From: GOTO Kentaro <gotoken@...>
Date: 2002-09-27T11:59:47+09:00
Subject: Re: 'kakasi': different style for hiragana and katakana?

At Fri, 27 Sep 2002 08:18:10 +0900,
Rudolf Polzer wrote:

> In my kakasi-renamer I want to change filenames to Roumaji as follows:
> 
>   hiragana
>   KATAKANA
>   Kanji

As far as I know, kakasi doesn't have such function.

> How to do that? I'm thinking of running kakasi multiple times, perhaps:
> 
> Katakana->ASCII
> upcase
> Everything else->ASCII
> 
> But this would also upcase existing ASCII characters in the file name.

You can use gsub. 

  require "kakasi"

  def eucjp_katakana_to_upcase(str)
    katakana_from = "\xa5\xa1"  # small a
    katakana_end = "\xa5\xf6"   # small ke
    katakana = /[#{katakana_from}-#{katakana_end}]+/em # em for euc&multiline
    str.gsub(katakana){Kakasi::kakasi("-ka -Ka", $&).upcase}
  end

  if __FILE__ == $0
    puts eucjp_katakana_to_upcase(ARGF.read)
  end

> Currently I have
> 
>     if Kconv::guess (name) != Kconv::UNKNOWN
>       euc = Kconv::toeuc (name)

Kconv::guess and Kconv::toeuc uses different algorithms to determine
encoding and  "if Kconv::guess (name) != Kconv::UNKNOWN" may not work. 

-- Gotoken