Akismet (http://akismet.com/) is a collaborative spam filtering service. Rakismet is easy Akismet integration with Rails and rack apps. TypePad's AntiSpam service and generic Akismet endpoints are supported.
Rakismet >= 1.0.0 work with Rails 3 and other Rack-based frameworks.
Rakismet <= 0.4.2 is compatible with Rails 2.
Add the Rakismet gem to your Gemfile
gem 'rakismet'
Once you've added the Rakismet gem to your Gemfile and installed it with bundle install
,
you'll need an API key. Head on over to http://akismet.com/signup/ and sign up
for a new username.
Configure the Rakismet key and the URL of your application by setting the following in application.rb:
config.rakismet.key = 'your wordpress key'
config.rakismet.url = 'http://yourdomain.com/'
or an initializer, for example config/initializers/rakismet.rb
:
YourApp::Application.config.rakismet.key = 'your wordpress key'
YourApp::Application.config.rakismet.url = 'http://yourdomain.com/'
If you wish to use another Akismet-compatible API provider such as TypePad's
antispam service, you'll also need to set config.rakismet.host
to your service
provider's endpoint.
If you want to use a proxy to access akismet (i.e. your application is behind a firewall), set the proxy_host and proxy_port option.
config.rakismet.proxy_host = 'http://yourdomain.com/'
config.rakismet.proxy_port = '8080'
If your Rails app is a multitenant application, you can specify your Rakismet url as a proc:
config.rakismet.url = Proc.new { ApplicationController.current_tenant.url }
First, introduce Rakismet to your model:
class Comment
include Rakismet::Model
end
With Rakismet mixed in to your model, you'll get three instance methods for interacting with Akismet:
spam?
submits the comment to Akismet and returns true if Akismet thinks the comment is spam, false if not.ham!
resubmits a valid comment that Akismet erroneously marked as spam (marks it as a false positive.)spam!
resubmits a spammy comment that Akismet missed (marks it as a false negative.)The ham!
and spam!
methods will change the value of spam?
but their
primary purpose is to send feedback to Akismet. The service works best when you
help correct the rare mistake; please consider using these methods if you're
moderating comments or otherwise reviewing the Akismet responses.
Rakismet sends the following information to the spam-hungry robots at Akismet:
author : name submitted with the comment
author_url : URL submitted with the comment
author_email : email submitted with the comment
comment_type : Defaults to comment but you can set it to trackback, pingback, or something more appropriate
content : the content submitted
permalink : the permanent URL for the entry the comment belongs to
user_ip : IP address used to submit this comment
user_agent : user agent string
referrer : referring URL (note the spelling)
By default, Rakismet just looks for attributes or methods on your class that match these names. You don't have to have accessors that match these exactly, however. If yours differ, just tell Rakismet what to call them:
class Comment
include Rakismet::Model
attr_accessor :commenter_name, :commenter_email
rakismet_attrs :author => :commenter_name, :author_email => :commenter_email
end
Or you can pass in a proc, to access associations:
class Comment < ActiveRecord::Base
include Rakismet::Model
belongs_to :author
rakismet_attrs :author => proc { author.name },
:author_email => proc { author.email }
end
You can even hard-code specific fields:
class Trackback
include Rakismet::Model
rakismet_attrs :comment_type => "trackback"
end
Akismet wants certain information about the request environment: remote IP, the
user agent string, and the HTTP referer when available. Normally, Rakismet
asks your model for these. Storing this information on your model allows you to
call the spam?
method at a later time. For instance, maybe you're storing your
comments in an administrative queue or processing them with a background job.
You don't need to have these three attributes on your model, however. If you choose to omit them, Rakismet will instead look at the current request (if one exists) and take the values from the request object instead.
This means that if you are not storing the request variables, you must call
spam?
from within the controller action that handles comment submissions. That
way the IP, user agent, and referer will belong to the person submitting the
comment. If you're not storing the request variables and you call spam?
at a
later time, the request information will be missing or invalid and Akismet won't
be able to do its job properly.
If you've decided to handle the request variables yourself, you can disable the middleware responsible for tracking the request information by adding this to your app initialization:
config.rakismet.use_middleware = false
Additionally, the middleware will send along additional env variables starting with HTTP_ to Akismet. If you wish to block any sensitive user information, use:
config.rakismet.excluded_headers = ['HTTP_COOKIE','HTTP_SENSITIVE']
excluded_headers will default to ['HTTP_COOKIE']
Rakismet can be configued to tell Akismet that it should operate in test mode - so Akismet will not change its behavior based on any test API calls, meaning they will have no training effect. That means your tests can be somewhat repeatable in the sense that one test won't influence subsequent calls.
You can configure Rakismet for test mode via application.rb:
config.rakismet.test = false # <- default
config.rakismet.test = true
Or via an initializer:
YourApp::Application.config.rakismet.test = false # <- default
YourApp::Application.config.rakismet.test = true
NOTE: When running in Rails, Rakismet will run in test mode when your Rails
environment is test
or development
, unless explictly configured otherwise.
Outside of Rails Rakismet defaults to test mode turned off.
If you want to see what's happening behind the scenes, after you call one of
@comment.spam?
, @comment.spam!
or @comment.ham!
you can check
@comment.akismet_response
.
This will contain the last response from the Akismet server. In the case of
spam?
it should be true
or false.
For spam!
and ham!
it should be
Feedback received.
If Akismet returned an error instead (e.g. if you left out
some required information) this will contain the error message.
Akismet needs enough information to decide if your test data is spam or not. Try to supply as much as possible, especially the author name and request variables.
Most people have the opposite problem, where Akismet doesn't think anything is spam. The only guaranteed way to trigger a positive spam response is to set the comment author to "viagra-test-123". To simulate a negative (not spam) result, set user_role to administrator, and all other required fields populated with typical values. The Akismet API will always return a false response.
If you've done this and spam?
is still returning false, you're probably
missing the user IP or one of the key/url config variables. One way to check is
to call @comment.akismet_response
. If you are missing a required field or
there was another error, this will hold the Akismet error message. If your comment
was processed normally, this value will simply be true
or false
.
Sure. Rakismet doesn't care what your persistence layer is. It will work with Datamapper, a NoSQL store, or whatever next month's DB flavor is.
Rakismet also has no dependencies on Rails or any of its components, and only uses a small Rack middleware object to do some of its magic. Depending on your framework, you may have to modify this slightly and/or manually place it in your stack.
You'll also need to set a few config variables by hand. Instead of
config.rakismet.key
, config.rakismet.url
, and config.rakismet.host
, set
these values directly with Rakismet.key
, Rakismet.url
, and Rakismet.host
.
If you have any implementation or usage questions, don't hesitate to get in touch: josh@vitamin-j.com.
Copyright (c) 2008 Josh French, released under the MIT license