Entity-style text parsing
Switch branches/tags
Latest commit 5973ed7 Jul 24, 2013 @soffes Update readme
Failed to load latest commit information.



Build Status Coverage Status Code Climate Dependency Status Gem Version

Entity-style text parsing. Quesadilla was extracted from Cheddar.

See the Cheddar text guide for more information about how to type entities.

Quesadilla's API is fully documented. Read the online documentation.


Add this line to your application's Gemfile:

gem 'quesadilla'

And then execute:

$ bundle

Or install it yourself as:

$ gem install quesadilla


To extract entites from text, simply call extract:

Quesadilla.extract('Some #awesome text')
# => {
#   display_text: "Some #awesome text",
#   display_html: "Some <a href=\"#hashtag-awesome\" class=\"tag\">#awesome</a> text",
#   entities: [
#     {
#       type: "hashtag",
#       text: "#awesome",
#       display_text: "#awesome",
#       indices: [5, 13],
#       hashtag: "awesome",
#       display_indices: [5, 13]
#     }
#   ]
# }


Quesadilla supports extracting various span-level Markdown features as well as automatically detecting links and GitHub-style named emoji. Here are the list of options you can pass when extracting:

Option Description Default
:markdown All Markdown parsing true
:markdown_code Markdown code tags true
:markdown_links Markdown links (including <http://soff.es> style links) true
:markdown_triple_emphasis Markdown bold italic true
:markdown_double_emphasis Markdown bold true
:markdown_emphasis Markdown italic true
:markdown_strikethrough Markdown Extra strikethrough true
:hashtags Hashtags true
:hashtags_validator Callable object to validate hashtags nil
:autolinks Automatically detect links true
:emoji GitHub-style named emoji true
:users User mentions false
:user_validator Callable object to validate usernames nil
:html Generate HTML representations for entities and the entire string true

Everything is enabled by deafult except user mentions. If you don't want to extract Markdown, you should call the extractor this like:

Quesadilla.extract('Some text', markdown: false)

You can also just disable strikethrough and still extract the rest of the Markdown entities if you want:

Quesadilla.extract('Some text', markdown_strikethrough: false)

Customizing HTML

If you want to change the generated HTML, you can create a custom renderer:

class CustomRenderer < Quesadilla::HTMLRenderer
  def hashtag(display_text, hashtag)
    %Q{<a href="http://example.com/tags/#{hashtag}" class="tag">#{display_text}</a>}

extraction = Quesadilla.extract('Some #awesome text', html_renderer: CustomRenderer)
extraction[:display_html] #=> 'Some <a href="http://example.com/tags/awesome" class="tag">#awesome</a> text'

Take a look at Quesadilla::HTMLRenderer for more details on creating a custom renderer.


To enable user mention extraction, pass users: true as an option. You can optionally pass a callable object to validate a username. Here's a quick example:

validator = lambda do |username|
  User.where('LOWER(username) = ?', username.downcase).first.try(:id)

extraction = extract('Real @soffes and fake @nobody', users: true, user_validator: validator)

Assuming there is a user named soffes in your database, it would extract @soffes. Assuming there isn't a user named nobody, that would remain plain text. Obviously feel free to do whatever you want here. Quesadilla makes no assumptions about your user system.


User and hashtag detection use the twitter-text gem. This has some limits that you may not expect such as usernames can't be more than 20 characters and hashtags can't contain certain characters.

Supported Ruby Versions

Quesadilla is tested under 1.9.2, 1.9.3, 2.0.0, JRuby 1.7.2 (1.9 mode), and Rubinius 2.0.0 (1.9 mode).


See the contributing guide.