Announcing the release of the External Search Sources Rails Engine
developed by Katipo Communications (http://katipo.co.nz) for Kete
(http://kete.net.nz), and funded by Digital New Zealand (http://
digitalnz.org.nz) and Katipo.
What does it do?
The engine allows you to define search sources that are based on web
services that return RSS feeds of their search results. An example
search source is the Digital New Zealand content metadata aggregation
service (http://search.digitalnz.org.nz/).
In your views, you can then call a helper method which is passed a
search term string that is defined by your code. Your defined search
sources will then be queried and their results will be presented.
This is an excellent way of displaying partner sites’ related content
along side your local search results.
There are a number of options that can be used to handle different
search sources’ boolean syntax and whether results should be cached
locally.
If the results feed for a search source returns thumbnails, they will
also be displayed.
How does it work?
The engine provides an administrative interface powered by
ActiveScffold (http://activescaffold.com/). Administrators can add
search sources which will be used to get results.
Put simply, a search source is an rss feed from another website. When
it needs to get results, it adds the search term to the end of a base
url. The base url needs to end in something similar to ‘&q=’ . That
resulting url is then fetched and parsed using a modified version of
Feedzirra (GitHub - kete/feedzirra: A feed fetching and parsing library that treats the internet like Godzilla treats Japan: it dominates and eats all.), and output to html,
which can then be styled using css to suit.
Requirements
Because we chose to go with a Rails engine and I18n support in this
case, Rails 2.3 or higher is required to run the engine.
The administrative interface requires ActiveScaffold be installed and
working as well.
Finally, the modified version of Feedzirra is needed (http://
GitHub - kete/feedzirra: A feed fetching and parsing library that treats the internet like Godzilla treats Japan: it dominates and eats all.). We have found cases where certain feed
elements weren’t being detected, so we have added these to our own
version of the gem and sent pull requests to the maintainers of the
original gem.
Assumptions
The engine has some assumptions for authorization that can be easily
reconfigured to suit your application.
It assumes you are using the authorization plugin (http://github.com/
DocSavage/rails-authorization-plugin) and acts as authenticated
(http://svn.techno-weenie.net/projects/plugins/acts_as_authenticated).
See the README.rdoc for more information.
So how do I install this?
Head on over to it’s Github repository and checkout the README.rdoc
file.
It has all the steps you need to get running with it as quickly as
possible. It also has a heap of documentation about configurable
settings, translating, styling id/classes, and testing.
Credits
This engine was co-designed by Kieran Pilkington (
) and Walter McGinnis ( walter (Walter McGinnis) · GitHub ).
Kieran did the implementation and testing. Walter provided guidance
and implementation feedback.