Blog

The last decade has seen our consumption of information skyrocket, but our capacity to manually organize our digital life simply crumble.

Think of it: ten years ago, we would create folders to organize our files, our music, our mail, but nowadays, we simply store our information into generic folders, and as long as each piece of information is properly tagged, we let the search engines organize our life. One could argue that this was predictable, because this is how the brain works. But there has been several formal studies on the subject, so this evolution wasn’t completely blind.

One of the key ingredients of Drupal’s success is the concept of generic nodes and attributes which we don’t care how and where they are stored. What is important is the capacity to retrieve information using parameters, through the Search and Views modules.

Drupal’s core search module can be replaced with Apache Solr, a web service which includes the Lucene engine also powering Alfresco. On web sites where Alfresco is used to store documents, this has the advantage of bringing one uniform search syntax to every search query. Furthermore, since Solr is called via a REST interface, you can install it on dedicated server, which means that your website performance won’t degrade during periods where the search activity is more intense.

Here are some other features of the Lucene engine:

1. Scored results

2. Highlighter (to show words found in context)

3. Fuzzy search

4. Query spellchecker

Solr turns Lucene into a web service and provides:

1. Faceting of query results

2. “More like this” plugin

3. “Did you mean” plugin

Installation

To install Solr, you must first setup Tomcat. If you have Alfresco already installed, you can run Solr from the same Tomcat instance.

Once Tomcat is installed, you need to download the binary distribution of Solr and copy the “example” folder to one area of your choice. On my system, my package manager installs Tomcat into /opt/local/sharejava/tomcat, therefore I chose to install Solr in a subdirectory of /opt/local/share/java/apache-solr/, next to the /opt/local/share/java/apache-ant directory:


tar -xzf apache-solr-1.4.0.tgz
cp apache-solr-1.4.0/example/solr /opt/local/share/java/apache-solr/example.dev
cp apache-solr-1.4.0/example/webapps/solr.war /opt/local/share/java/apache-solr/example.dev/bin/apachesolr-1.4.0.war


Now install the Drupal ApacheSolr module and copy schema.xml and solrconfig.xml to the conf folder of your Solr instance.


cp apachesolr/schema.xml /opt/local/share/java/apache-solr/example.dev/conf
cp apachesolr/solrconfig.xml /opt/local/share/java/apache-solr/example.dev/conf


Then create a startup script for Tomcat inside /opt/local/share/java/apache-solr/example.dev:


<?xml version="1.0" encoding="utf-8"?>
<Context docBase="/opt/local/share/java/apache-solr/example.dev/bin/apache-solr-1.4.0.war" debug="0" crossContext="true">
<Environment name="solr/home" type="java.lang.String" value="/opt/local/share/java/apache-solr/example.dev" override="true"/>
</Context>


Then symlink the configuration to your Tomcat instance:


ln -s /opt/local/share/java/apache-solr/example.dev/apache-solr-example.dev.xml /opt/local/share/java/tomcat6/conf/Catalina/local/apache-solr-example.dev.xml


Finally, restart Tomcat and configure Drupal ApacheSolr to call the Solr path /apache-solr-example.dev and you are good to go!

For more information:
Solr: http://lucene.apache.org/solr/features.html
Lucene: http://lucene.apache.org/
Drupal module: http://drupal.org/project/apachesolr