solr_scaffold_template -- generate a simple solr analysis filter

This is a simple generator to get a mostly-novice up-and-running with a custom analysis filter for their Solr project. See the base project solr_scaffold for more information about this kind of filter, as well as how to easily subclass solr.StrField with your own analysis (optionally changing the stored version of the field as well).

Note: The "right" way to do this is almost certainly by getting solr_scaffold into a maven repository and/or creating a maven archetypes, but that seems like a lot of work before we know if anyone cares.

Step 1: Generate

Supposed you want to make a filter to lowercase everything (ignoring that there are already better options).

What should we call it? Must be a java classname and hence start with an uppercase letter. So...Lowercasify
What package should it be in? Filters in solr in are something. something.solr.analysis, so I'll use com.billdueber.solr.analysis

git clone git@github.com:billdueber/solr_scaffold_template lowercasify
cd lowercasify
ruby generate.rb com.billdueber.solr.analysis Lowercasify

Step 2: Edit the filter

Two files were generated at the end of your package hierarchy under src/main/java: LowercasifyFilter.java and LowercasifyFilterFactory.java.

LowercasifyFilterFacoty.java is probably fine just as it is, and you can leave it alone.

LowercasifyFilter.java has a method in it, munge, which is where you put your string transformation logic. It can all live in there, or you can create other java files, pull other stuff in via the POM, etc.

package com.billdueber.solr.analysis;


import com.billdueber.solr_scaffold.analysis.SimpleFilter;
import org.apache.lucene.analysis.TokenStream;

/**
 *  For most cases, all you need to do is edit the `munge` method
 *  and leave the constructor alone.
**/

public class LowercasifyFilter extends SimpleFilter {

  public LowercasifyFilter(TokenStream aStream, Boolean echoInvalidInput) {
    super(aStream, echoInvalidInput);
  }

  @Override
  public String munge(String str) {
//    return str;
    return str.toLowerCase(); // I M SO SMRT!
  }

}

Step 3: Build it

mvn package

That was easy.

Step 4: Get the .jar files where your solr will find them.

There are two .jar files you'll need to grab

target/yourfilter-yourversion.jar
repo/com/billdueber/solr_scaffold/1.0/solr_scaffold-1.0.jar

These need to be put where your solr can find them. This is often controlled via the solrconfig.xml file; I put this line in mine

 <lib dir="${solr.core.config}/lib" regex=".*\.jar"/>

...and then I can have a lib/ directory right next to conf/ in my solr configuration.

Step 5: Use it in your schema.xml