Stringex

Some [hopefully] useful extensions to Ruby’s String class. It is made up of three libraries: ActsAsUrl, Unidecoder, and StringExtensions.

ActsAsUrl

This library is designed to create URI-friendly representations of an attribute, for use in generating urls from your attributes. Basic usage is just calling the method:

# Inside your model
acts_as_url :title

which will populate the url attribute on the object with the converted contents of the title attribute. This behavior can be customized by adding the following options to the arguments of the acts_as_url method:

:url_attribute: The name of the attribute to use for storing the generated url string. Default is :url
:scope: The name of model attribute to scope unique urls to. There is no default here.
:only_when_blank: If set to true, the url generation will only happen when :url_attribute is blank. Default is false (meaning url generation will happen always).
:sync_url: If set to true, the url field will be updated when changes are made to the attribute it is based on. Default is false.
:allow_slash: If set to true, the url field will not convert slashes. Default is false.
:allow_duplicates: If set to true, unique urls will not be enforced. Default is false. NOTE: This is strongly not recommended if you are routing solely on the generated slug as you will no longer be guaranteed to lookup the expected record based on a duplicate slug.
:limit: If set, will limit length of url generated. Default is nil.

In order to use the generated url attribute, you will probably want to override to_param like so, in your Model:

# Inside your model
def to_param
  url # or whatever you set :url_attribute to
end

Routing called via named routes like foo_path(@foo) will automatically use the url. In your controllers you will need to call Foo.find_by_url(params[:id]) instead of the regular find. Don’t look for params[:url] unless you set it explicitly in the routing, to_param will generate params[:id].

Note that if you add acts_as_url to an old model, the url database column will inititally be blank. To set this column for your old instances, you can use the initialize_urls method. So if your class is Post, just say Post.initialize_urls.

Unlike other permalink solutions, ActsAsUrl doesn’t rely on Iconv (which is inconsistent across platforms and doesn’t provide great transliteration as is) but instead uses a transliteration scheme (see the code for Unidecoder) which produces much better results for Unicode characters. It also mixes in some custom helpers to translate common characters into a more URI-friendly format rather than just dump them completely. Examples:

# A simple prelude
"simple English".to_url => "simple-english"
"it's nothing at all".to_url => "its-nothing-at-all"
"rock & roll".to_url => "rock-and-roll"

# Let's show off
"$12 worth of Ruby power".to_url => "12-dollars-worth-of-ruby-power"
"10% off if you act now".to_url => "10-percent-off-if-you-act-now"

# You don't even wanna trust Iconv for this next part
"kick it en Français".to_url => "kick-it-en-francais"
"rock it Español style".to_url => "rock-it-espanol-style"
"tell your readers 你好".to_url => "tell-your-readers-ni-hao"

Compare those results with the ones produced on my Intel Mac by a leading permalink plugin:

"simple English" # => "simple-english"
"it's nothing at all" # => "it-s-nothing-at-all"
"rock & roll" # => "rock-roll"

"$12 worth of Ruby power" # => "12-worth-of-ruby-power"
"10% off if you act now" # => "10-off-if-you-act-now"

"kick it en Français" # => "kick-it-en-francais"
"rock it Español style" # => "rock-it-espan-ol-style"
"tell your readers 你好" # => "tell-your-readers"

Not so great, actually.

Note: No offense is intended to the author of whatever plugins might produce such results. It’s not your faults Iconv sucks.

Unidecoder

This library converts Unicode [and accented ASCII] characters to their plain-text ASCII equivalents. This is a port of Perl’s Unidecode and provides eminently superior and more reliable results than Iconv. (Seriously, Iconv… A plague on both your houses! [sic])

You probably won’t ever need to run Unidecoder by itself. StringExtensions adds String#to_ascii which wraps all of Unidecoder’s functionality. For anyone interested, details of the implementation can be read about in the original implementation of Text::Unidecode. Extensive examples can be found in the tests.

Unidecoder module also provides localization options for Stringex. You can use this functionality by loading either a YAML file or Hash like the following code snippets:

# Using a Hash
Stringex.localize_from :en => {"é" => "ee"}

# Loading from a file. NOTE: The path to file should be absolute.
Stringex.localize_from "/path/to/yaml_file"

In both the YAML and pure Hash implementation, the end result should be a Hash with keys representing the locale and values being another Hash with those keys being the UTF character and the values the transliterated ASCII values. (I hope that made sense.) You can check the documentation for the Unidecoder module for more information about setting locales.

StringExtensions

A collection of extensions on Ruby’s String class. Please see the documentation for StringExtensions module for more information. There’s not much to explain about them really.

Note to users of CanCan

You’ll need to add a :find_by => :url to your load_and_authorize_resource. Here’s an example:

load_and_authorize_resource :class => "Whatever", :message => "Not authorized", :find_by => :url

Thanks & Acknowledgements

If it’s not obvious, some of the code for ActsAsUrl is based on Rick Olsen’s permalink_fu plugin. Unidecoder is a Ruby port of Sean Burke’s Text::Unidecode module for Perl. And, finally, the bulk of strip_html_tags in StringExtensions was stolen from Tobias Lütke’s Regex in Typo.