Installing the wordnet gem on OSX, as of February 9 2012

gistfile1.md

Getting the wordnet gem, version 0.0.5, working is a little tricky at the moment. It's being rewritten but isn't released yet and the dbd gem is in a similar state.

These steps got it working for me on OS X Lion and MRI 1.8.7. On 1.9.2 convertdb.rb segfaults.

WordNet

Download WordNet from http://wordnet.princeton.edu/wordnet/download/current-version/

$ ./configure
$ make
$ make install

BerkeleyDB

This uses homebrew (http://mxcl.github.com/homebrew/). If you use a different package manager then you will have to adjust the --with-db-dir below.

$ brew install berkeley-db

ruby-bdb gem

$ git clone https://github.com/knu/ruby-bdb
$ cd ruby-bdb
$ git checkout 6aa8054e022fa38055fd441de98b65fd392b8dcb
$ ruby extconf.rb --with-db-dir=/usr/local/Cellar/berkeley-db/5.1.19
$ make
$ make install

wordnet gem

$ gem install wordnet
$ gem which wordnet
$ cd your/wordnet/path
$ ruby convertdb.rb
$ mv ruby-wordnet/ `ruby -e "require 'rbconfig'; print Config::CONFIG['datadir']"`

Using it

$ irb
> require 'rubygems'
> require 'wordnet'
> lex = WordNet::Lexicon.new
> frog = lex.lookup_synsets('frog', WordNet::Noun)
> frog.first.hypernyms.first.overview
  => "amphibian [noun] -- (cold-blooded vertebrate typically living on land but breeding in water; aquatic larvae undergo metamorphosis into adult form)" 

@lramach I had the same error messages as you, and it was caused by my system defaulting to ruby 1.8.7, when I tried with 1.9.3 it worked fine.

Thanks that helped! I was able to install ruby-bdb with ruby version 1.9.2.
However, while trying to install the wordnet gem and setting up the database with 'ruby convertdb.rb' I am getting the following error:
.rvm/rubies/ruby-1.9.2-p290/lib/ruby/1.9.1/fileutils.rb:243:in `mkdir': Permission denied - /Library/Ruby/Gems/1.8/gems/wordnet-0.0.5/ruby-wordnet (Errno::EACCES)
I downloaded and installed WordNet-3.0 in a folder that did not require any permission.
Any idea how I can fix this? Thanks again for all the help!

Did you previously install the WordNet gem as a superuser? Perhaps try an uninstall of WordNet and then try to re-install.

Thanks for that suggestion!
I tried 'sudo ruby convert.rb' and it worked!

I completed all the steps, but when I run

require 'rubygems'
require 'wordnet'
lex = WordNet::Lexicon.new
I get the following error:

TypeError: can't convert Symbol into Integer
from /ruby-1.9.2-p290/gems/wordnet-0.0.5/lib/wordnet/lexicon.rb:93:in %' from /ruby-1.9.2-p290/gems/wordnet-0.0.5/lib/wordnet/lexicon.rb:93:ininitialize'
from (irb):3:in new' from (irb):3 from /ruby-1.9.2-p290/bin/irb:16:in

'

Did anyone encounter this error?

I finally got it working, after making a copy of the ruby-wordnet directory in the '..rvm/rubies/ruby-1.9.2-p290/share/' directory and using 'lex = WordNet::Lexicon.new(WordNet::Lexicon::DEFAULT_DB_ENV, :044)' to create the lexicon!

Thanks for all the help!

Great! Glad to hear you got it working.

I just realized that the 'ruby convertdb.rb' had resulted in a segmentation fault.

convertdb.rb: [BUG] Segmentation fault
ruby 1.9.2p290 (2011-07-09 revision 32553) [x86_64-darwin11.2.0]

-- control frame ----------

c:0001 p:0000 s:0002 b:0002 l:0006d8 d:0006d8 TOP

-- C level backtrace information -------------------------------------------

[NOTE]
You may have encountered a bug in the Ruby interpreter or extension libraries.
Bug reports are welcome.
For details: http://www.ruby-lang.org/bugreport.html

Although wordnet is working fine for some cases, it results in segmentation faults while trying to lookup_synset. I looked around but couldn't find a solution to this. Did anyone face a similar issue? Any help would be appreciated! thanks!

For me this segmentation fault occurred on exit - so the convertdb.rb script executed 'correctly'

Same here. However, when I run my code for looking up synsets (lookup_synsets), it throws me the following error:

.rvm/gems/ruby-1.9.2-p290/gems/wordnet-0.0.5/lib/wordnet/synset.rb:309:in initialize': uninitialized constant WordNet::Synset::Sync (NameError) from .rvm/gems/ruby-1.9.2-p290/gems/wordnet-0.0.5/lib/wordnet/lexicon.rb:233:innew'
from .rvm/gems/ruby-1.9.2-p290/gems/wordnet-0.0.5/lib/wordnet/lexicon.rb:233:in block in lookup_synsets_by_key' from .rvm/gems/ruby-1.9.2-p290/gems/wordnet-0.0.5/lib/wordnet/lexicon.rb:227:ineach'
from .rvm/gems/ruby-1.9.2-p290/gems/wordnet-0.0.5/lib/wordnet/lexicon.rb:227:in lookup_synsets_by_key' from .rvm/gems/ruby-1.9.2-p290/gems/wordnet-0.0.5/lib/wordnet/lexicon.rb:214:inlookup_synsets'
from /Aptana Studio 3 Workspace/NLP_Test_0.1/wordnetBasedSimilarity.rb:43:in compareStrings' from /Aptana Studio 3 Workspace/NLP_Test_0.1/wordnetBasedSimilarity.rb:404:in

'
/Aptana Studio 3 Workspace/NLP_Test_0.1/wordnetBasedSimilarity.rb: [BUG] Segmentation fault
ruby 1.9.2p290 (2011-07-09 revision 32553) [x86_64-darwin11.2.0]

-- control frame ----------

c:0001 p:0000 s:0002 b:0002 l:000308 d:000308 TOP

-- C level backtrace information -------------------------------------------

[NOTE]
You may have encountered a bug in the Ruby interpreter or extension libraries.

Bug reports are welcome.

Edit: The lookup code works fine while running it from the irb console.

try finishing your code by closing the WordNet::Lexicon connection (e.g. for me this is lex.close)

But you would add the lex.close after the lookup_synsets() line, correct?
My code is failing on the line containing the lookup!