From: Axel Etzold <AEtzold@...>
Date: 2008-09-04T01:01:14+09:00
Subject: Re: How to match words that rhyme?


-------- Original-Nachricht --------
> Datum: Wed, 3 Sep 2008 23:09:05 +0900
> Von: Lloyd Linklater <lloyd@2live4.com>
> An: ruby-talk@ruby-lang.org
> Betreff: Re: How to match words that rhyme?

> Here is a crude start based on an old version I wrote in Pascal.  It 
> should give you a place to start.   (Please do not be TOO brutal in 
> critique.  I am new with writing Ruby algorithms. Elegance will come.)
> 
> class Soundex
>   def soundsAlike word1, word2
>     soundex(word1) == soundex(word2)
>   end
> 
>   def soundex w
>     w.upcase!
>     temp1 = w[0,1]
>     1.upto(w.length - 1) { |i|
>       case w[i, 1]
>       when 'B','F','P','V'
>         temp1 += '1'
>       when 'C','G','J','K','Q','S','X','Z'
>         temp1 += '2'
>       when 'D','T'
>         temp1 += '3'
>       when 'L'
>         temp1 += '4'
>       when 'M', 'N'
>         temp1 += '5'
>       when 'R'
>         temp1 += '6'
>       end
>     }
>     i = 1
>     while i < temp1.length - 1 do
>       if temp1[i] == temp1[i + 1]
>         temp1[i] = ''
>       else
>         i += 1
>       end
>     end
>     temp1
>   end
> end
> 
> s = Soundex.new
> p s.soundsAlike('their', 'there')  # returns true
> -- 
> Posted via http://www.ruby-forum.com/.

Another solution, which could help the problem of non-phonetic writing in English is to use
the external text-to-speech program espeak:

http://espeak.sourceforge.net/

It can produce a kind of phonetic writing for English like so:

axel@alecrim:~$ espeak -v mb-en1 'dough' > res
axel@alecrim:~$ cat res
d	65
@U	294	 0 103 80 77 100 77
_	263
_	1
axel@alecrim:~$ espeak -v mb-en1 'tough' > res2
axel@alecrim:~$ cat res2
t	100
V	129	 0 103 80 77 100 77
f	142
_	263
_	1
axel@alecrim:~$ espeak -v mb-en1 'bow' > res3
axel@alecrim:~$ cat res3
b	65
@U	294	 0 103 80 77 100 77
_	263
_	1

... this works also in other languages for which the correspondence between pronounciation and
spelling isn't one-to-one, e.g., French:

axel@alecrim:~$ espeak -v mb-fr1 'eaux' > res4
axel@alecrim:~$ cat res4
o	202	 0 103 80 77 100 77
_	263
_	1
axel@alecrim:~$ espeak -v mb-fr1 'o' > res5
axel@alecrim:~$ cat res5
o	202	 0 103 80 77 100 77
_	263
_	1


espeak has many "voices" for different languages. For the above to work, you'll need to choose Mbrola
voices, because the text-to-speech system Mbrola (http://tcts.fpms.ac.be/synthesis/mbrola.html) needs
this phonetic text as input.

Best regards,

Axel 



-- 
Ist Ihr Browser Vista-kompatibel? Jetzt die neuesten 
Browser-Versionen downloaden: http://www.gmx.net/de/go/browser