This is a liquid implementation for both Node.js and browsers. Website: http://harttle.github.io/liquidjs/, Live Demo: https://jsfiddle.net/6u40xbzs/
Features
- Fully compatible to shopify, with all tags and filters implemented
- Support layout(extend) and include syntax
- In pure JavaScript with any-promise as the only one dependency
Differences
Though being compatible with Ruby Liquid is one of our priorities, there're still certain differences. You may need some configuration to get it compatible in these senarios:
- Dynamic file locating (enabled by default), which means layout/partial name can be an variable in liquidjs. See #51.
- Truthy and Falsy. All values except
undefined
,null
,false
are truthy, whereas in Ruby Liquid all exceptnil
andfalse
are truthy. See #26. - Number Rendering. Since JavaScript do not distinguish
float
andinteger
, we cannot either convert between them nor render regarding to their type. See #59.
- Usage
- API Spec
Install as Node.js dependency:
npm install --save liquidjs
Parse and Render:
var Liquid = require('liquidjs');
var engine = Liquid();
engine
.parseAndRender('{{name | capitalize}}', {name: 'alice'})
.then(console.log);
// outputs 'Alice'
Caching templates:
var tpl = engine.parse('{{name | capitalize}}');
engine
.render(tpl, {name: 'alice'})
.then(console.log);
// outputs 'Alice'
var engine = Liquid({
root: path.resolve(__dirname, 'views/'), // dirs to lookup layouts/includes
extname: '.liquid' // the extname used for layouts/includes, defaults ""
});
engine.renderFile("hello.liquid", {name: 'alice'})
.then(console.log) // outputs "Alice"
// which is equivalent to:
engine
.renderFile("hello", {name: 'alice'})
.then(console.log) // outputs "Alice"
// register liquid engine
app.engine('liquid', engine.express());
app.set('views', './views'); // specify the views directory
app.set('view engine', 'liquid'); // set to default
Here's an Express demo. When used with Express.js,
Express views
will be included when looking up
partials(includes and layouts).
You can get a dist file for browsers from
- Releases page for liquidjs, or
- unpkg.com: https://unpkg.com/liquidjs/dist/liquid.min.js
Here's the demo:
- JSFiddle: https://jsfiddle.net/6u40xbzs/
- Demo directory: /demo/browser/.
Note: For IE and Android UC browser, you will need a Promise polyfill.
// file: color.liquid
color: '{{ color }}' shape: '{{ shape }}'
// file: theme.liquid
{% assign shape = 'circle' %}
{% include 'color' %}
{% include 'color' with 'red' %}
{% include 'color', color: 'yellow', shape: 'square' %}
The output will be:
color: '' shape: 'circle'
color: 'red' shape: 'circle'
color: 'yellow' shape: 'square'
// file: default-layout.liquid
Header
{% block content %}My default content{% endblock %}
Footer
// file: page.liquid
{% layout "default-layout" %}
{% block content %}My page content{% endblock %}
The output of page.liquid
:
Header
My page content
Footer
- It's possible to define multiple blocks.
- block name is optional when there's only one block.
The full list of options for Liquid()
is listed as following:
-
root
is a directory or an array of directories to resolve layouts and includes, as well as the filename passed in when calling.renderFile()
. If an array, the files are looked up in the order they occur in the array. Defaults to["."]
-
extname
is used to lookup the template file when filepath doesn't include an extension name. Eg: setting to".html"
will allow including file by basename. Defaults to""
. -
cache
indicates whether or not to cache resolved templates. Defaults tofalse
. -
dynamicPartials
: if set, treat<filepath>
parameter in{%include filepath %}
,{%layout filepath%}
as a variable, otherwise as a literal value. Defaults totrue
. -
strict_filters
is used to enable strict filter existence. If set tofalse
, undefined filters will be rendered as empty string. Otherwise, undefined filters will cause an exception. Defaults tofalse
. -
strict_variables
is used to enable strict variable derivation. If set tofalse
, undefined variables will be rendered as empty string. Otherwise, undefined variables will cause an exception. Defaults tofalse
. -
trim_tag_right
is used to strip blank characters (including\t
, and\r
) from the right of tags ({% %}
) until\n
(inclusive). Defaults tofalse
. -
trim_tag_left
is similiar totrim_tag_right
, whereas the\n
is exclusive. Defaults tofalse
. See Whitespace Control for details. -
trim_value_right
is used to strip blank characters (including\t
, and\r
) from the right of values ({{ }}
) until\n
(inclusive). Defaults tofalse
. -
trim_value_left
is similiar totrim_value_right
, whereas the\n
is exclusive. Defaults tofalse
. See Whitespace Control for details. -
greedy
is used to specify whethertrim_left
/trim_right
is greedy. When set totrue
, all consecutive blank characters including\n
will be trimed regardless of line breaks. Defaults totrue
.
// Usage: {{ name | uppper }}
engine.registerFilter('upper', v => v.toUpperCase())
Filter arguments will be passed to the registered filter function, for example:
// Usage: {{ 1 | add: 2, 3 }}
engine.registerFilter('add', (initial, arg1, arg2) => initial + arg1 + arg2)
See existing filter implementations here: https://github.com/harttle/liquidjs/blob/master/filters.js
// Usage: {% upper name%}
engine.registerTag('upper', {
parse: function(tagToken, remainTokens) {
this.str = tagToken.args; // name
},
render: function(scope, hash) {
var str = Liquid.evalValue(this.str, scope); // 'alice'
return Promise.resolve(str.toUpperCase()); // 'Alice'
}
});
parse
: Read tokens fromremainTokens
until your end token.render
: Combine scope data with your parsed tokens into HTML string.
See existing tag implementations here: https://github.com/harttle/liquidjs/blob/master/tags/