Semantic HTML5 Backend For Asciidoctor

This project provides alternative HTML5 converter (backend) for Asciidoctor that focuses on correct semantics, accessibility and compatibility with common typographic CSS styles.

Goals

• Clean markup with correct HTML5 semantics.

• Good accessibility for people with disabilities.

• Compatibility with common typographic CSS styles when possible and especially with GitHub and GitLab.

• Full standalone converter without fallback to the built-in Asciidoctor converters.

• Easy to use and integrate into third-party projects.

• Well readable and maintainable code – this should be never sacrificed for performance (I’m looking at you, Asciidoctor!).

Non-goals

• Compatibility with existing Asciidoctor CSS styles.

Text Substitutions

Quotes

Asciidoctor provides syntax for so-called “curved quotation marks” (which are actually just the correct quotation marks), but the built-in converters outputs only one (hard-coded) type of the single/double quotation marks — that one used in English and few other languages. This converter outputs the correct type of the quotation marks based on the specified language (using standard lang attribute).

Name	Syntax	Languages (:lang:)	HTML	Rendered
double quotes	"`word`"	af, en, eo, ga, hi, ia, id, ko, mt, th, tr, zh	“word”	“word”
		bs, fi, sv	”word”	”word”
		cs, da, de, is, lt, sl, sk, sr	„word“	„word“
		hu, pl, nl, ro	„word”	„word”
single quotes	'`word`'	af, en, eo, ga, hi, ia, id, ko, mt, th, tr, zh	‘word’	‘word’
		bs, fi, sv	’word’	’word’
		cs, da, de, is, lt, sl, sk, sr	‚word‘	‚word‘
		nl	‚word’	‚word’
		hu, pl, ro	«word»	«word»

The default (fallback) type is the first one.

Replacements

Asciidoctor replaces -- with em dash (—) and does not provide any replacement for en dash (–). That’s not very convenient, because en dash is more common than em dash; at least in (British) English and Czech (actually we don’t use em dash at all). So this extension also modifies the replacements table: changes -- to en dash and adds --- for em dash.

Name	Syntax	Unicode	Rendered	Notes
En dash	--	–	–	Only replaced if between two word characters, between a word character and a line boundary, or flanked by spaces. When flanked by space characters (e.g. a -- b or a --- b), the normal spaces are replaced by thin spaces ( ).
Em dash	---	—	—

Other Enhancements

Image Block

The link attribute recognizes few special values:

link=self

Make the image a link with URL of the image itself – to open it in full size.

link=none / link=false

Suppress the effect of :html5s-image-default-link: self, i.e. remove the default image link.

Both block image and inline image supports additional attribute loading (see Lazy loading on MDN for more information).

Additional Inline Formatting Roles

del

[del]#deleted text# is rendered as <del>deleted text</del>.

ins

[ins]#inserted text# is rendered as <ins>inserted text</ins>.

strike

[strike]#inserted text# is rendered as <s>inserted text</s>. This is an alias for line-through.

Requirements

Note: This converter consists of Slim templates, but they are precompiled into pure Ruby code using asciidoctor-templates-compiler, so you don’t need Slim to use it!

Ruby

• Ruby 2.0+ or JRuby 9.1+

• Asciidoctor 1.5.7+

• thread_safe (not required, but recommended for Ruby MRI)

Node.js

• Node.js

• @asciidoctor/core >=2.0.0 <3.0.0

Installation

Ruby

Install asciidoctor-html5s from Rubygems:

gem install asciidoctor-html5s

or to get the latest development version:

gem install --pre asciidoctor-html5s

Node.js

Install asciidoctor-html5s from npmjs.com:

npm install --save asciidoctor-html5s

Usage

CLI

asciidoctor -r asciidoctor-html5s -b html5s FILE...

Node.js

// Load asciidoctor.js and asciidoctor-html5s.
const asciidoctor = require('@asciidoctor/core')()
const asciidoctorHtml5s = require('asciidoctor-html5s')

// Register the HTML5s converter and supporting extension.
asciidoctorHtml5s.register()

// Convert the content to HTML using html5s converter.
const content = "Hello, *world!*!"
const html = asciidoctor.convert(content, { backend: 'html5s' })
console.log(html)

Attributes

Extra attributes accepted by the asciidoctor-html5s:

html5s-force-stem-type

Ignore declared (e.g. :stem: asciimath, asciimath:[], …​) and default type of the stem macro/block and always use the one specified by this attribute.
Asciidoctor hard-codes the default stem type to “asciimath”, which is not supported by KaTeX.

html5s-image-default-link: self

Make every block image a link with the image’s source URL (i.e. user can click on the image to open it in full size), unless the link attribute is defined and is not none or false.

html5s-image-self-link-label

The link title and ARIA label for the block image link that points to the image file (i.e. href equals the image’s src). Default is Open the image in full size.

License

This project is licensed under MIT License. For the full text of the license, see the LICENSE file.