using thinking sphinx search in rails app

read

I am using acts_as_ferret and ferret server as of now with my rails application and it just works fine for me. The ONLY problem is performance as it takes a lot time to build index and to rebuild index when it gets screwed up and that’s where sphinx rocks!

To get started with thinking sphinx you need to install sphinx server first. for installation help click here.

To use sphinx search in rails we need to use either thinking sphinx gem or plugin that can be easily find on github.

Plugin installation

ruby script/plugin install  git://github.com/freelancing-god/thinking-sphinx.git

OR

## Gem installation

To install gem run the following command.

<div class='code-highlight'><pre class='code-highlight-pre'><div data-line='1' class='code-highlight-row numbered'><div class='code-highlight-line'>gem install thinking-sphinx –source http://rubygems.org/ </div></div><div data-line='2' class='code-highlight-row numbered'><div class='code-highlight-line'> </div></div><div data-line='3' class='code-highlight-row numbered'><div class='code-highlight-line'>If you’re upgrading, you should read this: </div></div><div data-line='4' class='code-highlight-row numbered'><div class='code-highlight-line'>http://freelancing-god.github.com/ts/en/upgrading.html </div></div><div data-line='5' class='code-highlight-row numbered'><div class='code-highlight-line'> </div></div><div data-line='6' class='code-highlight-row numbered'><div class='code-highlight-line'>Successfully installed thinking-sphinx-1.3.16 </div></div><div data-line='7' class='code-highlight-row numbered'><div class='code-highlight-line'>1 gem installed</div></div></pre></div>

Edit environment file and add following lines to it inside initializer block.

<div class='code-highlight'><pre class='code-highlight-pre'><div data-line='1' class='code-highlight-row numbered'><div class='code-highlight-line'>config.gem( </div></div><div data-line='2' class='code-highlight-row numbered'><div class='code-highlight-line'> ‘thinking-sphinx’, </div></div><div data-line='3' class='code-highlight-row numbered'><div class='code-highlight-line'> :lib => ‘thinking_sphinx’, </div></div><div data-line='4' class='code-highlight-row numbered'><div class='code-highlight-line'> :version => ‘1.3.16’ </div></div><div data-line='5' class='code-highlight-row numbered'><div class='code-highlight-line'> )</div></div></pre></div>

Edit Rakefile and add following lines to it.

require 'thinking_sphinx/tasks'

List rake tasks that should show up sphinx related tasks.

<div class='code-highlight'><pre class='code-highlight-pre'><div data-line='1' </div></div><div data-line='2' class='code-highlight-row numbered'><div class='code-highlight-line'>(in /home/sandi </div></div><div data-line='3' class='code-highlight-row numbered'><div class='code-highlight-line'>rake doc:plugin </div></div><div data-line='4' class='code-highlight-row numbered'><div class='code-highlight-line'>rake doc:plugin </div></div><div data-line='5' class='code-highlight-row numbered'><div class='code-highlight-line'>rake rails:upda </div></div><div data-line='6' class='code-highlight-row numbered'><div class='code-highlight-line'>rake rails:update:scripts </div></div><div data-line='7' class='code-highlight-row numbered'><div class='code-highlight-line'>rake stats </div></div><div data-line='8' class='code-highlight-row numbered'><div class='code-highlight-line'>rake test:units </div></div><div data-line='9' class='code-highlight-row numbered'><div class='code-highlight-line'>rake tmp:sockets:clear </div></div><div data-line='10' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:conf </div></div><div data-line='11' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:config </div></div><div data-line='12' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:in </div></div><div data-line='13' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:rebuild </div></div><div data-line='14' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:reindex </div></div><div data-line='15' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:restart </div></div><div data-line='16' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:run </div></div><div data-line='17' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:start </div></div><div data-line='18' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:stop </div></div><div data-line='19' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:version class='code-highlight-row numbered'><div class='code-highlight-line'>rake -T ts p/v1) s:acts_as_audited # Generate documentation for the acts_as… s:acts_as_ferret # Generate documentation for the acts_as… te:javascripts # Update your javascripts from your curre… # Add new scripts to the application scri… # Report code statistics (KLOCs, etc) fro… # Run tests for unitsdb:test:prepare / Ru… # Clears all files in tmp/sockets # Generate the Sphinx configuration file … # Generate the Sphinx configuration file … # Index data for Sphinx using Thinking Sp… # Stop Sphinx (if it’s running), rebuild … # Reindex Sphinx without regenerating the… # Restart Sphinx / Restart Sphinx # Stop if running, then start a Sphinx se… # Start a Sphinx searchd daemon using Thi… # Stop Sphinx using Thinking Sphinx’s set… # Output the current Thinking Sphinx vers…</div></div></pre></div>

Starting Sphinx

Start sphinx server this should give up an error.

<div class='code-highlight'><pre class='code-highlight-pre'><div data-line='1' class='code-highlight-row numbered'><div class='code-highlight-line'>rake ts:start </div></div><div data-line='2' class='code-highlight-row numbered'><div class='code-highlight-line'>(in /home/sandip/v1) </div></div><div data-line='3' class='code-highlight-row numbered'><div class='code-highlight-line'>Failed to start searchd daemon. Check /home/sandip/v1/log/searchd.log.</div></div></pre></div>

Adding indexes in models

define_index do
     # following fields are database fields
     # we can not add model methods in sphinx index 
     # sphinx fields allows ONLY model based associations
     # fields
      indexes customer_name
      indexes phone
      indexes mobile
      indexes other_phone
      indexes car_make
      indexes car_model
      indexes registration_no
end

Configure sphinx

rake ts:config
Generating Configuration to /home/sandip/v1/config/development.sphinx.conf
Index sphinx
 
rake ts:in
(in /home/sandip/v1)
Generating Configuration to /home/sandip/v1/config/development.sphinx.conf
Sphinx 0.9.8-rc2 (r1234)
Copyright (c) 2001-2008, Andrew Aksyonoff
 
using config file '/home/sandip/v1/config/development.sphinx.conf'...
indexing index 'call_core'...
collected 113521 docs, 2.1 MB
collected 0 attr values
sorted 0.1 Mvalues, 100.0% done
sorted 0.3 Mhits, 100.0% done
total 113521 docs, 2114790 bytes
total 6.102 sec, 346589.68 bytes/sec, 18604.78 docs/sec
distributed index 'call' can not be directly indexed; skipping.
Generating Configuration to /home/sandip/v1/config/development.sphinx.conf
Sphinx 0.9.8-rc2 (r1234)
Copyright (c) 2001-2008, Andrew Aksyonoff
 
using config file '/home/sandip/v1/config/development.sphinx.conf'...
indexing index 'call_core'...
collected 113521 docs, 2.1 MB
collected 0 attr values
sorted 0.1 Mvalues, 100.0% done
sorted 0.3 Mhits, 100.0% done
total 113521 docs, 2114790 bytes
total 3.628 sec, 582909.70 bytes/sec, 31290.34 docs/sec
distributed index 'call' can not be directly indexed; skipping.

Search using sphinx

ruby script console
 
>>Call.search 'sandip'
=> [#<call 00:00:00"="" 09:22:35",="" 1018,="" 18:10:00",="" 2009-02-10="" 2009-03-27="" 2009-04-07="" 2009-11-18="" 20:29:06",="" 6,="" created_at:="" id:="" insurance_expiry_date:="" is="" is_existing_customer:="" nil,="" note:="" phone="" ringing",="" scheduled_at:="" state_id:="" station_shift_customer_car_insurance_product_id:="" updated_at:="">, #<call 00:00:00"="" 11:13:33",="" 18:30:00",="" 2009-02-10="" 2009-02-16="" 2009-04-03="" 2009-11-18="" 20:34:52",="" 6,="" 6024,="" call="" created",="" created_at:="" customer="" id:="" insurance_expiry_date:="" is_existing_customer:="" new="" nil,="" note:="" scheduled_at:="" state_id:="" station_shift_customer_car_insurance_product_id:="" updated_at:="">] </call></call>

It returns array of records found. conditions are allowed by sphinx. If you need to add conditions on integer attributes then index block in model needs to have has method like author_id in following.

<div class='code-highlight'><pre class='code-highlight-pre'><div data-line='1' class='code-highlight-row numbered'><div class='code-highlight-line'>define_index do </div></div><div data-line='2' class='code-highlight-row numbered'><div class='code-highlight-line'> indexes content </div></div><div data-line='3' class='code-highlight-row numbered'><div class='code-highlight-line'> indexes :name, :sortable => true </div></div><div data-line='4' class='code-highlight-row numbered'><div class='code-highlight-line'> indexes comments.content, :as => :comment_content </div></div><div data-line='5' class='code-highlight-row numbered'><div class='code-highlight-line'> indexes [author.first_name, author.last_name], :as => :author_name </div></div><div data-line='6' class='code-highlight-row numbered'><div class='code-highlight-line'> </div></div><div data-line='7' class='code-highlight-row numbered'><div class='code-highlight-line'> has author_id, created_at </div></div><div data-line='8' class='code-highlight-row numbered'><div class='code-highlight-line'>end</div></div></pre></div>

using thinking sphinx search in rails app

Plugin installation

OR

Starting Sphinx

Adding indexes in models

Configure sphinx

Search using sphinx

Written by

Sandip Ransing

Supported by

Fun On Rails

Journal of a Web Developer #ruby #rails #JS