java_bin

This is an Apache Solr JavaBin format (binary format) implementation for Ruby.

Features

fast parse(2.5 - 5 times faster than ruby eval), and less network traffic. MRI 1.8.7 [data2] ruby eval parse. 5000 times. elapsed time 6.664981 javabin parse. 5000 times. elapsed time 1.472825 4.52530409247534 times

REE 1.8.7 [data2] ruby eval parse. 5000 times. elapsed time 4.844482 javabin parse. 5000 times. elapsed time 1.059044 4.57439162112245 times

YARV 1.9.2 pre1 [data2] ruby eval parse. 5000 times. elapsed time 6.659331328 javabin parse. 5000 times. elapsed time 1.261828473 5.277525012704322 times

rsolr/solr-ruby support.
pure and c extension code.
ruby 1.8/1.9/2.0 support.
solr 1.4/3.1 support.

Requirements

Ruby1.8.7 or later (include 1.9.x, 2.0.x)
(recommended) C compiler (gcc or vc++): you can also use java_bin without c extension, but 'pure' is 30 times slower than 'ext'.
JavaBin has been tested with MRI 1.8.7, REE 1.8.7, YARV 1.9.2, YARV 1.9.3 and 2.0.0 on CentOS6.3 (x86_64bit) with Apache Solr 1.4, 3.1

Install

(Linux) gem sources -a http://gemcutter.org gem install java_bin

(Windows) fix (path to)\ruby\lib\ruby\1.8\i386-mswin32\config.h #if _MSC_VER != 1200 ↓ #if _MSC_VER < 1200

start => visual studio c++ 200x xxxx edition => visual studio tools => visual studio 200x command prompt gem sources -a http://gemcutter.org gem install java_bin or cd (path to)\ext\java_bin\ext ruby extconf.rb nmake

Simple Usage

require 'rubygems' require 'java_bin'

@parser = JavaBin.parser.new result = @parser.parse( javabin format string )

With RSolr/Solr-Ruby

By RSolr/Solr-Ruby support, a parser automatically uses javabin format instead of ruby format.

require 'rubygems' require 'rsolr' require 'java_bin'

solr = RSolr.connect solr.select :q => ':'

Valgrind/Memory Leaks

In this project we use Valgrind (http://www.valgrind.org) to ensure that the native extension is leak-free.
On any Debian-based Linux you simply need to install the valgrind package:

apt-get install valgrind

On homebrew-powered OSX, simply type:

brew install valgrind

Then invoke the Rake task. If any errors are present, it will fail.

rake valgrind

TODO

more parse speed
license
shared string over parsing
builder(writer)

References

javabin http://wiki.apache.org/solr/javabin
javabin codec (for java) http://svn.apache.org/viewvc/lucene/solr/trunk/src/common/org/apache/solr/common/util/JavaBinCodec.java?view=markup
rsolr http://github.com/mwmitchell/rsolr/
solr-ruby http://wiki.apache.org/solr/solr-ruby

Note on Patches/Pull Requests

Fork the project.
Make your feature addition or bug fix.
Add tests for it. This is important so I don't break it in a future version unintentionally.
Commit, do not mess with rakefile, version, or history. (if you want to have your own version, that is fine but bump version in a commit by itself I can ignore when I pull)
Send me a pull request. Bonus points for topic branches.

Author

Toshinori Kajihara mailto:[email protected]