[ann] double_metaphone.rb

ruby-talk
1

http://tfletcher.com/lib/double_metaphone.rb

Adapted from the PHP version. There's at least one person looking for
it:

http://groups.google.com/group/comp.lang.ruby/search?group=comp.lang.ruby&q=double+metaphone

Bug reports/fixes welcome!

2

Tim Fletcher wrote:

http://tfletcher.com/lib/double_metaphone.rb

Adapted from the PHP version. There's at least one person looking for
it:

Very cool! Maybe cook up a gem?

I've played with it for a few minutes, and look!

   require 'double_metaphone'

   module SpellChecker
     def self.words
       @words ||= {}
     end

     def self.add_word(word)
       word.downcase!
       raise ArgumentError unless word =~ /^[a-z]+$/
       words[word.to_sym] = DoubleMetaphone[word]
     end

     def self.valid? word
       words.has_key? word.to_sym
     end

     def self.check(word)
       word.downcase!

       if valid? word
         puts "`#{word}' is valid"
       else
         puts "`#{word}' is invalid. did you mean one of these?"
         key = DoubleMetaphone[word]
         words.select{|w, k| k == key}.each do |alt|
           puts " * #{alt.first}"
         end
       end
     end
   end

A bit rough, but hey.

Thanks mate!

Daniel

3

I do have one comment though: perhaps have the key be a symbol instead of a string, since we're probably not gonna alter it. E.g.

   DoubleMetaphone["foo"] => [:F, nil]

It may just be a personal preference, though.

Cheers,
Daniel

4

Not sure it's worth a gem, as it's quite easy just to drop in.

As for symbols/strings, the string is already there and generated, so
might as well return it. It's easy enough to use to_sym, if that's what
your application needs/prefers.

Forgot to mention: if anyone's looking for the regular (non Double)
Ruby metaphone, it's at http://po-ru.com/projects/metaphone/

5

Tim Fletcher wrote:

As for symbols/strings, the string is already there and generated, so
might as well return it. It's easy enough to use to_sym, if that's what
your application needs/prefers.

The string gets GC'ed, though. If you're going to store lots of keys, it may be better to store them as symbols.

Cheers,
Daniel

6

Oh, that's one of mine. Perhaps we should package Metaphone, Double
Metaphone, and Levenshtein[1] together - perhaps with Soundex[2] too.

Paul.

1. http://po-ru.com/projects/levenshtein/
2. http://raa.ruby-lang.org/project/soundex/

···

On 18/08/06, Tim Fletcher <twoggle@gmail.com> wrote:

Not sure it's worth a gem, as it's quite easy just to drop in.

As for symbols/strings, the string is already there and generated, so
might as well return it. It's easy enough to use to_sym, if that's what
your application needs/prefers.

Forgot to mention: if anyone's looking for the regular (non Double)
Ruby metaphone, it's at http://po-ru.com/projects/metaphone/

7

Not sure whether you're referring to the function or the memoization -
is it the memo hash _keys_ you think should be stored as symbols, or
the double_metaphone return _values_?

8

Tim Fletcher wrote:

Not sure whether you're referring to the function or the memoization -
is it the memo hash _keys_ you think should be stored as symbols, or
the double_metaphone return _values_?

Both.

Cheers,
Daniel