Permalinks plugin for Assemble, the static site generator for Grunt.js, Yeoman and Node.js. This plugin enables powerful and configurable URI patterns, Moment.js for parsing dates, much more.
(TOC generated by verb using markdown-toc)
Install with npm:
$ npm install --save grunt-assemble-permalinks
Also see the Gruntfile for example usage.
From the same directory as your project's Gruntfile and package.json, install this plugin with the following command:
npm install grunt-assemble-permalinks --save-dev
Once that's done, just add permalinks
, the name of this module, to the plugins
option in the Assemble task:
module.exports = function(grunt) {
// Project configuration.
grunt.initConfig({
assemble: {
options: {
plugins: ['grunt-assemble-permalinks', 'other/plugins/*'],
permalinks: {
structure: ':year/:month/:day/:foo/index.html'
}
},
files: {'blog/archives/': ['archives/*.hbs']}
}
});
grunt.loadNpmTasks('grunt-assemble');
grunt.registerTask('default', ['assemble']);
};
If everything was installed and configured correctly, you should be ready to go!
Replacement patterns for dynamically constructing permalinks, as well as the corresponding directory structures.
Permalinks are appended to the dest directory. So given this config:
{
assemble: {
blog: {
options: {
permalinks: {
structure: ':year/:month/:day/:basename:ext'
}
},
files: {
'blog/archives/': ['archives/*.hbs']
}
}
}
}
// the generated directory structure and resulting path would look something like:
//=> 'blog/archives/2011/01/01/an-inspiring-post.html'
This plugin comes with a number of built-in replacement patterns that will automatically parse and convert the built-in variables into the appropriate string. Since Assemble provides a number of generic variables for accessing page data, such as basename
, ext
, filename
etc., this plugin simply dynamically builds the replacement patterns from those generic variables.
Barring a few exceptions (_page
, data
, filePair
, page
, pageName
), you should be able to use any valid variable that is on the page context in your replacement patterns.
For example, assuming we have a file, ./templates/overview.hbs
:
:ext
: would result in thedest
extension:.html
:extname
: alias for:ext
.:basename
: would result inoverview
:filename
: would result in the dest file name,overview.html
:pagename
: alias for:filename
.:category
: Slugified version of the very first category for a page.
A few special replacement patterns were created for this lib.
Automatically adds sequential, "padded" numbers based on the length of the pages array in the current target.
For example, given the structure :num-:basename
:
- 1-9 pages would result in
1-foo.html
,2-bar.html
,3-baz.html
and so on. - 1,000 pages would result in
0001-foo.html
,0002-bar.html
,0003-baz.html
, ...1000-quux.html
.
Adds sequential digits. Similar to :num
, but the number of digits is determined by the number of zeros defined.
Example:
:00
will result in two-digit numbers:000
will result in three-digit numbers:00000000
will result in eight-digit numbers, and so on...
Adds randomized characters based on the pattern provided in the parentheses. The first parameter defines the pattern you wish to use, and an optional second parameter defines the number of characters to generate.
For example, :random(A, 4)
(whitespace insenstive) would result in randomized 4-digit uppercase letters, like, ZAKH
, UJSL
... and so on.
no second parameter
If a second parameter is not provided, then the length()
of the characters used in the first parameter will be used to determine the number of digits to output. For example:
:random(AAAA)
is equivelant to:random(A, 4)
:random(AAA0)
and:random(AA00)
and:random(A0A0)
are equivelant to:random(A0, 4)
valid characters (and examples)
:random(aa)
: results in double-digit, randomized, lower-case letters (abcdefghijklmnopqrstuvwxyz
):random(AAA)
: results in triple-digit, randomized, upper-case letters (ABCDEFGHIJKLMNOPQRSTUVWXYZ
):random(0, 6)
: results in six-digit, randomized nubmers (0123456789
):random(!, 5)
: results in single-digit randomized, valid non-letter characters (~!@#$%^&()_+-={}[];\',.
):random(A!a0, 9)
: results in nine-digit, randomized characters (any of the above)
The order in which the characters are provided has no impact on the outcome.
If you have some patterns you'd like to implement, if you think they're common enough that they should be built into this plugin, please submit a pull request.
Adding patterns is easy, just add a patterns: []
property to the permalinks
option, then add any number of patterns to the array. For example, let's say we want to add the :project
variable to our permalinks:
options: {
permalinks: {
structure: ':year/:month/:day/:project/:slug:ext',
patterns: []
}
}
...
Since :project
is not a built-in variable, we need to add a replacement pattern so that any permalinks that include this variable will actually work:
options: {
permalinks: {
structure: ':year/:month/:day/:project/:slug:ext',
patterns: [
{
pattern: ':project',
replacement: '<%= pkg.name %>'
}
]
}
}
...
Any string pattern is acceptable, as long as a :
precedes the variable, but don't forget that there must also be a matching property in the context or Assemble will might an error (or worse, not). In other words, when you add a replacement pattern for :foo
, it's good practice to make sure this property exists:
---
foo: bar
slug:
---
This plugin uses the incredibly feature rich and flexible moment.js for parsing dates. Please consult the moment.js documentation for usage information and for the full list of available options.
For the date variables to work, a date
property must exist on the page object.
---
date: 2014-01-29 3:45 PM
---
Or
pages: [
{
data: {
title: 'All about permalinks, the novel.',
description: 'This rivoting sequel to War & Peace will have you sleeping in no time.'
date: '2014-01-29 3:45 PM'
},
content: ""
}
]
:year
: The year of the date, four digits, for example2014
:month
: Month of the year, for example01
:day
: Day of the month, for example13
:hour
: Hour of the day, for example24
:minute
: Minute of the hour, for example01
:second
: Second of the minute, for example59
For the following examples, let's assume we have a date in the YAML front matter of a page formatted like this:
---
date: 2014-01-29 3:45 PM
---
(note that this property doesn't have to be in YAML front matter, it just needs to be in the page.data
object, so this works fine with options.pages
collections as well.)
:date
: Eqivelant to the full date:YYYY-MM-DD
. Example:2014-01-29
:YYYY
: The full year of the date. Example:2014
:YY
: The two-digit year of the date. Example:14
:year
: alias forYYYY
:MMMM
: The full name of the month. ExampleJanuary
.:MMM
: The name of the month. Example:Jan
:monthname
: alias forMMMM
:MM
: The double-digit number of the month. Example:01
:M
: The single-digit number of the month. Example:1
:month
: alias forMM
:mo
: alias forM
:day
: alias forDD
:DD
: The double-digit day of the month. Example:29
:D
: The double-digit day of the month. Example:29
:dddd
: Day of the week. Example:monday
:ddd
: Day of the week. Example:mon
:dd
: Day of the week. Example:Mo
:d
: Day of the week. Example:2
:HH
: The double-digit time of day on a 24 hour clock. Example15
:H
: The single-digit time of day on a 24 hour clock. Example3
:hh
: The double-digit time of day on a 12 hour clock. Example03
:h
: The single-digit time of day on a 12 hour clock. Example3
:hour
: alias forHH
:mm
: Minutes. Example:45
.:m
: Minutes. Example:45
.:min
: Alias formm
|m
.:minute
: Alias formm
|m
.
:ss
: Seconds. Example:09
.:s
: Seconds. Example:9
.:sec
: Alias forss
|s
.:second
: Alias forss
|s
.
Note that this plugin does not currently modify actual links inside pages, so that will need to be addressed separately. I'm also willing to look at options for incorporating that into this plugin.
Type: String
Default: undefined
The permalink pattern to use for building paths and generated files. Permalink structures are appended to the dest
defined for the current target.
For example, let's say we use the following pattern on a few blog posts: foo.hbs
, bar.hbs
, and baz.hbs
:
options: {
permalinks: {
structure: ':year/:month/:day/:basename:ext',
},
files: {
'./blog/': ['./templates/blog/*.hbs']
}
}
// results in
// => './blog/2014/01/01/foo.html'
// => './blog/2014/01/01/bar.html'
// => './blog/2014/01/01/baz.html'
Note that permalink structures will be ignored for files with the basename index
. See Issue #20 for more info.
Type: String
Default: undefined
The following presets are currently available:
pretty
: expands to:basename/index:html
.dayname
: expands to:YYYY/:MM/:DD/:basename/index:ext
.monthname
: expands to:YYYY/:MM/:basename/index:ext
.
In a nutshell, a preset is simply a pre-defined permalink structure
, so instead of having to type out :foo/:bar/:baz/basename:html
, you can just use pretty
. Presets expand into permalink structures following this pattern:
dest + preset
//=> dest + :bar/index:html
Additionally, if a structure
is also defined, the preset
will be appended to it.
dest + structure + preset
//=> dest + :foo + :bar/index:html
If you would like to see another preset, please submit an issue.
Type: Array
Default: ["YYYY-MM-DD"]
Array of custom date formats for Moment.js to use for parsing dates.
options: {
permalinks: {
dateFormats: ["YYYY-MM-DD", "MM-DD-YYYY", "YYYY-MM-DDTHH:mm:ss.SSS"]
},
files: {
...
}
}
Type: String
Default: en
Set the "global" language for Moment.js to use for converting dates:
options: {
permalinks: {
structure: ':year/:month/:day/:basename:ext',
lang: 'fr'
}
files: {
'blog/': ['templates/blog/*.hbs']
}
}
...
//=> blog/2013/mars/13/my-post.html
Type: Array
Default: ['_page', 'data', 'filePair', 'page', 'pageName']
Properties to omit from the context for processing replacement patterns. I wanted to use this for omitting the default properties, but I decided to expose this as an option in case it comes in useful to someone else.
options: {
permalinks: {
exclusions: ["foo", "bar"],
},
files: {
...
}
}
Pretty links involve saving an index.html
to each directory, with the tile, file name, slug, or some other variable as the :basename
of the directory. For example:
assemble: {
blog: {
options: {
permalinks: {
structure: ':basename/:index.html'
}
},
files: [
{expand: true, cwd: 'templates/', src: ['*.hbs'], dest: 'blog/', ext: '.html'}
]
}
}
which results in something like:
dest + /my-node-js-post/index.html
dest + /my-javascript-post/index.html
dest + /my-assemble-post/index.html
Also see the Gruntfile for example usage.
Presets allow you to achieve certain permalinks structures without having to explicitly define each URL segment. For example, in the previous example we created pretty URLs., Here is how we would do the same with presets
:
options: {
permalinks: {
preset: 'pretty'
},
files: {
'./blog/': ['./templates/blog/*.hbs']
}
}
The above example won't necessarily save a whole lot of time, but it's a nice way of ensuring that you're getting pretty links with whatever permalinks structure you define. To some, this might be particularly useful when "stacked" with more complex permalink structures, e.g.:
options: {
permalinks: {
preset: 'pretty',
structure: ':archives/:categories',
},
files: {
'./blog/': ['./templates/blog/*.hbs']
}
}
which expands to: ./blog/:archives/:categories/:basename:/index:ext
, and would result in:
./blog/archives/categories/foo/index.html
Also see the Gruntfile for example usage.
In most cases your generated HTML will have the .html
extension, then using :index.html
is probably fine. But if you happen to switch back and forthing between projects that alternate between .htm
and .html
, you can use :index:ext
instead.
Also see the Gruntfile for example usage.
You don't have to use slashes (/
) only in your permalinks, you can use -
or _
wherever you need them as well. For example, this is perfectly valid:
:YYYY_:MM-:DD/:slug:category:foo/:bar/index.html
Warning, this should be obvious, but make sure not to use a .
in the middle of your paths, especially if you use Windows.
Also see the Gruntfile for example usage.
You can even dynamically build up strings using Lo-Dash templates:
---
date: 1-1-2014
# Dynamically build the slug for example
area: business
section: finance
slug: <%= area %>-<%= section %>
---
With this config:
blog: {
options: {
permalinks: {
structure: ':year/:month/:day/:slug/:title.html'
}
},
files: {
'blog/': ['posts/*.hbs']
}
}
Would render to:
blog/2014/01/01/business-finance/index.html
Also see the Gruntfile for example usage.
Keep in mind that the date is formatted the way you want it, you don't need to follow these examples. Also, some of these variables will only work if you add that property to your pages, and setup the replacement patterns.
:YYYY/:MM/:DD/news/:id/index:ext
//=> dest + '/2014/01/01/news/001/index.html'
:YYYY/:MM/:DD/:mm/:ss/news/:id/index:ext
//=> dest + '/2014/01/01/40/16/news/001/index.html'
:year/:month/:day/:basename:ext
//=> dest + '/2014/01/01/my-post.html'
blog/:year-:month-:day/:basename:ext
//=> dest + 'blog/2014-01-01/my-post.html'
:date/:basename:ext
//=> dest + '2014-01-01/my-post.html'
:year/:month/:day/:category/index.html
//=> dest + '/2014/01/01/javascript/index.html'
:year/:month/:day/:slug/index.html
//=> dest + '/2014/01/01/business-finance/index.html'
Also see the Gruntfile for example usage.
Permalinks are important for SEO. but you should spend some time thinking about the strategy you want to use before you decide on a URL structure.
Yep, that's what I said. There are plenty of valid use cases for using date-based URL's. This plugin offers a number of date-based patterns, and we leverage Moment.js a lot. Still, I recommend that you avoid using a date-based permalink structure for your blog or documentation, because there is a good chance it will do more harm than good over the long term.
Date-based URL's tend to decrease click through rates on older articles. Think about it, who prefers reading out of date content? So use a URL strategy that doesn't go out of its way to emphasize the date, and you'l keep your posts feeling like fresh content.
Numeric or :id
based permalinks are better than date-based, but they don't really offer much usability or SEO benefit.
The best structure is one that:
- provides the highest degree of semantic relevance to the content, and
- is useful to both search engines and humans
Here are some example permalink structures, pick the one you like or feel free to use something else:
:author
:category/:author
Since the :author
variable isn't actually built in, you'll need to add it as a custom replacement pattern. But you could use :filename
, :pagename
, :basename
and so on. The important thing to remember is that the name counts.
If you need to use a custom variable, such as :author
or :title
, just add it like this:
var _str = require('underscore.string');
{
assemble: {
options: {
permalinks: {
structure: ':author:ext',
patterns: [
{
pattern: ':author',
replacement: '<%= _str.slugify(pkg.author.name) %>'
}
]
}
},
files: {},
}
}
- grunt-assemble-anchors: Assemble plugin for creating anchor tags from headings in generated html using Cheerio.js. | homepage
- grunt-assemble-contextual: Generates a JSON file with the context of each page. Basic plugin to help see… more | homepage
- grunt-assemble-decompress: Assemble plugin for extracting zip, tar and tar.gz archives. | homepage
- grunt-assemble-download: Assemble plugin for downloading files from GitHub. | homepage
- grunt-assemble-i18n: Plugin for adding i18n support to Assemble projects. | homepage
- grunt-assemble-lunr: Assemble plugin for adding search capabilities to your static site, with lunr.js. | homepage
- grunt-assemble-navigation: Assemble navigation plugin. Automatically generate Bootstrap-style, multi-level side nav. See the sidenav on assemble.io for… more | homepage
- grunt-assemble-permalinks: Permalinks plugin for Assemble, the static site generator for Grunt.js, Yeoman and Node.js. This plugin… more | homepage
- grunt-assemble-sitemap: Sitemap plugin for Assemble | homepage
- grunt-assemble-toc: Assemble middleware for adding a Table of Contents (TOC) to any HTML page. | homepage
- grunt-assemble-wordcount: Assemble plugin for displaying wordcount and average reading time to blog posts or pages. | homepage
- grunt-assemble: Static site generator for Grunt.js, Yeoman and Node.js. Used by Zurb Foundation, Zurb Ink, H5BP/Effeckt… more | homepage
Pull requests and stars are always welcome. For bugs and feature requests, please create an issue.
(This document was generated by verb-generate-readme (a verb generator), please don't edit the readme directly. Any changes to the readme must be made in .verb.md.)
To generate the readme and API documentation with verb:
$ npm install -g verb verb-generate-readme && verb
Install dev dependencies:
$ npm install -d && npm test
Jon Schlinkert
Copyright © 2016, Jon Schlinkert. Released under the MIT license.
This file was generated by verb, v0.9.0, on July 19, 2016.