In my use case, I’d like to filter a list of questions by a group of categories on Answerbag. This is used primarily for partner sites who’ve opted into our white label cobrand system and have selected a list of categories to restrict the site to. It affects search results, as the goal is to limit searches to these categories.

Technically, this can be done on 2 levels. First, it might be done in code, where we bring back a massive result set and filter out the questions that don’t satisfy our requirement. Ideally it can be done on the search engine side, which would reduce network overhead and eliminate a few really weird bugs that can crop up as a result.

It turns out this is pretty easy – but you must send it through as a single filter query, and not separate params for each filter.

To limit your filters to certain categories, you’d add something along these lines to your GET request, assuming you’ve stored categories by their numeric primary key:

fq=category_id:(24 79 84 693 694 1270 2309 2821 3284)

As usual, comments and thoughts are encouraged.

• Make a copy of the example directory found in the solr directory. To make obvious, I just named it after the site I’m working on.
• You’ll want to edit the solr/conf/schema.xml file found in there. It’s actually a really good sample file to get started with, so you’ll learn a whole bunch by starting there. Information on the solr schema. You’ll need to modify the configuration file for solr to function correctly with your schema.
• If you’re using a PHP and curl, and you get the error “missing content stream”, you need to make sure you set your curl header
  

  curl_setopt($c, CURLOPT_HTTPHEADER, array(”Content-Type: text/xml”));

• At one point, I received the error “Unexpected EOF in prolog” – this was because I originally had a “\n” at the end of my header (see previous note). Removing it solved my problem.

Here’s a good solr starting point that covers the schema and field types.

Let’s get started. You’ll need to check to see if Java is up to date. Run the below to find out.

java -version

You need to be running at least Java 1.5. Next is to check to see if Ant is installed. Do that with.

ant -version

I’m doing this on my Mac using Leopard. Here’s some instructions on setting up Ant if you aren’t using Leopard.

Java was up to date, and ant is built in. Sweet. However, I need to install JUnit. I’ll do that first.

Download JUnit. You’ll need to put it somewhere that’s accessible by the $CLASSPATH variable. There’s more information on the JUnit FAQ.

I threw it in /usr/share and left the name as junit-4.4.jar and set my CLASSPATH to point to that file (not the directory)

export CLASSPATH=$CLASSPATH:/usr/share/junit-4.4.jar

I ran that, as well as put it in my /etc/bashrc file (which you must be root to edit) so I don’t have to deal with it again.

Compile Solr

Switch back to the directory containing the solr files, and run:

ant compile

You should see something like this:

Buildfile: build.xml

init-forrest-entities:

checkJunitPresence:

compile:
[javac] Compiling 185 source files to /Users/jhaddad/src/apache-solr-1.2.0/build
[javac] Note: Some input files use or override a deprecated API.
[javac] Note: Recompile with -Xlint:deprecation for details.
[javac] Note: Some input files use unchecked or unsafe operations.
[javac] Note: Recompile with -Xlint:unchecked for details.

BUILD SUCCESSFUL
Total time: 3 seconds

I then ran:

ant dist

Which outputted something like this:

Buildfile: build.xml

init-forrest-entities:

checkJunitPresence:

compile:

make-manifest:
[mkdir] Created dir: /Users/jhaddad/src/apache-solr-1.2.0/build/META-INF

dist-jar:
[jar] Building jar: /Users/jhaddad/src/apache-solr-1.2.0/dist/apache-solr-1.2.1-dev.jar

dist-war:
[war] Building war: /Users/jhaddad/src/apache-solr-1.2.0/dist/apache-solr-1.2.1-dev.war

dist:

BUILD SUCCESSFUL
Total time: 0 seconds

You can run the example by going to the example directory and running

java -jar start.jar

Then go here: http://localhost:8983/solr/admin/ and check out your admin.

Load a few sample docs by going here:

/example/exampledocs

and running

java -jar post.jar solr.xml monitor.xml

I will post a follow up on how to get Solr running in Tomcat, as well as examples on how to use the server.