" for a video immediately displays the video dialog box.
Only widgets (images, videos and the like) may be specified as types for singletons. For a standalone rich-text editor that doesn't allow any widgets, just limit the set of controls to those that are not widgets:
{{ aposArea(page, 'main', { controls: [ 'style', 'bold', 'italic', 'createLink' ] }) }}
It's common to want to do something special if an area or singleton is empty, especially if the user does not have editing privileges. You can detect that:
{% if (not edit) and aposSingletonIsEmpty({ area: page.areas.sidebarVideo, type: 'video' }) %}
<p>Default placeholder video might go here</p>
{% endif %}
aposAreaIsEmpty
is also available. (Singletons are stored as areas but aposSingletonIsEmpty is correctly written to detect whether a widget of the proper type is present.)
"What if I'm building a site with lots of pages? Each page might have two or three areas. Is there an efficient way to get them all at once?"
Sure! Apostrophe provides a getPage
method for this purpose.
If you pass the slug /about
to getPage
, and areas exist with the following slugs:
/about:main
/about:sidebar
/about:footer
Then getPage
will fetch all of them and deliver an object like this:
{
slug: '/about',
areas: {
main: {
slug: '/about/:main',
content: 'main content'
}
sidebar: {
slug: '/about/:sidebar',
content: 'sidebar content'
}
footer: {
slug: '/about/:sidebar',
content: 'footer content'
}
}
}
"Those page objects look useful. Can I store other stuff in those? Page titles and so forth?" Yes. You can write your own mongo code to set additional properties on the objects in the pages collection. Apostrophe won't mind. And you can pass those properties when calling putPage
. Just make sure you retrieve the entire page object with getPage
or a suitable MongoDB query before saving it with putPage
so you don't lose data.
Apostrophe's putArea and getArea methods are written to automatically spot slugs containing a ":" and update or fetch an area within the areas property of a page in the pages collection, rather than creating a freestanding area object in the areas collection.
The apostrophe-pages module uses this method to deliver complete pages automatically for you. In most cases this is what you'll want to do. In rarer cases you'll write your own routes that need to deliver content. See the sandbox project and the apostrophe-pages
module for examples.
TODO: expand this section. Permissions on pages are enforced via lists of groups and people who have been given editing or viewing privileges, and admins can always edit everything. The permissions method usually should not be overridden because it would be incompatible with the fast database queries that locate pages the user is permitted to see or edit based on permissions.
You can hide edit buttons by passing edit: false
to the aposArea
function, and you should if the user doesn't have that privilege. But that doesn't actually prevent clever users from making form submissions that update areas. Fortunately Apostrophe's apos.permissions function is used to verify that the current user (based on req.user, if present) may carry out the requested action.
You can extend apos with additional widgets, and the apostrophe-twitter and apostrophe-rss modules provide working examples. New widget types can even have custom loaders that bring in additional data on the server side as needed. It's neat stuff.
The apostrophe-pages module extends Apostrophe with full blown support for trees of editable web pages (like having a static site, except of course that your users can edit it gorgeously and intuitively).
The apostrophe-snippets module provides a way to maintain a library of reusable content and insert it anywhere on the site. This widget is also designed as a basis for creating modules like apostrophe-blog, apostrophe-events and apostrophe-map with a minimum of code duplication.
Apostrophe has built-in support for minifying its CSS and JS. This is done via the uglify-js
and clean-css
modules, which are pure JS and have good performance given the complexity of what they do.
By default, minification does not occur. This allows developers to click refresh and immediately see their changes, and also avoids the hassles of debugging minified code.
To turn on minification, just pass the minify: true
option to Apostrophe. The aposScripts
and aposStylesheets
locals will then load a single minified file of each type.
Pass that option only in staging and production environments. We recommend resisting the urge not to minify on a staging server, because you need a truly faithful production-like environment to avoid surprises in production. You should minify on staging if you minify in production.
Apostrophe automatically minifies CSS and JS on the first request received (in each process), then reuses the result. There is no need to "clear the cache" or rebuild assets with grunt. Your production deployment process should always involve restarting Apostrophe, which will be the case if you use Stagecoach as seen in our sandbox project.
One of the great challenges of full-stack development is to pass data and call functions in browser-side javascript from server-side code without making a mess. Apostrophe can help here. Apostrophe has features that help you create a single block of clean JavaScript code and data assignment statements without escaping problems or a zillion separate script
elements.
Apostrophe adds req.pushCall
and req.pushData
to the Express request object. You call req.pushCall
like so:
req.pushCall('something.func(?)', { any: { json: [5, 7] } });
You can do this as many times as you need to.
You can then call:
apos.getCalls(res)
To obtain a block of JavaScript code ready to be popped into a script
tag at the end of the body
element. It's easy to pass that to your templates. If you are using the apostrophe-pages
module, you don't even have to do that, because it is already passed as the calls
property.
Every ?
in the pattern (the first argument) is replaced by a correctly JSON-encoded representation of each additional argument. You can use as many ?
s as you need.
JSON-encoding is a great way to avoid escaping bugs, but sometimes you do want one of your parameters to be inserted literally, for instance to pass a constructor function name dynamically. To do that, use @
rather than ?
:
req.pushCall('@(?)', 'SomeFunction', 'SomeData')
Note that req.pushCall
pushes a call to be invoked just in the current HTTP request's response. To make calls that will be included in the calls
property for every request, make a call like this:
apos.pushGlobalCall('myblog.setup(?)', options)
Apostrophe's modules use this approach heavily for browser-side initialization.
You can use both ?
(escaped via JSON) and @
(inserted literally) as placeholders here too.
Note: If you are not using the apostrophe-pages
module, you'll need to call apos.getCalls(req)
and apos.getGlobalCalls()
to get two blocks of JavaScript source code ready to insert at the end of the body inside a script
element. If you are using apostrophe-pages
this is done for you, so your layout can just output both sets of calls (already concatenated together) via the calls
property.
Apostrophe's apostrophe-pages
module automatically makes these calls available as a single block of code ready to insert into a script tag at the end of the body.
Well, we've already given you a way to pass data, since you can pass any arguments you like to any browser-side JavaScript code via the pushCall
mechanism and JSON-escape the arguments automatically. However, there is an alternative way to pass data.
By calling req.pushData({ ... })
and/or apos.pushGlobalData({ ... })
, you can pass objects to the browser which will be merged into the browser-side JavaScript object apos.data
. The merge operation is carried out recursively via jQuery.extend
. That means that you can set some sub-properties on one call and set additional sub-properties on another.
Apostrophe's modules use this mechanism to push options for constructors like AposBlogPosts
to scoop up later.
If you are not using the apostrophe-pages
module to render the results, you'll need to call apos.getData(req)
and apos.getGlobalData()
to get the JavaScript blocks you'll want to invoke at the end of the body. You should do so before doing the same for getCalls
and getGlobalCalls
. If you are using apostrophe-pages
, you can just take advantage of the data
property already supplied to your page template.
req.pushData
and req.pushGlobalData
are not as flexible as req.pushCall
and req.pushGlobalCall
. In many cases you'll want to use the latter.
Keep in mind that all data passed via any of these mechanisms must be JSON-friendly. You cannot pass server-side function objects to browser-side code. That's just life in JavaScriptLand.
It is possible to include Nunjucks macros and template files from one module in a template that resides in another, or in a project-level page template, layout template, etc.
Each module registers itself for this purpose. By convention, the name of a module for this purpose is the same as the name of the module, minus the "apostrophe-" prefix if any. The name is always-hyphenated, neverCamelCased.
The syntax is:
{% include "modulename:file.html" %}
For example, the apostrophe-schemas
module relies on this feature to allow you to include its macros for use in other modules:
{% include "schemas:schemaMacros.html" %}
If you are using apostrophe-site
to structure a project (and you should be), project-level overrides of an Apostrophe module like this:
lib/modules/apostrophe-blog/views/blogMacros.html
Are automatically respected when you write:
{% include "blog:blogMacros.html" %}
In most cases you will not need the cross-module includes feature, but if you are creating a module that delivers a "mixin" intended to augment other modules without the need for subclassing its usefulness will immediately become clear.
Apostrophe refrains from loading most of its JavaScript and CSS for logged-out users. This is a good thing because it renders pages faster.
However sometimes those features are really useful for anonymous users. For instance, our apostrophe-moderator
module lets the public submit new events, articles and so forth. That depends on the ability to present forms and edit content the same way a logged-in user would.
To do that, just use the apos.requireScene
method in your browser-side JavaScript code. The code in your callback is guaranteed to have access to all the JavaScript, CSS and DOM template assets that logged-in users see:
apos.requireScene('user', function() {
// Do great stuff like using `apostrophe-schemas` to process forms
});
Since scene upgrades are calcualted during Apostrophe's minification process when the server starts up, it won't take long to load the assets and start running your callback. Do make sure you have set minify: true
in data/local.js
on your production server.
We often need to carry out command line tasks, such as database migrations, with access to the same database and capabilities that regular Apostrophe code has access to. Apostrophe makes it really easy to register command line tasks as part of your application.
If the command line is:
node app.js
The app will start listening for connections as normal, while if it is:
node app.js apostrophe:migrate
The app will execute that task.
To see the available tasks, just type:
node app.js apostrophe:help
This will list all of the registered tasks.
You'll find that app.js
invokes this code just before listening on the port:
// Command line tasks
if (apos.startTask()) {
// Chill and let the task run until it's done,
// don't try to listen or exit
return;
}
appy.listen();
All you have to do is extend this by passing an object to apos.startTask
. Each property of that object is a "task group" with one or more tasks. Each "task group" is an object with one or more task functions. Task functions receive three arguments: the apos
object, an argv
object with any command line options, and a callback to be invoked when the task is done.
Confused? Here's how to implement a single task called project:init
:
// Earlier in app.js
var myTasks = {
project: {
init: function(apos, argv, callback) {
// Do time consuming, asynchronous things!
// When we're finished:
return callback(null);
}
}
};
// In the listen function at the end of app.js
if (apos.startTask(myTasks)) {
return;
}
appy.listen();
This structure allows for projects with many tasks.
Pass something other than null when invoking the callback.
The argv
object comes from the optimist module. You can pick up command line arguments with it in a super-easy, super-friendly way. See the optimist module for more information.
Good question. Lots of Apostrophe methods expect a req
object, because permissions are tied to the identity of the user, and certain types of caching are tied to the lifetime of a request.
If your task needs to call a function like snippets.get
which requires a req
object, call this function to get a req
object that always has unlimited permissions:
apos.getTaskReq()
The apos
object is an EventEmitter. In English, that means you can write:
apos.on('task:apostrophe:migrate:before', function() {
apos.taskBusy();
// Do a variety of things asynchronously
apos.taskDone();
}
Or:
apos.on('task:apostrophe:migrate:after', function() {
apos.taskBusy();
// Do a variety of things asynchronously
apos.taskDone();
}
Apostrophe needs to know when your task is finished. But you're likely to call asynchronous functions (functions with callbacks), so Apostrophe can't just assume that the task is finished when the function returns.
apos.taskBusy()
solves this problem. If your task needs to do lots of asynchronous stuff, call apos.taskBusy()
. The process will not exit until you call apos.taskDone()
.
Your event handler can call apos.taskFailed()
in this situation.
Apostrophe is a work in progress. See github issues for details on some of our plans as well as the Google Group.
That's it! You should have everything you need to enable rich content editing on your sites. If not, open an issue on github and we'll talk. See also the above roadmap.
Tom Boutell