Rouge is a pure-ruby syntax highlighter. It can highlight 100 different languages, and output HTML or ANSI 256-color text. Its HTML output is compatible with stylesheets designed for pygments.
If you'd like to help out with this project, assign yourself something from the issues page, and send me a pull request (even if it's not done yet!). Bonus points for feature branches.
First, take a look at the pretty colors.
# make some nice lexed html
source = File.read('/etc/bashrc')
formatter = Rouge::Formatters::HTML.new
lexer = Rouge::Lexers::Shell.new
formatter.format(lexer.lex(source))
# Get some CSS
Rouge::Themes::Base16.mode(:light).render(scope: '.highlight')
# Or use Theme#find with string input
Rouge::Theme.find('base16.light').render(scope: '.highlight')
As of Rouge 2.0, you are encouraged to write your own formatter for custom formatting needs. Builtin formatters include:
Rouge::Formatters::HTML.new
- will render your code with standard class names for tokens, with no div-wrapping or other bells or whistles.Rouge::Formatters::HTMLInline.new(theme)
- will render your code with no class names, but instead inline the styling options into thestyle=
attribute. This is good for emails and other systems where CSS support is minimal.Rouge::Formatters::HTMLLinewise.new(formatter, class: 'line-%i')
This formatter will split your code into lines, each contained in its own div. Theclass
option will be used to add a class name to the div, given the line number.Rouge::Formatters::HTMLPygments.new(formatter, css_class='codehilite')
wraps the given formatter with div wrappers generally expected by stylesheets designed for Pygments.Rouge::Formatters::HTMLTable.new(formatter, opts={})
will output an HTML table containing numbered lines. Options are:start_line: 1
- the number of the first lineline_format: '%i'
- asprintf
template for the line number itselftable_class: 'rouge-table'
- a CSS class for the tablegutter_class: 'rouge-gutter'
- a CSS class for the guttercode_class: 'rouge-code'
- a CSS class for the code column
Rouge::Formatters::HTMLLegacy.new(opts={})
is a backwards-compatibility class intended for users of rouge 1.x, with options that were supported then. Options are:inline_theme: nil
- use an HTMLInline formatter with the given themeline_numbers: false
- use an HTMLTable formatterwrap: true
- use an HTMLPygments wrappercss_class: 'codehilite'
- a CSS class to use for the pygments wrapper
Rouge::Formatters::Terminal256.new(theme)
theme
must be an instnce ofRouge::Theme
, or aHash
structure with:theme
entry
Print a trace of the lex on stdout
Allows you to specify which language the template is inside
CSS selector that styles are applied to, e.g. Rouge::Themes::MonokaiSublime.render(scope: 'code')
Rouge aims to be simple to extend, and to be a drop-in replacement for pygments, with the same quality of output. Also, Rouge ships with a rougify
command which allows you to easily highlight files in your terminal:
$ rougify foo.rb
$ rougify style monokai.sublime > syntax.css
- No need to spawn Python processes.
- We're faster in almost every measure
- The HTML output from Rouge is fully compatible with stylesheets designed for pygments.
- The lexers are implemented with a dedicated DSL, rather than being hand-coded.
- Rouge supports every language CodeRay does and more.
require 'redcarpet'
require 'rouge'
require 'rouge/plugins/redcarpet'
class HTML < Redcarpet::Render::HTML
include Rouge::Plugins::Redcarpet # yep, that's it.
end
If you have :fenced_code_blocks
enabled, you can specify languages, and even options with CGI syntax, like php?start_inline=1
, or erb?parent=javascript
.
Rouge is only for UTF-8 strings. If you'd like to highlight a string with a different encoding, please convert it to UTF-8 first.
- Middleman: middleman-syntax (@bhollis)
- Middleman: middleman-rouge (@Linuus)
- RDoc: rdoc-rouge (@zzak)
- Rouge::Rails: render code samples in your rails views (@jacobsimeon)
If you're here to implement a lexer for your awesome language, there's a good chance you don't already have a ruby development environment set up. Follow the instructions on the wiki to get up and running. If you have trouble getting set up, let me know - I'm always happy to help.
You can test the core of Rouge simply by running rake
(no bundle exec
required), or rake spec TEST=spec/xxx_spec.rb
to run a single test file.
It's also set up with guard
, if you like.
To test a lexer visually, run rackup
from the root and go to localhost:9292/#{some_lexer}
where some_lexer
is the tag or an alias of a lexer you'd like to test. If you add ?debug=1
, helpful debugging info will be printed on stdout.
is at http://rubydoc.info/gems/rouge/frames.
You can probably learn a lot just by reading through the existing lexers. Basically, a lexer consists of a collection of states, each of which has several rules. A rule consists of a regular expression and an action, which yields tokens and manipulates the state stack. Each rule in the state on top of the stack is tried in order until a match is found, at which point the action is run, the match consumed from the stream, and the process repeated with the new lexer on the top of the stack. Each lexer has a special state called :root
, and the initial state stack consists of just this state.
Here's how you might use it:
class MyLexer < Rouge::RegexLexer
state :root do
# the "easy way"
# simple rules
rule /0x[0-9a-f]+/, Num::Hex
# simple state stack manipulation
rule /{-/, Comment, :next_state
rule /-}/, Comment, :pop!
# the "flexible way"
rule /abc/ do |m|
# m is the match, for accessing match groups manually
# you can do the following things:
pop!
push :another_state
push # assumed to be the current state
state? :some_state # check if the current state is :some_state
in_state? :some_state # check if :some_state is in the state stack
# yield a token. if no second argument is supplied, the value is
# taken to be the whole match.
# The sum of all the tokens yielded must be equivalent to the whole
# match - otherwise characters will go missing from the user's input.
token Generic::Output, m[0]
# calls SomeOtherLexer.lex(str) and yields its output. See the
# HTML lexer for a nice example of this.
# if no second argument is supplied, it is assumed to be the whole
# match string.
delegate SomeOtherLexer, str
# the context object is the lexer itself, so you can stash state here
@count ||= 0
@count += 1
# advanced: push a dynamically created anonymous state
push do
rule /.../, Generic::Output
end
end
rule /(\w+)(:)/ do
# "groups" yields the matched groups in order
groups Name::Label, Punctuation
end
end
start do
# this is run whenever a fresh lex is started
end
end
If you're creating a lexer that's very similar to a different lexer, you can use subclassing (see C/C++/ObjC and also QML/Javascript for examples):
class MyLexer < OtherLexer
# independent states
state :my_state do ... end
# override states
state :your_state do ... end
# prepend rules to states
prepend :parent_state do ... end
# append rules to states
append :parent_state do ... end
end
Please don't submit lexers that are largely copy-pasted from other files.
I don't get paid to maintain rouge. If you've found this software useful, consider dropping a tip in the bucket.
Rouge is released under the MIT license. Please see the LICENSE
file for more information.