Haml Coffee is a JavaScript templating solution that uses Haml as markup, understands inline CoffeeScript and generates a JavaScript function that renders to HTML. It can be used in client-side JavaScript applications that are using Backbone.js, Spine.js, JavaScriptMVC, KnockoutJS and others, or on the server-side in frameworks like Express.

You can try Haml Coffee online by visiting Haml Coffee Online.

Haml Coffee is available in NPM and can be installed with:

$ npm install haml-coffee

Please have a look at the CHANGELOG when upgrading to a newer Haml Coffee version with npm update.

There are different packages available to integrate Haml Coffee into your workflow:

• CoffeeScriptHaml Syntax highlighting for .hamlc files in Sublime Text.

• grunt-haml for projects using Grunt.
• gulp-haml-coffee for projects using Gulp.
• hem-haml-coffee for projects using Hem.
• stitch-haml-coffee for projects using Stitch.
• Mincer the Sprockets inspired web assets compiler.

• haml_coffee_assets for projects using Rails.
• guard-haml-coffee for projects using Guard.

• Haml Coffee compiler (minified) for compiling in the browser.

The browser distribution doesn't come bundled with CoffeeScript, so you'll have to make sure you've included it before requiring haml-coffee.

You can compile a Haml Coffee template to a JavaScript function and execute the function with the locals to render the HTML. The following code

hamlc = require 'haml-coffee'
tmpl = hamlc.compile '%h1= @title'
html = tmpl title: 'Haml Coffee rocks!'

will create the HTML <h1>Haml Coffee rocks!</h1>.

The compile function can take the compiler options as second parameter to customize the template function:

hamlc.compile '%h1= @title'
  cleanValue: false
  escapeHtml: false

See the compiler options for detailed information about all the available options and browse the codo generated Haml Coffee API documentation.

You can configure Express to use Haml Coffee as template engine.

Starting with version 1.4.0, Haml Coffee has support for Express 3 and can be registered as view engine as follows:

express = require 'express'
app     = express()

app.engine 'hamlc', require('haml-coffee').__express

Alternatively you can also use consolidate.js to register the engine:

express = require 'express'
cons    = require 'consolidate'
app     = express()

app.engine 'hamlc', cons['haml-coffee']

Starting with version 0.5.0, Haml Coffee has support for Express 2 and can be registered as view engine as follows:

express = require 'express'

app = express.createServer()
app.register '.hamlc', require('haml-coffee')

Alternatively you can also use consolidate.js to register the engine:

express = require 'express'
cons    = require 'consolidate'

app = express.createServer()
app.register '.hamlc', cons['haml-coffee']

Express 2 uses a layout file layout.hamlc by default and you have to insert the rendered view body into the layout like this:

!!!
%head
  %title Express App
%body
  != @body

Now you can create a Haml Coffee view

%h1= "Welcome #{ @name }"
%p You've rendered your first Haml Coffee view.

that you can render with:

app.get '/', (req, res) ->
  res.render 'index.hamlc', name: 'Express user'

Express 3 has removed layout support, but you can get it back by installing express-partials and configure it as middleware:

partials = require 'express-partials'
app.use partials()

It's possible to use Haml Coffee as the default template engine by setting the view engine:

app.configure ->
  app.set 'view engine', 'hamlc'

which allows you to omit the .hamlc extension when rendering a template:

app.get '/', (req, res) ->
  res.render 'index', name: 'Express user'

With Express 3, you can set global compiler options by using app.locals:

app.locals.uglify = true

which is the same as:

res.render view, { uglify: true }

See the compiler options for detailed information about all the available options.

After the installation you will have a haml-coffee binary that can be used to compile single templates and even compile multiple templates recursively into a single file.

$ haml-coffee
Usage: node haml-coffee

Options:
  -i, --input                        Either a file or a directory name to be compiled
  -o, --output                       Set the output filename
  -n, --namespace                    Set a custom template namespace
  -t, --template                     Set a custom template name
  -b, --basename                     Ignore file path when generate the template name
  -e, --extend                       Extend the template scope with the context
  -r, --render                       Render to standalone HTML

The following section describes only the options that are unique to the command line tool.

You can see all the available options by executing haml-coffee --help and have a look at the compiler options for detailed information about all the options.

The input and output are optional and you can also directly redirect the streams.

You can either specify a single template or a directory with the -i/--input argument. When you supply a directory, templates are being searched recursively:

$ haml-coffee -i template.haml

This will generate a template with the same name as the file but the extension changed to .jst. The above command for example would generate a template named template.jst.

A valid Haml Coffee template must have one of the following extensions: .haml, .html.haml, .hamlc or .html.hamlc.

You can specify a single output file name to be used instead of the automatic generated output file name with the -o/--output argument:

$ haml-coffee -i template.haml -o t.js

This creates a template named t.js. You can also set a directory as input and give an output file name for concatenating all templates into a single file:

$ haml-coffee -i templates -o all.js

This will create all the templates under the templates directory into a single, combined output file all.js.

Each template will register itself by default under the window.HAML namespace, but you can change the namespace with the -n/--namespace argument:

$ haml-coffee -i template.haml -n exports.JST

Each template must have a unique name under which it can be addressed. By default the template name is derived from the template file name by stripping off all extensions and remove illegal characters. Directory names are converted to nested namespaces under the default namespace. For example, a template named user/show-admin.html.haml will result in a template that can be accessed by window.HAML['user/show_admin'].

Given the -b/--basename argument, the deduced template name will not include the path to the template. For example, a template named user/show-admin.html.haml will result in a template that can be accessed by window.HAML['show_admin'] instead of window.HAML['user/show_admin'].

With the -t/--template argument you can set a template name manually:

$ haml-coffee -i template.haml -n exports.JST -t other

This will result in a template that can be accessed by exports.JST['other'].

By extending the template scope with the context, you can access your context data without @ or this:

%h2= title

This effect is achieved by using the with statement. Using with is forbidden in ECMAScript 5 strict mode.

You can use Haml Coffee on the command line to enter a template and stop it with Ctrl-D:

$ haml-coffee -p amd
%h1 Hello AMD
^D

which will output the AMD module source code to the console. You either have to set the placement option to amd or give it a template name like

$ haml-coffee -t name
%p JST rocks!
^D

which will output the JST source code. Now you can also redirect files like:

$ haml-coffee -t name < input.hamlc > output.jst

Haml Coffee implements the Haml Spec to ensure some degree of compatibility to other Haml implementations and the following sections are fully compatible to Ruby Haml:

• Plain text
• Multiline: |
• Element names: %
• Attributes: {} or ()
• Class and ID: . and #, implicit div elements
• Self-closing tags: /
• Doctype: !!!
• HTML comments: /, conditional comments: /[], Haml comments: -#
• Running CoffeeScript: -, inserting CoffeeScript: =
• CoffeeScript interpolation: #{}
• Whitespace preservation: ~
• Whitespace removal: > and <
• Escaping \
• Escaping HTML: &=, unescaping HTML: !=
• Filters: :plain, :javascript, :css, :cdata, :escaped, :preserve
• Boolean attributes conversion
• Haml object reference syntax: []

Please consult the official Haml reference for more details.

Haml Coffee supports both Ruby 1.8 and Ruby 1.9 style attributes. So the following Ruby 1.8 style attribute

%a{ :href => 'http://haml-lang.com/', :title => 'Haml home' } Haml

can also be written in Ruby 1.9 style:

%a{ href: 'http://haml-lang.com/', title: 'Haml home' } Haml

HTML style tags are also supported:

%a( href='http://haml-lang.com/' title='Haml home') Haml

Haml Coffee supports a small subset of the Ruby Haml helpers. The provided helpers will bind the helper function to the template context, so it isn't necessary to use =>.

Surrounds a block of Haml code with strings, with no whitespace in between.

!= surround '(', ')', ->
  %a{:href => "food"} chicken

produces the HTML output

(<a href='food'>chicken</a>)

Appends a string to the end of a Haml block, with no whitespace between.

click
!= succeed '.', ->
  %a{:href=>"thing"} here

produces the HTML output

click
<a href='thing'>here</a>.

Prepends a string to the beginning of a Haml block, with no whitespace between.

!= precede '*', ->
  %span.small Not really

produces the HTML output

*<span class='small'>Not really</span>

Haml Coffee supports object references, but they are implemented slightly different due to the underlying runtime and different code style for CoffeeScript.

Square brackets contain a CoffeeScript object or class that is used to set the class and id of that tag. The class is set to the object’s constructor name (transformed to use underlines rather than camel case) and the id is set to the object’s constructor name, followed by the value of its id property or its #to_key or #id functions (in that order). Additionally, the second argument (if present) will be used as a prefix for both the id and class attributes.

For example:

%div[@user, 'greeting']
  Hello

is compiled to:

<div class='greeting_user' id='greeting_user_15'>
  Hello!
</div>

If the user object is for example a Backbone model with the id of 15. If you require that the class be something other than the underscored object’s constructor name, you can implement the #hamlObjectRef function on the object:

:coffeescript
  class User
    id: 23
    hamlObjectRef: -> 'custom'

%div[new User()]
  Hello

is compiled to:

<div class='custom' id='custom_23'>
  Hello!
</div>

Haml Coffee supports currently a single directive that extends the Haml syntax.

You can use the +include directive to include another template:

%h1 Include
+include 'partials/test'

This will look up the specified template and include it. So if the partial partials/test contains

%p Partial content

The final result will be

<h1>Include</h1>
<p>Partial content</p>

Haml and CoffeeScript are a winning team, both use indention for blocks and are a perfect match for this reason. You can use CoffeeScript instead of Ruby in your Haml tags and the attributes.

It's not recommended to put too much logic into the template.

When you define an attribute value without putting it into quotes (single or double quotes), it's considered to be CoffeeScript code to be run at render time. By default, attributes values from CoffeeScript code are escaped before inserting into the document. You can change this behaviour by setting the appropriate compiler option.

HTML style attributes are the most limited and can only assign a simple variable:

%img(src='/images/demo.png' width=@width height=@height alt=alt)

Both the @width and @height values must be passed as locals when rendering the template and alt must be defined before the %img tag.

Ruby style tags can be more complex and can call functions:

%header
  %user{ :class => App.currentUser.get('status') }= App.currentUser.getDisplayName()

Attribute definitions are also supported in the Ruby 1.9 style:

%header
  %user{ class: App.currentUser.get('status') }= App.currentUser.getDisplayName()

More fancy stuff can be done when use interpolation within a double quoted attribute value:

%header
  %user{ class: "#{ if @user.get('roles').indexOf('admin') is -1 then 'normal' else 'admin' }" }= @user.getDisplayName()

But think twice about it before putting such fancy stuff into your template, there are better places like models, views or helpers to put heavy logic into.

You can define your attributes over multiple lines and the next line must not be correctly indented, so you can align them properly:

%input#password.hint{ type: 'password', name: 'registration[password]',
                      data: { hint: 'Something very important', align: 'left' } }

In the above example you also see the usage for generating HTML5 data attributes.

You can run any CoffeeScript code in your template:

- for project in @projects
  - if project.visible
    .project
      %h1= project.name
      %p&= project.description

There are several supported types to run your code:

• Run code without insert anything into the document: -
• Run code and insert the result into the document: =

All inserted content from running code is escaped by default. You can change this behaviour by setting the appropriate compiler option.

There are three variations to run code and insert its result into the document, two of them to change the escaping style chosen in the compile option:

• Run code and do not escape the result: !=
• Run code and escape the result: &=
• Preserve whitespace when insert the result: ~

Again, please consult the official Haml reference for more details. Haml Coffee implements the same functionality like Ruby Haml, only for CoffeeScript.

If you use the CoffeeScript interpolation without explicitly run code with = and -, the interpolation output is not being escaped. You can manually force escaping by using the HTML escape reference $e:

%p
  foo #{ $e '<bar>' }

will be rendered to

<p>
  foo &lt;bar&gt;
</p>

Running code must be placed on a single line and unlike Ruby Haml, you cannot stretch a it over multiple lines by putting a comma at the end.

However, you can use multiline endings | to stretch your code over multiple lines to some extend:

- links = {          |
    home: '/',       |
    docs: '/docs',   |
    about: '/about'  |
  }                  |

%ul
  - for name, link of links
    %li
      %a{ href: link }= name

Please note, that since the line is concatenated before the compilation, you cannot omit the curly braces and the commas in the above example, like you'd do in normal CoffeeScript code. Therefore it's recommended to use the CoffeeScript filter to have real multiline code blocks:

:coffeescript
  links =
    home: '/'
    docs: '/docs'
    about: '/about'

%ul
  - for name, link of links
    %li
      %a{ href: link }= name

You can also create functions that generate Haml:

- sum = (a, b) ->
  %div
    %span= a
    %span= b
    %span= a+b
= sum(1,2)
= sum(3,4)

or pass generated HTML output through a function for post-processing.

= postProcess ->
  %a{ href: '/' }

The content of the :coffeescript filter is run when the template is rendered and doesn't output anything into the resulting document. This comes in handy when you have code to run over multiple lines and don't want to prefix each line with -:

%body
  :coffeescript
    tags = ['CoffeeScript', 'Haml']
    project = 'Haml Coffee'
  %h2= project
  %ul
    - for tag in tags
      %li= tag

The following section describes all the available compiler options that you can use through the JavaScript API, as Express view option or as argument to the command line utility.

The command line arguments may be slightly different. For example instead of passing --escape-html=false you have to use the --disable-html-escaping argument. You can see a list of all the command line arguments by executing haml-coffee --help.

The HTML options change the way how the generated HTML will look like.

• Name: 'format'
• Type: String
• Default: html5

The Haml parser knows different HTML formats to which a given template can be rendered and it must be one of:

• xhtml
• html4
• html5

Doctype, self-closing tags and attributes handling depends on this setting. Please consult the official Haml reference for more details.

• Name: uglify
• Type: Boolean
• Default: false

All generated HTML tags are properly indented by default, so the output looks nice. This can be helpful when debugging. You can skip the indention by setting the uglify option to false. This save you some bytes and you'll have increased rendering speed.

• Name: escapeHtml
• Type: Boolean
• Default: true

The reserved HTML characters ", ', &, < and > are converted to their HTML entities by default when they are inserted into the HTML document from evaluated CoffeeScript.

You can always change the escaping mode within the template to either force escaping with &= or force unescaping with !=.

• Name: escapeAttributes
• Type: Boolean
• Default: true

All HTML attributes that are generated by evaluating CoffeeScript are also escaped by default. You can turn of HTML escaping of the attributes only by setting escapeAttributes to false. You can't change this behaviour in the template since there is no Haml markup for this to instruct the compiler to change the escaping mode.

• Name: cleanValue
• Type: Boolean
• Default: true

Every output that is generated from evaluating CoffeeScript code is cleaned before inserting into the document. The default implementation converts null or undefined values into an empty string and marks real boolean values with a hidden marker character. The hidden marker character is necessary to distinguish between String values like 'true', 'false' and real boolean values true, false in the markup, so that a boolean attribute conversion can quickly convert these values to the correct HTML5/XHTML/HTML4 representation.

• Name: preserve
• Type: String
• Default: textarea,pre

The preserve option defines a list of comma separated HTML tags that are whitespace sensitive. Content from these tags must be preserved, so that the indention has no influence on the displayed content. This is simply done by converting the newline characters to their equivalent HTML entity.

• Name: autoclose
• Type: String
• Default: meta,img,link,br,hr,input,area,param,col,base

The autoclose option defines a list of tag names that should be automatically closed if they have no content.

• Name: placement
• Type: String
• Default: global

The placement option defines where the template function is inserted upon compilation.

Possible values are:

• global
  Inserts the optionally namespaced template function into window.HAML.

• 'standalone'
  Returns the template function without wrapping it

• amd
  Wraps the template function into a define() statement to allow async loading via AMD.

See AMD support for more information.

• Name: dependencies
• Type: Object
• Default: { hc: 'hamlcoffee' }

The dependencies option allows you to define the modules that must be required for the AMD template define function. The object key will be the function parameter name of the module the object value defines. See AMD support for more information.

• Name: hyphenateDataAttrs
• Type: Boolean
• Default: true

Convert underscores to hyphens for data attribute keys, see the Ruby Haml reference.

Haml Coffee provides helper functions for HTML escaping, value cleaning and whitespace preservation, which must be available at render time. By default every generated template function is self-contained and includes all of the helper functions.

However you can change the reference to each helper function by providing the appropriate compiler option and there are good reasons to do so:

• You want to reduce the template size and provide all the helpers from a central place.
• You want to customize a helper function to better fit your needs.

To change these functions, simply assign the new function name to one of the following options:

• customHtmlEscape: Escape the reserved HTML characters into their equivalent HTML entity.
• customPreserve: Converting newlines into their HTML entity.
• customFindAndPreserve: Find whitespace sensitive tags and preserve their content.
• customCleanValue: Clean the value that is returned after evaluating some inline CoffeeScript.
• customSurround: Surrounds a block of Haml code with strings, with no whitespace in between.
• customSucceed: Appends a string to the end of a Haml block, with no whitespace between.
• customPrecede: Prepends a string to the beginning of a Haml block, with no whitespace between.
• customReference: Creates the Haml object reference.

The customSurround, customSucceed and customPrecede are bound to the template context.

You can find a default implementation for all these helper functions in Haml Coffee Assets.

• Global dependencies
• Trivial dependency detection

Haml Coffee has built in AMD support by setting the placement option to amd. This will generate a module definition for the JavaScript template. The dependencies options can be used to provide a mapping of module names to parameters. To illustrate this, the default value will result in the following module declaration:

define ['hamlcoffee'], (hc) ->

When the template contains a require call in the form of

 - require 'module'
 - require 'deep/nested/other'

it will be added to the module definition list

define ['hamlcoffee', 'module', 'deep/nested/other'], (hc, module, other) ->

allowing you to render a partial template:

!= module()
!= other()

Of course the require call can have different quotes or parenthesises, allowing you to directly require and render:

!= require("another/other")()

By default Haml Coffee AMD templates depend on the hamlcoffee module that provides the client side helpers needed to render the template. You need to supply your own module, but you can grab a copy from Haml Coffee Assets AMD helpers and adapt it to your needs. Another option is to remove the module from the module dependencies, but that's usually not what you want because it duplicates the needed function within every template.

By default Haml Coffee outputs pre-compiled templates for rendering dynamic content.

There is a render helper that can be used as follows to get the standalone HTML rendering of a HAML template.

hamlc = require 'haml-coffee'
text = hamlc.render '%h1= @title', {title: 'hi'}
text
'<h1>hi</h1>'

Haml Coffee uses Grunt for development, which you can install with NPM:

$ npm install

and run Grunt to automatically run the Jasmine specs on file modification:

$ grunt

Feel free to take a look at the crispy changelog instead of crawling through the commit history.

Haml Coffee in the Rails asset pipeline:

• haml-coffee-assets

• Michael Kessler (@netzpirat, FlinkFinger)
• Sebastion Deutsch (@sippndipp, 9elements)
• Jan Varwig (@agento, 9elements)

See all contributors on the contributor page.

(The MIT License)

Copyright (c) 2011 9elements, 2011-2013 Michael Kessler

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.