Using WebSolr with Java on Heroku

This is a sample project that demonstrates how to use WebSolr add on with Java on Heroku. WebSolr is essentially a hosted version of Apache Solr which is a really fast and flexible search and indexing platform that can search/index any data. This project also uses the JQuery Autocomplete plugin and SolrJ, a Java API to communicate with any Solr instance.

Note: This project is by no means a demonstration of all of Solr's capabilties. Please refer to the Solr wiki for more information on additional features and functionality that you can leverage with Solr.

Pre-requisites

• Basic knowledge of Java and Java servlets
• Knowledge of JQuery and using plugins
• Understanding of Solr and Solr Queries.

How does it work?

Most of the magic happens in two main components:

1. index.jsp : This contains the JQuery autocomplete code and includes the Autocomplete Javascripts and CSS.
2. SolrProxy.java : This is a servlet that acts as a proxy the WebSolr index URL that is exposed on provision the add-on to your heroku account.

Provisioning WebSolr

You can provision the WebSolr add-on to your Heroku app by using the command:

$ heroku addons:add websolr:cobalt

Note: WebSolr is not a free add-on. If you provision their "cobalt" add-on you will be charged $20 per month.

Once you provision this add-on, WebSolr will create a "default" index and add the WebSolr index URL to your Heroku config. To view the WebSolr index URL use the command:

$ heroku config
...
WEBSOLR_URL         => http://index.websolr.com/solr/...

Once you have the WebSolr index URL, you and send your Solr queries to this URL. To modify the Solr schema.xml and make updates to your Solr cofiguration, navigate to the WebSolr add-on under your app and that will open up a page to all the Solr indexes that you have defined. On click a specific index, you can look at the schema, synonym configurations, stop words etc.

Using SolrJ

SolrJ is a pure java client to communicate with your Solr instance. Refer to http://wiki.apache.org/solr/Solrj#Adding_Data_to_Solr on how to add data to your index using Solr. Refer to http://wiki.apache.org/solr/Solrj#Reading_Data_from_Solr on how to use Solr queries using SolrJ.

Using SolrJ and WebSolr

SolrProxy.java is the main Java servlet that does the communication with the WebSolr index URL. The code snippet that performs the query is below:

  SolrQuery query = new SolrQuery();
  query.setQuery( "text:"+queryParam );
  query.set("wt", "json");
  query.setFields("id","StreetAddress");
  query.setFacet(true);
  query.addFacetField("State");
  
  Long start,end;
  start = System.currentTimeMillis();
  QueryResponse queryResponse = server.query(query);
  end = System.currentTimeMillis();

To see a working example of this, go to http://websolr-java.herokuapp.com