#!/usr/local/bin/ruby -Ke -T
# -*- coding: euc-jp -*-
##!/usr/bin/ruby -Ke -T
## Copyright (C) 2002, 2003 Mikio NAKAJIMA <minakaji@osaka.email.ne.jp>
## Copyright (C) 2002 Kazuhiro NISHIYAMA <zn@mbf.nifty.com>
## Copyright (C) 2004 Kimura Fuyuki <fuyuki@hadaly.org>
## Copyright (C) 2004 MITA Yuusuke <clefs@mail.goo.ne.jp>
##
## Author: Mikio NAKAJIMA <minakaji@osaka.email.ne.jp>
## Maintainer: SKK Development Team <skk@ring.gr.jp>
## Version: $Id: skkform.rb.in,v 1.5 2010/11/04 14:18:00 czkmt Exp $
## Keywords: japanese, dictionary
## Last Modified: $Date: 2010/11/04 14:18:00 $
##
## This program is free software; you can redistribute it and/or modify
## it under the terms of the GNU General Public License as published by
## the Free Software Foundation; either version 2, or (at your option)
## any later version.

## This program is distributed in the hope that it will be useful,
## but WITHOUT ANY WARRANTY; without even the implied warranty of
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
## General Public License for more details.

## You should have received a copy of the GNU General Public License
## along with this program, see the file COPYING.  If not, write to the
## Free Software Foundation Inc., 59 Temple Place - Suite 330, Boston,
## MA 02111-1307, USA.
##
### Commentary:
# http://www.rubycgi.org/cgi_explanation/2-7.htm を参考にした。
#
# アスキー刊の「オブジェクト指向スクリプト言語ruby」の 121 頁から拝借した
# csv_split 関数については、その著書の一人であるまつもとゆきひろさんに、
# パブリックドメイン扱いで良いとの許可を得た(2002/07/10)。
#
# <log format> -- csv format
# midasi,candidate,hinsi,yourei,goodict,daijisen,date,search_engine_cache,remote_host,remote_addr,daijisen_url_cache,goo_hit_number
# めしあg,召し上,その他,,○,×,07/09/02 08:17:56,"めしあがる,召し上がる",localhost,127.0.0.1
# ちょうば,跳馬,その他,,○,×,07/09/02 06:28:08,,localhost,127.0.0.1
# おじぎそう,含羞草,その他,,○,○,04/04/03 14:45:54,,localhost,127.0.0.1
# <TODO>

Debug = false

# $COMPAT_VERSION = "1.6"
# require "./cgisup"
require 'cgi'
require 'cgi-lib' if Debug
require 'jcode'
require 'kconv'
require 'socket'
require 'timeout'

CR = "\015"
LF = "\012"
CRLF = CR + LF
BR = "<br>"
# for debugging
if Debug
  SUCCESSIVE_ACCESS_SECONDS = 0
  TMPDIRECTORY = '/tmp/'
  DICDIRECTORY = '/home/minakaji/progs/ring/skk/dic/'
  PRIVATEBINDIR = '/usr/local/bin/'
  LOCKDIRNAME = '/home/minakaji/progs/ring/skk/web/test/registdic.log.lock'
  LOGDIRECTORY = '/home/minakaji/progs/ring/skk/web/test/'
  Logdirectory = './'
else
  SUCCESSIVE_ACCESS_SECONDS = 10
  TMPDIRECTORY = '/circus/openlab/skk/tmp/'
  DICDIRECTORY = '/circus/openlab/skk/skk/dic/'
  PRIVATEBINDIR = '/usr/home/minakaji/bin/'
  ALTERNATIVEBINDIR = '/usr/home/czkmt/bin/'
  LOCKDIRNAME = '/circus/openlab/skk/tmp/registdic.log.lock/'
  LOGDIRECTORY = '/circus/openlab/skk/log/'
  # Dir.glob(pattern) doesn't work if Logdirectory is a full path representation...?  (newcomments.cgi and search.cgi)
  # local tests could go w/o it, but is left just in case.
  Logdirectory = './log/'
end
LOGFILENAME = LOGDIRECTORY + 'registdic.log'
Logfilename = Logdirectory + 'registdic.log'


# cvs repositry 上には具体的な IP を登録しない。openlab.jp/skk に置くときに必要に応じ書く。
# BLACK_LIST = "/xxx.xxx.xxx.xxx|yyy.yyy.yyy.yyy/"
BLACK_LIST = ""

# reducing queries by reordering them in order of frequency
Okurigana = {
  "a" => ["あ"],
  "b" => ["び", "びる", "ぶ","ばしい", "ばむ", "ば",  "ばる", "べる", "べ", "びせる", "びせ", "ぼす", "ぼ"],
  "c" => ["ち", "ちる", "っちょろい"],
  "d" => ["だ", "ぢ", "づ", "で", "ど"],
  "e" => ["え", "える"],
  "f" => ["ふ", "へ", "ほ"],
  "g" => ["が", "がり", "がる", "ぎ", "ぎる", "ぐ", "げ", "げる", "ご"],
  "h" => ["は", "ひ", "ふ", "へ", "ほ"],
  "i" => ["い", "いる"],
  "j" => ["ざる", "じ", "じる"],
  "k" => ["き", "きる", "く", "ける", "か", "かう", "かし", "かしい", "かす", "かぶ", "かる", "きい", "くす", "け", "こ", "こす", "こえる", "こる"],
  "m" => ["む", "み", "め", "める", "ます", "まり", "まる", "まれる", "まえる", "ま", "も", "めく", "もしい", "もる"],
  "n" => ["ぬ", "に", "な", "ない", "なう", "なる", "ね", "ねる", "の", "ん"],
  "o" => ["お"],
  "p" => ["ぱ", "ぴ", "ぷ", "ぺ", "ぽ", "っぽい"],
  "r" => ["る", "り", "りる", "れ", "れる", "ろ", "ら", "らう", "らえる", "らか", "らぐ", "らしい", "らす", "らむ", "るい", "れむ", "れやか", "ろす", "ろし", "ろしい"],
  "s" => ["す", "する", "し", "しい", "さ", "さい", "さえる", "しむ", "せ", "せる", "そ"],
  "t" => ["つ" ,"ちる", "てる", "ち", "た", "たい", "たる", "っ", "って", "て", "と", "とす", "つく"],
  "u" => ["う", "うい"],
  "w" => ["わしい", "わす", "わる", "わせ", "わせる", "われる", "わ", "わう", "を"],
  "x" => ["ぁ", "ぃ", "ぅ", "ぇ", "ぉ", "ゃ", "ゅ", "ょ", "ゎ", "ゐ", "ゑ"],
  "y" => ["ゃ", "や", "やか", "やす", "ゅ", "ゆ", "ょ", "よ"],
  "z" => ["ず", "ずる", "じ", "じる", "ぜる", "ぜ", "ざる", "ざ", "ぞ"],
}

GyakuhikiOkurigana = [
  ["あ" , "a"],
  ["ば" , "b"], ["び" , "b"], ["ぶ" , "b"], ["べ" , "b"], ["ぼ" , "b"],
  ["だ" , "d"], ["ぢ" , "d"], ["づ" , "d"], ["で" , "d"], ["ど" , "d"],
  ["え" , "e"],
  ["が" , "g"], ["ぎ" , "g"], ["ぐ" , "g"], ["げ" , "g"], ["ご" , "g"],
  ["は" , "h"], ["ひ" , "h"], ["ふ" , "h"], ["へ" , "h"], ["ほ" , "h"],
  ["い" , "i"],
  ["じ" , "j"],
  ["か" , "k"], ["き" , "k"], ["く" , "k"], ["け" , "k"], ["こ" , "k"],
  ["ま" , "m"], ["み" , "m"], ["む" , "m"], ["め" , "m"], ["も" , "m"],
  ["な" , "n"], ["に" , "n"], ["ぬ" , "n"], ["ね" , "n"], ["の" , "n"], ["ん" , "n"],
  ["お" , "o"],
  ["ぱ" , "p"], ["ぴ" , "p"], ["ぷ" , "p"], ["ぺ" , "p"], ["ぽ" , "p"],
  ["ら" , "r"], ["り" , "r"], ["る" , "r"], ["れ" , "r"], ["ろ" , "r"],
  ["さ" , "s"], ["し" , "s"], ["す" , "s"], ["せ" , "s"], ["そ" , "s"],
  ["た" , "t"], ["ち" , "t"], ["っ" , "t"], ["つ" , "t"], ["て" , "t"], ["と" , "t"],
  ["う" , "u"],
  ["わ" , "w"], ["を" , "w"],
  ["ぁ" , "x"], ["ぃ" , "x"], ["ぅ" , "x"], ["ぇ" , "x"], ["ぉ" , "x"],
  ["ゃ" , "x"], ["ゅ" , "x"], ["ょ" , "x"], ["ゎ" , "x"], ["ゐ" , "x"], ["ゑ" , "x"],
  ["ゃ" , "y"], ["や" , "y"], ["ゅ" , "y"], ["ゆ" , "y"], ["ょ" , "y"], ["よ" , "y"],
  ["ざ" , "z"], ["じ" , "z"], ["ず" , "z"], ["ぜ" , "z"], ["ぞ" , "z"]
]

Hinsi = {
  "etc" => "その他",
  "geographic" => "地名",
  "firstname" => "姓",
  "lastname" => "名",
  "name" => "その他の人名(フルネームなど)",
  "propernoun" => "その他の固有名詞",
  "assoc" => "連想",
  "wrong" => "誤登録",
}

LogFields = {
  "key" => 0,
  "candidate" => 1,
  "hinsi" => 2,
  "yourei" => 3,
  "goodict" => 4,
  "yahoodict" => 5,
  "date" => 6,
  "search_engine_cache" => 7,
  "rh" => 8,
  "ra" => 9,
  "u" => 10,
  "goo" => 11,

  # aliases
  "midasi" =>0,
  "word" =>1,
  "daijisen" => 5,
  "goohits" => 11,
}

class Cvsdic
  def initialize(file)
    # @ldic = '/circus/openlab/skk/skk/dic/SKK-JISYO.L'
    @dic = file
  end

  def include?(key, word)
    if !(key && !key.empty?)
      return $errmsg = "<p><b>ERROR:</b> 見出し語がありません!"
    elsif !(word && !word.empty?)
      return $errmsg = "<p><b>ERROR:</b> 候補がありません!"
    end
#    expr2 = PRIVATEBINDIR + 'skkdic-expr2'
    expr2 = ALTERNATIVEBINDIR + 'skkdic-expr2'
    tmpdir = TMPDIRECTORY
    tmpfile_base = tmpdir + 'skkdic-expr2_' + "#{Process.pid}"
    result_string = false
    key = Kconv.toeuc(key)
    word = Kconv.toeuc(word)
    tmpfile0 = tmpfile_base + '.00'
    tmpfile1 = tmpfile_base + '.01'
    File.open(tmpfile0, "w+") do |file_handler|
      file_handler.print "#{key} /#{word}/\n"
    end
    command = expr2  + ' ' + tmpfile0 + ' ^ ' + @dic + ' >' + tmpfile1
    begin
      escaped_word = word.gsub(/[\[\]\{\}\(\)\|\+\*\.\^\$\&\\]/, '\\\\\&')
      print "<html><body>word=#{word}, escaped_word=#{escaped_word}<br>" if Debug
      if system(command)
	File.foreach(tmpfile1) do |line|
	  if (line =~ /#{key} \/#{escaped_word}(;[^\/]+)*\//o)
	    return true
	  end
	end
      end
    rescue Exception
      print "<html><body>ERROR: #$!\n</body></html>"
    ensure
      File.unlink(tmpfile0)
      File.unlink(tmpfile1)
    end
    return false
  end

  def include_key(key)
    if !(key && !key.empty?)
      return $errmsg = "<p><b>ERROR:</b> 見出し語がありません!"
    end
    tmpdir = TMPDIRECTORY
    tmpfile_base = tmpdir + 'grep_' + "#{Process.pid}"
    result_string = false
    key = Kconv.toeuc(key)
    tmpfile0 = tmpfile_base + '.00'
    tmpfile1 = tmpfile_base + '.01'
    File.open(tmpfile0, "w+") do |file_handler|
      escaped_key = key.gsub(/[\[\]\{\}\(\)\|\+\*\.\^\$\&]/, '\\\\\&')
      file_handler.print "^#{escaped_key} "
    end
    command = '/usr/bin/grep -f ' + tmpfile0 + ' ' + @dic + ' >' + tmpfile1
    begin
      if system(command)
	File.foreach(tmpfile1) do |line|
	  return line
	end
      end
    rescue Exception
      print "<html><body>ERROR: #$!\n</body></html>"
    ensure
      File.unlink(tmpfile0)
      File.unlink(tmpfile1)
    end
    return nil
  end
end

class String
  def to_katakana
    self.gsub(/う゛/, '\\1ヴ').tr('ぁ-ん', 'ァ-ン')
  end

  def to_hiragana
    self.gsub(/ヴ/, 'う゛').tr('ァ-ン', 'ぁ-ん')
  end

  def cut_off_prefix_postfix
    self.sub(/^[<>\?]([ーぁ-ん]+)$/, '\1').sub(/^([ーぁ-ん]+)[<>\?]$/, '\1')
  end

  # from 「オブジェクト指向スクリプト言語ruby」p121
  def csv_split(delimiter = ',')
    csv = []
    data = ""
    self.split(delimiter).each do |d|
      if data.empty?
	data = d
      else
	data += delimiter + d
      end
      if /^"/ =~ data
	if /[^"]"$/ =~ data or '""' == data
	  csv << data.sub(/^"(.*)"$/, '\1').gsub(/""/, '"')
	  data = ''
	end
      else
	csv << d
	data = ''
      end
    end
    raise "cannot decode CSV\n" unless data.empty?
    csv
  end

  def csv_quote
    self.gsub(/"/, '\\"').sub(/.*,.*/, '"\&"')
  end

  def csv_unquote
    self.sub(/^\"(.+)\"$/, '\1')
  end

  # 09/30/04 => 04/09/30
  def mdy2ymd
    self.sub(/^([0-9]*)\/([0-9]*)\/([0-9]*)/, '\3/\1/\2')
  end
end

#ファイルロッククラス
class FileLockDir
  def lock(lockdir)
    Dir.mkdir(lockdir)
    return true
  rescue Errno::EEXIST
    return false
  end

  def unlock(lockdir)
    Dir.rmdir(lockdir)
    return true
  rescue
    return false
  end
end

class Goo
  def search(key)
    hits = tries = 0
    key = Kconv.toeuc(key)
    key = CGI::escape(key)
    timeout(20) {
      begin
	tries += 1
	sock = TCPSocket.open("search.goo.ne.jp", 80)
	sock.printf("GET /web.jsp?DC=1&MT=\"%s\" HTTP/1.0\r\n\r\n", key)
	sock.readlines.each do |line|
	  line = Kconv.toeuc(line)
	  if (/<b>約*([,0-9]+)<\/b>件中/ =~ line)
	    hits = $1.gsub(/,/, '').to_i
	    return hits if hits > 0
	  # 0hits, or system error?
	  elsif (/<br>該当する検索結果が見当たりません。/ =~ line)
	  #elsif (/<br>検索システムから検索結果を正常に取得できなかった可能性があります。/ =~ line) next
	    return 0
	  end
	end
      ensure
	sock.close if sock
      end while tries < 3 && hits < 1
    }
    "unknown"
  end
end

# &kind=
GooDictType = {
  "japanese" => "jn",
  "jn" => "jn",
  "ej" => "ej",
  "je" => "je",
  #"all" => n/a
}
# &mode=
GooDictMode = {
  "forward" => 0,
  "full" => 1,
  "backward" => 2,
  "body" => 3,
  #"partial" => n/a
}
# &dtype=
YahooDictType = {
  "japanese" => 0,
  "jn" => 0,
  "ej" => 1,
  "je" => 1,
  "english" => 1,
  "all" => 2,
}
# &stype=
YahooDictMode = {
  "forward" => 0,
  "full" => 1,
  "partial" => 2,
  "backward" => 3,
  "body" => 4,
}

# superclass for misc. web dictionary service classes
class WebDict
  def initialize
    @found = '×'
    @cache = ''
  end

  def set_found(string)
    @found = string
  end

  def found
    @found
  end

  def cache
    @cache
  end

  def _search(key, type, host, query, flags = "")
    @found = '×'
    @cache = ''
    v = Array.new
    return v if key.nil?
    timeout(20) {
      begin
	sock = TCPSocket.open(host, 80)
	sock.print(query)
	sock.readlines.each do |line|
	  temp = _pickup(key, type, Kconv.toeuc(line))
	  v = v + temp if !(temp.empty?)
	  if flags =~ /generous/
	    temp = _pickup(key, type, Kconv.toeuc(line), flags)
	    v = v + temp if !(temp.empty?)
	  end
	end
	if !(v.empty?)
	  v.each do |candidate|
	    temp = candidate.gsub(/(.)(\1)/, '\1' + "々")
	    if !v.include?(temp)
	      v.push(temp)
	    end
	  end
	end
      ensure
	sock.close if sock
      end
    }
    if Debug
      print "<html><body>key=#{key}<br>"
      p key,type,host,query,flags
      print "<br>"
      p v
      print "<br>"
    end
    v
  end

  # NOTE: "generous" flag is a replacement of 'kind=jnr' / 'dtype=-1' kludges
  # that handled to_hiragana conversion
  def search_with_okuri_process(key, candidate, type = "japanese", mode = "full", flags = "generous")
    #key := 'そぶr', candidate := '素振'
    return false unless /^([ーぁ-ん]+)([a-z])$/ =~ key
    key = $1 #'そぶ'
    # perfect match always fails if forward match does
    return false if !(okurigana = Okurigana[$2])
    return false if mode == "full" && search(candidate, type, "forward", flags).empty?
    okurigana.each do |char|
      if search(key + char, type, mode, flags).include?(candidate + char)
	return (key + char), (candidate + char)
      end
    end
    false
  end
end

class GooDict < WebDict
  def search(key, type = "japanese", mode = "full", flags = "generous")
    key = key.cut_off_prefix_postfix
    _search(key, type, "dictionary.goo.ne.jp", "GET /search.php?MT=#{CGI::escape(key)}&kind=#{GooDictType[type]}&mode=#{GooDictMode[mode]}&DC=1 HTTP/1.0\r\n\r\n", flags)
  end

  def _pickup(key, type, line, flags = "")
    switch = GooDictType[type]
    v = Array.new
    if (switch == 'ej')
      if (/^<font color=\"#[a-z0-9]+\"><b>([^<>]+)<\/b><\/font>/ =~ line)
	v.push($1.gsub(/・/, ''))
      end
    elsif (switch == 'jn' && flags =~ /generous/)
      #<b>パンや</b> <sub>1</sub> 【―屋】</font>
      if /<b>([-・ーヴァ-ン]+)([-ぁ-ん]+)<\/b>[　― ーヴぁ-んァ-ン]+ *<sub>[ 0-9]+<\/sub> 【―[^―]+】/ =~ line
	temp1 = $1
	temp2 = $2
	temp1 = temp1.to_hiragana.sub(/-/, '')
	temp2 = temp2.sub(/-/, '')
	v.push(temp1 + temp2)
      elsif /<b>([-ぁ-ん]+)([-・ーヴァ-ン]+)<\/b>[　― ーヴぁ-んァ-ン]+ *<sub>[ 0-9]+<\/sub> 【[^―]+―】/ =~ line
	temp1 = $1
	temp2 = $2
	temp1 = temp1.sub(/-/, '')
	temp2 = temp2.to_hiragana.sub(/-/, '')
	v.push(temp1 + temp2)
	#<b>ようさい</b> <sub>0</sub> 【溶<sup>▼</sup>滓/<sup>▼</sup>鎔<sup>▼</sup>滓】
	#<b>ほうき</b>　はう―  <sub>1</sub> 【邦<sup>▼</sup>畿】</font>
      elsif /<b>([・ーヴァ-ンぁ-ん]+)<\/b>[　― ーぁ-ん]+ *<sub>[ 0-9]+<\/sub> 【.+】/ =~ line
	temp1 = $1
	temp1 = temp1.sub(/・/, '')
	v.push(temp1)
	@cache = temp1
	v.push(temp1.to_hiragana)
      end
    elsif (switch == 'jn')
      if (/【(.*)】/ =~ line) ||
	  #<font color="#cc3333"><b>マイル</b> <sub>1</sub> [mile]</font>
	  (/^<font color=\"#[a-z0-9]+\"><b>([^<>]+)<\/b> <sub>[0-9]+<\/sub> \[#{Regexp.quote(key)}\]<\/font>/ =~ line)
	target = $1
	target.gsub!(/<sup>▼<\/sup>/, '')
	target.gsub!(/<sup>▽<\/sup>/, '')
	while /([^】]+)】[^【]+【([^】]+)/ =~ target
	  target = $1 + "・" + $2
	end
	target.gsub!(/\//, '・')
	target.split("・").each {|word|
	  word.gsub!(/<[^<>]+>|〈|〉|《|》/, '')
	  if /(.+)（(.+)）/ =~ word
	    #【行う（行なう）】
	    v.push($1).push($2)
	  elsif /\((.+)\)/ =~ word
	    # <B>かばやき  【蒲焼(き)】  </B></TD>
	    # <B>くりごと  【繰(り)言】  </B>
	    temp = word
	    while /(.*)\(([^()]+)\)(.*)/ =~ temp
	      # with all characters with parens
	      temp = $1 + $2 + $3
	    end
	    v.push(temp)
	    temp = word
	    while /(.*)\(([^()]+)\)(.*)/ =~ temp
	      # without characters with parens
	      temp = $1 + $3
	    end
	    v.push(temp)
	    # XXX with/without characters with parens
	  else
	    v.push(word)
	  end
	}
      end
    end
    v
  end
end

# not used for the time being; needs rewriting anyway
class LycosDict < WebDict
  def search(key)
    @found = '×'
    v = Array.new
    return v if key.nil?
    key = key.cut_off_prefix_postfix
    url = false
    timeout(20) {
      begin
	sock = TCPSocket.open("jiten.www.infoseek.co.jp", 80)
	sock.sync = true
	url = "/Kokugo?qt=#{CGI::escape(Kconv.toeuc(key))}&sm=0&pg=result_k.html&col=KO"
	sock.printf("GET %s HTTP/1.0\r\nHost: jiten.www.infoseek.co.jp\r\n\r\n" , url)
	array = sock.readlines
	#XXX
	midasi= key.split("").join("-\?")
	array.each do |line|
	  line = Kconv.toeuc(line)
	  #    <FONT COLOR=#FF0000><b>おりまぜる</b>  【織(り)交ぜる】<font size=5>
	  if (/<FONT COLOR=#FF0000><b>#{midasi}<\/b>/ =~ line)
	    #<b>きゅうしょく</b>　きう―   【休職】<font size=5>
	    #<b>おりまぜる</b>	【織(り)交ぜる】<font size=5>
	    while (/<b>#{midasi}<\/b>[　 ]+[^<>]*【([^<>]+)】<font size=[0-9]+>/ =~ line)
	      temp = $1
	      line = $'
	      v = _pickup(temp, v)
	    end
	    #<big><b>きゅうしょく</b>　きう―  <font size=1>0</font> 【休職】</big>
	    #<big><b>ちてき-ざいさんけん</b> <font size=1>6</font> 【知的財産権】</big>
	  elsif (/<big><b>#{midasi}<\/b>[^<>]+<font size=1>[0-9]+<\/font> 【([^<]+)】<\/big>/ =~ line)
	    v = _pickup($1, v)
	  end
	end
      ensure
	sock.close if sock
      end
    }
    v
  end

  def _pickup(base, v)
    base.split("・").each do |word|
      # 織(り)交ぜる
      if /(.*)\(([ぁ-ん])\)(.*)/ =~ word
	temp = word
	while /(.*)\(([ぁ-ん])\)(.*)/ =~ temp
	  # with all characters with parens
	  temp = $1 + $2 + $3
	end
	v.push(temp) if !v.include?(temp)
	temp = word
	while /(.*)\(([ぁ-ん])\)(.*)/ =~ temp
	  # without characters with parens
	  temp = $1 + $3
	end
	v.push(temp) if !v.include?(temp)
	# XXX with/without characters with parens
      else
	v.push(word) if !v.include?(word)
      end
    end
    v
  end

  def search_with_okuri_process(key, candidate)
    #key := 'そぶr', candidate := '素振'
    return false unless /^([ーぁ-ん]+)([a-z])$/ =~ key
    key = $1 #'そぶ'
    if !(okurigana = Okurigana[$2])
      return false
    else
      okurigana.each do |char|
	v = search(key + char)
	if v.include?(candidate + char)
	  return [(key + char), (candidate + char)]
	end
      end
      false
    end
  end
end

# NOTE: now dic.yahoo.co.jp searches Daijirin if not matched in Daijisen
class YahooDict < WebDict
  def search(key, type = "japanese", mode = "full", flags = "generous")
    key = key.cut_off_prefix_postfix
    _search(key, type, "dic.yahoo.co.jp", "GET /bin/dsearch?p=#{CGI::escape(key)}&stype=#{YahooDictMode[mode]}&dtype=#{YahooDictType[type]}\r\n\r\n", flags)
  end

  def _pickup(key, type, line, flags = "")
    dtype = YahooDictType[type]
    v = Array.new
    if (dtype == 1) #英和・和英
      # <b>a・lo・ha</b>&nbsp;[発音記号]
      if (/^<b>(.+)<\/b>&nbsp;\[.+\]/ =~ line)
	v.push($1.gsub(/・/, ''))
      end
    elsif (dtype == 0 && flags =~ /generous/) #国語
      #temp = line.scan(/>[‐・ーヴァ-ンぁ-ん]+&nbsp;【([^<>]+)】</)
      temp = line.scan(/>[‐・ーヴァ-ンぁ-ん]+(?:&nbsp;)?【(.+?)】</)
      temp.each do |target|
	next if target.empty?
	target = target[0]
	target.gsub!(/<sup>.<\/sup>/, '')
	target.gsub!(/\//, '・')
	target.split("・").each {|word|
	  if /(.+)（(.+)）$/ =~ word
	    #【行う（行なう）】
	    v.push($1).push($2)
	  elsif /（(.+)）/ =~ word
	    #【<sup>×</sup>蒲焼（き）】
	    #【繰（り）言】
	    temp1 = word
	    while /(.*)（([^（）]+)）(.*)/ =~ temp1
	      # with all characters with parens
	      temp1 = $1 + $2 + $3
	    end
	    v.push(temp1)
	    temp1 = word
	    while /(.*)（([^（）]+)）(.*)/ =~ temp1
	      # without characters with parens
	      temp1 = $1 + $3
	    end
	    v.push(temp1)
	    # XXX with/without characters with parens
	  else
	    v.push(word)
	  end
	}
      end
    elsif (dtype == 0) #国語
      #temp = line.scan(/>([‐・ーヴァ-ンぁ-ん]+)&nbsp;【[^<>]+】</)
      temp = line.scan(/>([‐・ーヴァ-ンぁ-ん]+)(?:&nbsp;)?【[^<>]+】</)
      temp.each do |target|
	if !target.empty?
	  v.push(target[0].gsub(/[‐・]/, '').to_hiragana)
	  @cache = target[0].gsub(/[‐・]/, '').to_hiragana
	end
      end
    end
    v
  end
end

#class LogDict
def _print_line_header
  print <<-EOF
  <table border="1"><tbody><tr>
  <td align="center"><b>登録日時</b></td>
  <td align="center"><b>エントリ (見出し語 /単語/) </b></td>
  <td align="center"><b>品詞</b></td>
  <td align="center"><b>用例等</b></td>
  <td align="center"><a href="http://dictionary.goo.ne.jp"><b>大辞林</b></a></td>
  <td align="center"><a href="http://dic.yahoo.co.jp/"><b>大辞泉</b></a></td>
  <td align="center"><b><a href="http://search.goo.ne.jp">goo</a>ヒット数</b></td>
  <td align="center"><b>コメント</b></td>
  </tr>
  EOF
end

def _print_line(line, showcomment = false)
  rawkey,rawcandidate,hinsi,yourei,goodict,daijisen,date,search_engine_cache,rh,ra,url_cache,goo = line.csv_split
  rawkey.gsub!(/\\"/, '"')
  rawcandidate.gsub!(/\\"/, '"')
  yourei.gsub!(/\\"/, '"')
  _print_part1(rawkey, rawcandidate, hinsi, yourei, date, search_engine_cache)
  _print_goodict(goodict, rawkey, rawcandidate, search_engine_cache)
  _print_daijisen(daijisen, rawkey, rawcandidate, search_engine_cache)
  _print_goo(rawkey, rawcandidate, search_engine_cache, goo)
  _print_comment(rawkey, rawcandidate, showcomment)
end

def _print_part1(rawkey, rawcandidate, hinsi, yourei, date, search_engine_cache)
  if search_engine_cache && !search_engine_cache.empty?
    search_engine_cache = search_engine_cache.csv_unquote
    google_key, google_candidate = search_engine_cache.csv_split
  else
    google_key = rawkey.sub(/^([ーぁ-ん]+)[a-z]$/, '\1').cut_off_prefix_postfix
    google_candidate = rawcandidate
  end
  yourei = CGI::escapeHTML(yourei).gsub(/http:\/\/[-,.+~;\?\&=%\/\.0-9A-Za-z_#]+/, '<a href="\&">\&</a>')
  print "<tr><td>#{date}</td><td><a href=\"http://www.google.co.jp/search?q=%22#{CGI::escape(Kconv.tosjis(google_key))}%22+%22#{CGI::escape(Kconv.tosjis(google_candidate))}%22&ie=Shift_JIS&lr=lang_ja\">#{CGI::escapeHTML(rawkey)}</a> /<a href=\"http://www.google.co.jp/search?q=%22#{CGI::escape(Kconv.tosjis(google_candidate))}%22&ie=Shift_JIS&lr=lang_ja\">#{CGI::escapeHTML(rawcandidate)}</a>/</td><td>#{hinsi}</td><td>#{yourei}</td>"
end

def _print_goodict(goodict, rawkey, rawcandidate, search_engine_cache)
  if goodict == "×"
    print "<td><a href=\"http://dictionary.goo.ne.jp/search.php?MT=#{CGI::escape(rawcandidate)}&kind=jn&mode=0\">×</a></td>"
  elsif goodict == "△"
    print "<td><a href=\"http://dictionary.goo.ne.jp/search.php?MT=#{CGI::escape(rawkey)}&kind=ej&mode=1\">△</a></td>"
  else
    if search_engine_cache && !search_engine_cache.empty?
      search_engine_cache = search_engine_cache.csv_unquote
      goo_key,goo_candidate = search_engine_cache.csv_split
      #      elsif /^[ーぁ-ん]+[a-z]$/ =~ rawkey
      #	goo_key, goo_candidate = @goodict.search_with_okuri_process(rawkey, rawcandidate)
      #	if !goo_key
      #	  goo_key, goo_candidate = @goodict.search_with_okuri_process(rawkey, rawcandidate, 'jnr')
      #	end
    else
      goo_key = rawkey.cut_off_prefix_postfix
      goo_candidate = rawcandidate
    end
    if (goo_key && !goo_key.empty?) || (goo_candidate && !goo_candidate.empty?)
      if /[ーヴァ-ン]/ =~ goo_candidate
	print "<td><a href=\"http://dictionary.goo.ne.jp/search.php?MT=#{CGI::escape(goo_candidate)}&kind=jn&mode=1\">#{goodict}</a></td>"
      else
	print "<td><a href=\"http://dictionary.goo.ne.jp/search.php?MT=#{CGI::escape(goo_key)}&kind=jn&mode=1\">#{goodict}</a></td>"
      end
    else
      print "<td>×</td>"
    end
  end
end

def _print_daijisen(daijisen, rawkey, rawcandidate, search_engine_cache)
  if daijisen == "×"
    print "<td><a href=\"http://dic.yahoo.co.jp/bin/dsearch?p=#{CGI::escape(rawcandidate)}&stype=2&dtype=2\">×</a></td>"
  elsif daijisen == "△"
    print "<td><a href=\"http://dic.yahoo.co.jp/bin/dsearch?p=#{CGI::escape(rawkey)}&stype=1&dtype=1\">△</a></td>"
  else
    if search_engine_cache && !search_engine_cache.empty?
      search_engine_cache = search_engine_cache.csv_unquote
      daijisen_key,daijisen_candidate = search_engine_cache.csv_split
      #      elsif /^[ーぁ-ん]+[a-z]$/ =~ rawkey
      #	daijisen_key, daijisen_candidate = @daijisen.search_with_okuri_process(rawkey, rawcandidate)
      #	if !daijisen_key
      #	  daijisen_key, daijisen_candidate = @daijisen.search_with_okuri_process(rawkey, rawcandidate, 1, 0)
      #	end
    else
      daijisen_key = rawkey.cut_off_prefix_postfix
      daijisen_candidate = rawcandidate
    end
    if (daijisen_key && !daijisen_key.empty?) || (daijisen_candidate && !daijisen_candidate.empty?)
      if /[ーヴァ-ン]/ =~ daijisen_candidate
	print "<td><a href=\"http://dic.yahoo.co.jp/bin/dsearch?p=#{CGI::escape(daijisen_candidate)}&stype=1&dtype=0\">#{daijisen}</a></td>"
      else
	print "<td><a href=\"http://dic.yahoo.co.jp/bin/dsearch?p=#{CGI::escape(daijisen_key)}&stype=1&dtype=0\">#{daijisen}</a></td>"
      end
    else
      print "<td>×</td>"
    end
  end
end

def _print_goo (key, candidate, search_engine_cache, goo)
  if (!goo || goo.empty? || goo == "\n")
    goo = 'unknown'
  end
  if search_engine_cache && !search_engine_cache.empty?
    search_engine_cache = search_engine_cache.csv_unquote
    key, candidate = search_engine_cache.csv_split
  end
  print "<td><a href=\"http://search.goo.ne.jp/web.jsp?MT=%22#{CGI::escape(candidate)}%22\">#{goo}</a></td>"
end

def _print_comment_contents(logfilename, lines)
  # if not given, count them here
  if lines.nil?
    lines = 0
    File.foreach(logfilename.untaint) do |line|
      lines += 1
    end
  end
  File.foreach(logfilename.untaint) do |line|
    date,ip,host,comment,rate = line.csv_split
    comment = comment.csv_unquote
    comment = comment.gsub(/http:\/\/[-,.+~;\?\&=%\/\.0-9A-Za-z_#]+/, '<a href="\&">\&</A>')
    print "<b>#{lines}：#{date}  登録希望度：#{rate}</b><br>#{comment}<br><br>\n"
    lines -= 1
  end
end

def _print_comment(key, word, showcomment = false)
  logfilename = LOGDIRECTORY + CGI::escape(key) + '+' + CGI::escape(word) + '.log'
  if File.exist?(logfilename.untaint)
    lines = total_rate = 0
    File.foreach(logfilename.untaint) do |line|
      date,remote_addr,remote_host,comment,rate = line.chomp.csv_split
      total_rate += rate.to_i
      lines += 1
    end

    #comments = '○'
    comments = lines.to_s + '/' + total_rate.to_s
  else
    comments = '新'
  end
  print "<td><a href=\"registcomment.cgi?midasi=#{CGI::escape(key)}&word=#{CGI::escape(word)}\">#{comments}</a></td></tr>\n"

  # Experimental - display comments directly
  # XXX assumes _print_part1; I know it's not the place to be,
  # but it is here so that we don't have to count "lines" twice
  if comments != '新' && showcomment
    print "<tr><td>-</td><td align=center>〃</td><td>-</td><td>"
    _print_comment_contents(logfilename, lines)
    print "</td><td>-</td><td>-</td><td>-</td><td>-</td></tr>"
  end
end
#end

def has_in_log(key, candidate)
  begin
    key_mod = key.gsub(/\\/, '\\\\\\\\').gsub(/"/, '\\\\\\"').gsub(/[\(\)\[\]\*\.\^\$]/, '\\\\\&')
    candidate_mod = candidate.gsub(/\\/, '\\\\\\\\').gsub(/"/, '\\\\\\"').gsub(/[\(\)\[\]\*\.\^\$]/, '\\\\\&')
    print key, " , ", key_mod , " , ", candidate, " , ", candidate_mod, "." if Debug
    lineno = 0
    File.foreach(LOGFILENAME) do |line|
      lineno += 1
      return lineno, line if line =~ /^#{key_mod},"?#{candidate_mod}"?,/
    end
    return nil,nil
  rescue Errno::ENOENT
  end
end

# end of registdic.cgi