Serif is a static site generator and blogging system powered by markdown files.

To get started with Serif based on a site skeleton:

gem install serif     # install serif
cd path/to/some/place # go to where you'll be creating your site directory
serif new             # create an initial site skeleton

# ... edit your files how you want them ...

serif generate        # generate the site based on the source files
serif dev             # serve up the site for local testing purposes

Now visit http://localhost:8000/ to view the site.

• Basic usage
• Content and site structure
• Publishing drafts
• Updating posts
• Archive pages
• Configuration
• Deploying
• Custom tags and filters
• Template variables
• License and contributing
• Developing Serif
• Changes and what's new

Installation is via RubyGems. If you don't have Ruby installed and are new to Ruby, I recommend using RVM.

$ gem install serif

$ cd path/to/site/directory
$ serif generate

This runs a very simple web server that is mainly designed to test what the site will look like and let you make changes to stuff like CSS files without having to regenerate everything. Changes to post content will not be detected (yet).

$ cd path/to/site/directory
$ serif dev

Once this is run, visit http://localhost:8000.

You can generate a skeletal directory to get you going, using the new command.

$ cd path/to/site/directory
$ serif new

The structure of a Serif site is something like this:

.
├── _site
├── _layouts
│   └── default.html
├── _drafts
│   ├── some-draft
│   └── another-unfinished-post
├── _posts
│   ├── 2012-01-01-a-post-you-have-written
│   ├── 2012-02-28-another-post
│   └── 2012-03-30-and-a-third
├── _templates
│   ├─── post.html
│   └─── archive_page.html
├── _trash
├── _config.yml
├── css
│   └── ...
├── js
│   └── ...
├── images
│   └── ...
├── 404.html
├── favicon.ico
├── feed.xml
└── index.html

This is where generated content gets saved. You should serve files out of here, be it with Nginx or Apache. You should assume everything in this directory will get erased at some point in future. Don't keep anything in it!

This is where layouts for the site go. The file default.html is used by default, and individual files can override this by setting a Layout: foo header, which will use _layouts/foo.html instead.

Drafts go in _drafts, posts go in _posts. Simple enough.

Posts must have filenames in the format of YYYY-MM-DD-your-post. Drafts do not have a date part, since they're drafts and not published.

All files in these directories are assumed to be written in Markdown, with simple HTTP-style headers. The Markdown renderer is Kramdown, with Smarty for punctuation tweaks, and Pygments to get syntax highlighting for GitHub-flavoured fenced code blocks (although you'll need your own CSS).

Here's an example post:

Title: A title of a post
Created: 2012-01-01T14:30:00+00:00

Something something.

1. A list
2. Of some stuff
3. Goes here

End of the post

The headers are similar to Jekyll's YAML front matter, but here there are no formatting requirements beyond Key: value pairs. Header names are case-insensitive (so title is the same as Title), but values are not.

File headers like Title: Some title can contain any header, but certain headers have special meaning in Serif.

Note that while it is possible for you to manually handle timestamp values, it is recommended that you rely on using the value of now for Update and Publish.

If you change the permalink value for a published post, you will break any inbound URLs as well as, e.g., any feeds that rely on the URL as a unique persistent ID.

Two templates are available:

• post.html, which will be used to render individual post pages.
• archive_page.html, which will be used to render individual archive pages.

Both must be valid Liquid templates.

Deleted drafts go in here just in case you want them back.

Used for configuration settings.

Here's a sample configuration:

permalink: /blog/:title
archive:
  enabled: yes
  url_format: /blog/:year/:month

If a permalink setting is not given in the configuration, the default is /:title. There are the following options available for permalinks:

NOTE: if you change the permalink value, you will break existing URLs for published posts, in addition to, e.g., any feed ID values that depend on the post URL never changing.

Any other file in the directory's root will mostly be copied over exactly as-is.

An exception is any file ending in .html or .xml. These files are assumed to contain Liquid markup and will be processed as such. Header values will be available on page.

For example, this would work as an about.html:

<h1>All about me</h1>
<p>Where do I begin? {{ 'Well...' }}</p>

And so would this:

x: y

<h1>All about me</h1>
<p>Where do I begin? Well... First, x is {{ page.x }}.</p>

If you have a file like feed.xml that you wish to not be contained within a layout, specify layout: none in the header for the file:

layout: none

<?xml version="1.0" encoding="utf-8"?>
<!-- ... -->

To automatically publish a draft, add a publish: now header to the draft:

title: A draft that will be published
publish: now

This is a draft that will be published now.

On the next site generation (serif generate) this draft will be automatically published, using the current time as the creation timestamp.

When you update a post, you need to remember to change the updated time. As luck would have it, Serif takes care of timestamps for you! Just use a header of update: now at the top of your published post after making your changes:

title: My blog post
Created: 2013-01-01T12:01:30+00:00
update: now

Now the next time the site is generated, the timestamp will be updated:

title: My blog post
Created: 2013-01-01T12:01:30+00:00
Updated: 2013-03-18T19:03:30+00:00

By default, archive pages are made available at /archive/:year/month, e.g., /archive/2012/11. Individual archive pages can be customised by editing the _templates/archive_page.html file, which is used for each month.

Within the archive_page.html template, you have access to the variables month, which is a Ruby Date instance, and posts for the posts within that month.

To disable archive pages, or configure the URL format, see the section on configuration.

To link to archive pages, there is a site.archive template variable available in all pages. The structure of site.archive is a nested map starting at years:

{
  "posts" => [...],
  "years" => [
    {
      "date" => Date.new(2012),
      "posts" => [...],
      "months" => [
        { "date" => Date.new(2012, 12), "archive_url" => "/archive/2012/12", "posts" => [...] },
        { "date" => Date.new(2012, 11), "archive_url" => "/archive/2012/11", "posts" => [...] },
        ...
      ]
    }
  ]
}

Using site.archive, you can iterate over years, then iterate over months and use archive_url to link to the archive page for that given month within the year.

Configuration goes in _config.yml and must be valid YAML. Here's a sample configuration with available options:

permalink: /posts/:title
archive:
  enabled: yes
  url_format: /archive/:year/:month

permalink is the URL format for individual post pages. The default permalink value is /:title. For an explanation of the format of permalinks, see above.

archive contains configuration options concerning archive pages. enabled can be used to toggle whether archive pages are generated. If set to no or false, no archive pages will be generated. By default, this value is yes.

The archive url_format configuration option is the format used for archive pages. The default value is /archive/:year/:month. This must include both year and month. Visiting, e.g., /archive/2012/11 would render posts made in November 2012. See the section on archive pages above for more details.

To serve the site, set any web server to use /path/to/site/directory/_site as its root. NOTE: URLs generated in the site do not contain .html "extensions" by default, so you will need a rewrite rule. Here's an example rewrite for nginx:

error_page 404 @not_found_page;

location / {
	index  index.html index.htm;

	try_files $uri.html $uri $uri/ =404;
}

location @not_found_page {
	rewrite .* /404.html last;
}

Use ENV=production serif generate to regenerate the site for production.

These tags can be used in templates, in addition to the standard Liquid filters and tags. For example:

{{ post.title | smarty }}

{{ post.content | markdown }}

{% file_digest foo.css prefix:- %}

• markdown

  e.g., {{ post.content | markdown }}.

  This runs the given input through a Markdown + SmartyPants renderer, with fenced codeblocks enabled.

• smarty

  e.g., {{ post.title | smarty }}.

  This runs the given input through a SmartyPants processor, so quotes, dashes and ellipses come out better. Note that the markdown filter already does SmartyPants processing.

• strip

  Strips trailing and leading whitespace.

  e.g., {{ " hello " | strip }} will render as hello.

• xmlschema

  e.g., {{ post.created | xmlschema }}.

  Takes a Time value and returns an ISO8601-format string, as per Ruby's Time#xmlschema definition.

  Example output: 2013-02-16T23:55:22+01:00

  If the time value is in UTC: 2013-02-16T22:55:22Z

• encode_uri_component

  e.g., example.com/foo?url=http://mysite.com{{ post.url | encode_uri_component }}

  Intended to provide the functionality of JavaScript's encode_uri_component() function. Essentially: encodes the entire input so it is usable as a query string parameter.

• file_digest <path> [prefix:<prefix>]

  (Note: For this tag to return anything, ENV=production must be set as an environment variable.)

  Computes a hex digest of the contents of <path>, optionally prefixed with <prefix>. <path> is delimited by whitespace.

  Useful for URL fingerprinting for long-lived caching.

In addition to those mentioned above, such as the archive page variables, there are others. This may not be an exhaustive list.

These should be available in any template:

• {{ site }} --- a container for the site itself, containing:
  • {{ site.posts }} --- the published posts of the site
  • {{ site.latest_update_time }} --- a Ruby Time instance for the latest time that any post was updated. Useful for RSS/Atom feeds.
  • {{ site.archives }} --- a nested hash structure that groups posts by month. See above for how to use it.
• {{ draft_preview }} -- Set to true if this is part of generating a draft preview.
• {{ post_page }} -- Set to true if this is part of generating a regular published post.

A published post has variables like post.url and post.title. Here's a list of what's available wherever {{ post }} is available. If the variable is, e.g., post.url, use as {{ post.url }}.

These are available on individual post pages, in _template/post.html.

These are set when processing archive pages.

Serif is released under the MIT license. See LICENSE for details.

The quickest way to get changes contributed:

1. Visit the GitHub repository for Serif.
2. Fork the repository.
3. Check out a branch on the latest master for your change: git checkout -b master new-feature --- do not make changes on master! Make sure that anything added or changed has a test in the spec/ directory. Use the existing files as examples. All tests for new/changed behaviour should pass.
4. Send a pull request on GitHub, including a description of what you've changed.

• ./bin/serif {dev,generate} to run Serif commands.
• rspec to run the tests.
• rake docs will generate HTML documentation in docs/. Open docs/index.html in a browser to start.

• lib/serif/ is generally where files go.
• spec/ contains the test files.

See CHANGELOG.