handroll - A website generator for software artisans¶

handroll is a static website generator that uses markup languages like Markdown, reStructuredText, and Textile.

handroll development is done on GitHub. Announcements and discussions happen on Google Groups.

Installation¶

handroll is available for download from PyPI. You can install it with pip. handroll is currently supported on Python 3.5, 3.4, 3.3, 2.7, and PyPy.

$ pip install handroll

Usage¶

When inside a website’s source directory, the following command will generate results and store them in output. Use handroll -h to see all the options.

$ handroll build
Complete.

Features¶

handroll follows a “batteries included” philosophy for generating static websites. One goal is to support a wide range of tools to cater to many diverse interests. The list below isn’t exhaustive, but it provides a good idea of what handroll is capable of doing.

Start a new site with a single command using Scaffolds for immediate results.
Convert Markdown to HTML.
Convert reStructuredText to HTML.
Convert Textile to HTML.
Convert Sass to CSS.
Copy static assets like CSS or JavaScript.
Track blog entries and automatically generate a feed (see Blog extension).
Generate a proper Atom XML feed from metadata stored in JSON.
Run a development server with the watch flag to monitor your site and update the output immediately as you make changes (see Development server).
Find the site source root so you don’t have to. If you’re anywhere in your site’s source, calling handroll without the site input parameter will trigger handroll to look for your site’s root directory.
Store global configuration in a configuration file (see Configuration). You’ll never need to specify the output directory again.
Keep extra data for templates in a separate front matter section in YAML format (see Front matter).
Templates can use the Jinja2 template engine.
Content is only updated when either a template or the source file is newer than the existing output file. This eliminates wasted regeneration on unchanged content.
Be extensible for users who want to write their own plugins (see Composers and Extensions).
Provide timing information to see file processing time.
Translated to many different languages.

The remaining documentation provides additional details about all listed features.

Documentation¶

Configuration¶

handroll supports an optional handroll.conf file that can be stored at the root of the site’s directory. This ini style file provides configuration information that handroll will use while generating the output. For example:

[site]
outdir = ~/mblayman.github.io

Arguments provided on the command line will override the equivalent configuration file option.

`site` section¶

The outdir option will determine the output directory.

The outdir permits relative paths. One useful pattern with relative paths is to set outdir = .. as the value. Source and output can exist in a single repository or directory. Putting the output at the root of a repository makes it easy to deploy the entire project as a website. When generating output or watching the source directory, handroll is aware of the source and allows the two directories to coexist without interference.

If a tilde character (~) is supplied, it will be expanded to the user’s home directory.

The with_blog option set to true, on, yes, or 1 will enable the blog extension. See Blog extension for setup information.

Front matter¶

Source documents like Markdown files can have additional data added to them. This data is stored in a front matter section at the top of a source document. handroll will read the extra data and pass it along to the template. In the template, the data will be accessible by whatever name was provided. An example Markdown source document would look like:

---
title: A Sample Site Title
bonus: It's a secret to everybody.
---
## Another heading

This is text in the body.

You may also include the YAML directive (e.g., %YAML 1.1). The following example is equally valid.

%YAML 1.1
---
title: A Sample Site Title
bonus: It's a secret to everybody.
---
## Another heading

This is text in the body.

Note: When using front matter, handroll does not infer the title from the first line of the document. If a title is desired, the attribute must be explicitly added to the front matter.

Scaffolds¶

handroll can generate a new site using a single command. This will help you get started in a snap. After making a site with the scaffold command, you can immediately run handroll to get a functioning website.

$ handroll scaffold default mysite

Scaffolds can include any content. The default scaffold includes a base template, an index file in Markdown, a configuration file, and a sample CSS file.

To list the available scaffolds,

$ handroll scaffold

Development server¶

handroll comes with a built-in development server. The server helps develop websites even faster by watching the changes you make. As files in your site are created, modified, or moved, the server will update your output with each change.

The development server is available with the watch command. The server will make your site accessible on http://localhost:8000.

Here is an example:

matt@eden:~/handroll/sample$ handroll watch
Serving /home/matt/handroll/sample/output at http://localhost:8000/.
Press Ctrl-C to quit.
Generating HTML for /home/matt/handroll/sample/index.md ...

Composers¶

handroll uses a plugin system to decide how to process each file type. The plugins are called composers. A composer is provided a source file and can produce whatever output it desires. handroll will load each available composer using setuptools entry points. handroll loads the class and constructs a Composer instance by invoking a no parameter constructor.

class handroll.composers.Composer(config)¶

Interface for all composers

__init__(config)¶: Each composer is given the configuration when instantiated.

compose(catalog, source_file, out_dir)¶

Compose whatever appropriate output is generated by the composer.

Parameters:	catalog – the `TemplateCatalog` source_file – the filename of the source out_dir – the directory to store output

get_output_extension(filename)¶

Get the extension of the output file generated by this composer.

The filename is required because some composers may vary their extension based on the filename.

A plugin should be added to the handroll.composers entry point group. For example, the MarkdownComposer plugin included by default defines its entry point in setup.py as:

entry_points={
    'handroll.composers': [
        '.md = handroll.composers.md:MarkdownComposer',
    ]
}

This entry point registers the MarkdownComposer class in the handroll.composers.md module for the .md file extension. The example is slightly confusing because the entry point name and the package are the same so here is a fictious example.

A composer class called FoobarComposer defined in another.package for the .foobar file extension would need the following entry point.

entry_points={
    'handroll.composers': [
        '.foobar = another.package:FoobarComposer',
    ]
}

Built-in composers¶

class handroll.composers.atom.AtomComposer(config)¶

Compose an Atom feed from an Atom metadata file (.atom).

The AtomComposer parses the metadata specified in the source file and produces an XML Atom feed. AtomComposer uses parameters that are needed by Werkzeug’s AtomFeed API. Refer to the Werkzeug documentation for all the available options.

The dates in the feed should be in RfC 3339 format (e.g., 2014-06-13T11:39:30).

Here is a sample feed:

{
  "title": "Sample Feed",
  "url": "http://some.website.com/archive.html",
  "id": "http://some.website.com/feed.xml",
  "author": "Matt Layman",
  "entries": [
    {
      "title": "Sample C",
      "updated": "2014-05-04T12:00:00",
      "url": "http://some.website.com/c.html",
      "summary": "A summary of the sample post"
    },
    {
      "title": "Sample B",
      "updated": "2014-03-17T12:00:00",
      "url": "http://some.website.com/b.html",
      "summary": "A summary of the sample post"
    },
    {
      "title": "Sample A",
      "updated": "2014-02-23T00:00:00",
      "url": "http://some.website.com/a.html",
      "summary": "A summary of the sample post"
    }
  ]
}

class handroll.composers.j2.Jinja2Composer(config)¶

Compose any content from a Jinja 2 template files (.j2).

The Jinja2Composer takes a template file and processes it through the Jinja 2 renderer. The site configuration is provided to the context for access to global data.

The output file uses the same name as the source file with the .j2 extension removed.

class handroll.composers.CopyComposer(config)¶

Copy a source file to the destination.

CopyComposer is the default composer for any unrecognized file type. The source file will be copied to the output directory unless there is a file with an identical name and content already at the destination.

class handroll.composers.md.MarkdownComposer(config)¶

Compose HTML from Markdown files (.md).

The first line of the file will be used as the title data for the template. All following lines will be converted to HTML and sent to the template as the content data.

The MarkdownComposer supports syntax highlighting using Pygments. Code can be specified using “fenced code” triple backticks.

```python
class Foo(object):
    '''This sample code would be highlighted in a Python style.'''
```

Use pygmentize to create your desired CSS file. Refer to the Pygments documentation for more information.

$ pygmentize -S default -f html > pygments.css

The MarkdownComposer generates better typographical quotes by using the SmartyPants library.

class handroll.composers.rst.ReStructuredTextComposer(config)¶

Compose HTML from reStructuredText files (.rst).

The first line of the file will be used as the title data for the template. All following lines will be converted to HTML and sent to the template as the content data.

class handroll.composers.sass.SassComposer(path=None)¶

Compose CSS files from Sass files (.scss or .sass).

Sass is a CSS preprocessor to help manage CSS files. The Sass website has great documentation to explain how to use it.

Because Sass is not written in the same language as handroll, it must be installed separately before it can be used. Check out the installation options.

class handroll.composers.txt.TextileComposer(config)¶

Compose HTML from Textile files (.textile).

The first line of the file will be used as the title data for the template. All following lines will be converted to HTML and sent to the template as the content data.

Extensions¶

In addition to Composers, handroll has an extension system to plug in other functionality. Users enable extensions by adding with_* = true to their site section in the configuration file, where * is the name of the extension. For example, the blog extension is named blog, and with_blog = true will enable it.

Extension authors can use the base Extension to create new extensions. Extensions are never directly called, but an extension can connect to one of handroll’s Signals.

class handroll.extensions.base.Extension(config)¶

A base extension which hooks handler methods to handroll’s signals.

on_frontmatter_loaded(source_file, frontmatter)¶

Handle the frontmatter_loaded signal.

Activate this handler by setting handle_frontmatter_loaded to True in the extension subclass.

Parameters:	source_file – Absolute path of the source file frontmatter – Dictionary of parsed frontmatter

on_post_composition(director)¶

Handle the post_composition signal.

Activate this handler by setting handle_post_composition to True in the extension subclass.

Parameters:	director – The director instance

on_pre_composition(director)¶

Handle the pre_composition signal.

Activate this handler by setting handle_pre_composition to True in the extension subclass.

Parameters:	director – The director instance

Extension authors can include new extensions by adding to the handroll.extensions entry point. For example, handroll includes the following entry point in setup.py:

entry_points={
    'handroll.extensions': [
        'blog = handroll.extensions.blog:BlogExtension',
    ]
}

Built-in extensions¶

Blog extension¶

The blog extension allows you to automatically generate an atom feed of blog entries. It can also create an entry list for one of your pages.

Enable the blog extension by adding with_blog = True to the site section of your configuration file.

Atom feed¶

The extension requires some additional information to create a valid atom feed. Add a blog section to your configruation file with the following fields:

atom_author - The author of the blog
atom_id - A unique identifier for the atom feed. One suggestion is to use the URL of the feed itself. For example, http://www.mattlayman.com/feed.xml.
atom_title - The title of the blog
atom_url - The URL for the feed. For example, http://www.mattlayman.com/archive.html.

To create the atom feed, you need to specify an output path using the atom_output option. The path provided is relative to the output directory.

[blog]
atom_output = feed.xml

In this example, the atom feed would be stored in the root of the output directory with a filename of feed.xml.

List page¶

To create a blog list page, add a list_template option to your blog section. If you include list_template, then you must also include list_output. list_output is a path relative to the output directory.

When the blog extension generates the list page, the context will receive a blog_list. The blog_list is an HTML fragment of list item tags. There is one list item tag for every post.

Here is a possible sample template.

<html>
<body>
  <ul>
    {{ blog_list }}
  </ul>
</body>
</html>

And here is some possible output.

<html>
<body>
  <ul>
    <li><a href="/another_post.html">Another post</a></li>
    <li><a href="/a_post.html">First post!</a></li>
  </ul>
</body>
</html>

For more complex formatting, the actual blog posts are provided in the context as posts.

Blog post frontmatter¶

A source file is marked as a blog post by setting blog: True in the front matter. The blog front matter has required and optional fields.

Required fields:

title - The title of the post
date - The published date of the post. The date should be in RfC 3339 format (e.g., 2015-07-15T12:00:00Z).

Optional fields:

summary - A summary of the post.

Sitemap extension¶

The sitemap extension generates a sitemap of your site’s HTML content. The generated file will be stored in the root of the output directory as sitemap.txt.

Enable the sitemap extension by adding with_sitemap = True to the site section of your configuration file.

Signals¶

handroll fires various signals while running. These signals provide hooks for extensions to execute additional code. The list of signals is provided below.

frontmatter_loaded¶

frontmatter_loaded fires whenever a file contains a front matter section (see Front matter). Any handler function that connects to the signal will be called with:

source_file - The absolute path to the file containing front matter.
frontmatter - The front matter dictionary that was loaded.

pre_composition¶

pre_composition fires before processing the entire site. When the watcher is running (see Development server), the signal will fire before handling any file or directory change. Any handler function that connects to the signal will be called with:

director - The director instance that processed the site.

post_composition¶

post_composition fires after processing the entire site. When the watcher is running (see Development server), the signal will fire after handling any file or directory change. Any handler function that connects to the signal will be called with:

director - The director instance that processed the site.

Templates¶

Your source content (e.g., Markdown or reStructuredText) is read and converted into HTML. After content is converted to HTML, it is passed to a template system as a variable called content. Each template system can then insert the HTML content into a template.

handroll supports multiple template systems. Templates are stored in a templates directory at the root of your site. Alternatively, if you have very simple needs, you can use a template.html file at your site’s root.

Any template used from the templates directory must be specified using front matter (see Front matter) or the default template.html will be used. This sample Markdown file uses a string template.

%YAML 1.1
---
title: With a different template
template: different.html
---
## Another heading

This is using a different string template.

String templates¶

Any template using the .html extension (including the default template.html) will uses Python’s built-in string templates. String templates are limited to the capabilities of the standard library, but they can support basic needs.

Jinja2 templates¶

Any template using the .j2 extension will use the Jinja2 template language. handroll works with Jinja’s template inheritance system and the majority of Jinja’s other features.

Translation¶

For Translators¶

Translation for handroll is done on handroll’s Transifex project. All contributions from any language are welcome. The project should be configured to automatically accept requests to join the translation team.

Translators will be added to the AUTHORS list of contributors for contributions of any size (unless you specifically do not want to be listed).

If you have any problems, please feel free to report an issue on Github. If there is a new language that you want to translate for, please file an issue so it can be added (there is a small amount of software setup to make a new language work with handroll).

For Developers¶

All user interface strings (aside from flag names) should be translated. Strings are wrapped in a conventional way for gettext using the alias _. The gettext method is configured in the handroll.i18n module. Any new code would use that method for strings. For example:

from handroll.i18n import _

translated = _('This string is translated.')

The handroll.pot file is the template that the translators use as the basis for translation. handroll uses Babel to generate the pot file. A new pot file is generated by executing python setup.py extract_messages. The updated pot file is automatically synced to Transifex after it is pushed to GitHub.

Deployment¶

Download the po files. The mo files are automatically generated with the sdist command.

$ python transifex.py
$ python setup.py sdist

Releases¶

Version 3.1, Released December 26, 2016¶

Processs Jinja 2 templates for any file with a .j2 extension with the built-in Jinja2Composer.
Add SitemapExtension to generate sitemaps.
Move version information into the handroll package so it is available at runtime.
Perform continuous integration testing on OS X.
Include posts in the blog feed list to permit more complex list rendering.
Remove support for Python 2.6

Version 3.0, Released March 7, 2016¶

Replaced all flag based commands with sub-commands. This change means all interaction now happens through handroll build, handroll watch, and handroll scaffold.

Version 2.1, Released October 18, 2015¶

Create a site quickly with the new scaffold command (e.g., handroll -s default new_site)
Use the SmartyPants library to generate better quotation marks for Markdown.
Composers can be forced to compose with the --force flag.
Translated to Arabic.
Relax the frontmatter requirement and don’t force the inclusion of the YAML directive (e.g., %YAML 1.1).
Support Python 3.5.
An output directory can be a relative path.

Version 2.0, Released July 25, 2015¶

Added an extension interface for plugin authors to integrate with various events.
Added a blog extension to automatically generate an Atom XML feed and blog listing page.
Translated to Greek.

Version 1.5, Released February 24, 2015¶

Translated to Dutch.

Version 1.4, Released December 1, 2014¶

A development server (accessible from the watch flag) will monitor a site and generate new output files as the source is modified.
Sass support for .scss and .sass files.
Add internationalization (i18n).
Translated to French, German, Italian, Portuguese, and Spanish.
Skip certain directories that should not be in output (like a Sass cache).
Moved project to a GitHub organization to separate from a personal account.
Include documentation in the release.
Massive unit test improvements (100% coverage).

Version 1.3, Released September 3, 2014¶

Update the appropriate output only when a template or content was modified.
Use Jinja templates or standard Python string templates.
Provide YAML formatted front matter to add any data to a template.

Version 1.2, Released July 2, 2014¶

Add a basic configuration file to specify the output directory.
A search for the site root is done when no site path is provided.
Add timing reporting to find slow composers.
Update Textile version to enable Python 3 support.
Generate Atom feeds.
Drop 3.2 support. Too many dependencies do not support it.

Version 1.1, Released June 1, 2014¶

Skip undesirable file types (e.g., Vim .swp files).
Use Markdown code highlighting (via Pygments) and fenced code extensions.
All input and output is handled as UTF-8 for better character encoding.
Run against Python versions 2.6 through 3.4 using Travis CI.
Add a plugin architecture to support composers for any file type.
Provide HTML docs at Read the Docs.
Textile support for .textile files.
ReStructuredText support for .rst files.
Support PyPy.

Version 1.0, Released May 4, 2014¶

Initial release of handroll
Copy all file types.
Convert Markdown to HTML.

handroll - A website generator for software artisans¶

Installation¶

Usage¶

Features¶

Documentation¶

Configuration¶

site section¶

Front matter¶

Scaffolds¶

Development server¶

Composers¶

Built-in composers¶

Extensions¶

Built-in extensions¶

Blog extension¶

Atom feed¶

List page¶

Blog post frontmatter¶

Sitemap extension¶

Signals¶

frontmatter_loaded¶

pre_composition¶

post_composition¶

Templates¶

String templates¶

Jinja2 templates¶

Translation¶

For Translators¶

For Developers¶

Deployment¶

Releases¶

Version 3.1, Released December 26, 2016¶

Version 3.0, Released March 7, 2016¶

Version 2.1, Released October 18, 2015¶

Version 2.0, Released July 25, 2015¶

Version 1.5, Released February 24, 2015¶

Version 1.4, Released December 1, 2014¶

Version 1.3, Released September 3, 2014¶

Version 1.2, Released July 2, 2014¶

Version 1.1, Released June 1, 2014¶

Version 1.0, Released May 4, 2014¶

`site` section¶