#!/usr/local/bin/ruby
# review search (html decode)
# 2003.2.7 ver.0.1 by ippo - 2003.2.18 ver.0.?

#require 'simpleuri'
require 'webagent'
require 'kconv'

class ReviewSearch
  def initialize
    @script_url = ENV['SCRIPT_URI'].to_s
#    @script_url = 'http://' + ENV['REMOTE_HOST'].to_s + ENV['REQUEST_URI'].to_s
    #REMOTE_HOST = REMOTE_ADDR ?
    @stderr = $stderr
  end

#========== 取得・出力 ========== ========== ==========

  def decode_html( url = nil, flag_decode = nil, reviewer = nil, flag_name = nil, mode = nil, bodyfile = nil )
    head, file = File.split(url)
    if url.sub(/^http:\/\//,'') !~ /\//
      url = url + '/'
    else
      url = url + '/' if file !~ /\./
    end
    
    list = []
    return list if url.nil?
    message = 'CHK; access; ' + url
    @messageKcode  = 3 #1-jis,2-euc,3-sjis
    @insystemKcode = 2 #euc
#    @stderr.puts message.kconv(@messageKcode)
    body = bodyfile
    if body.nil?
      agent = WebAgent.new()
      agent.uri = url
      agent.get()
      body = agent.body.kconv(@insystemKcode)
      body = engine_a_decode(body, flag_name)
    end
    case flag_decode
    when 'amazon'
      list = decode_amazon(body, flag_name)
    when 'bk1'
      list = decode_bk1(body, flag_name)
    when 'none'
      list = decode_none(body, flag_name)
    when 'bk1_bibid'
      list = decode_bk1_bibid(body, flag_name)
    else
      list = decode_refer(body, flag_name)
    end
    reviewer = reviewer.to_s.strip
    if reviewer != ''
      list.each_index{ |i| list[i][-1] = reviewer }
    end
    if mode.nil?
      url_org = URI.parse(url)
      list.each_index{ |i|
        if flag_name == 'name'
          name = list[i][1].to_s
          list[i][1] = url.sub(/#.*$/,'')
          list[i][1] += '#' + name if name != ''
        else
          list[i][1] = url_org.merge(list[i][1]).to_s
        end
      }
    else
      head, file = File.split(url)
      file.gsub!(/\#.*$/,'')
      list.each_index{ |i|
        if flag == 'name'
          if list[i][1].to_s == ''
            list[i][1] = file
          else
            list[i][1] = file + '#' + list[i][1]
          end
        end
      }
    end
=begin
    if flag_name == 'name'
      head, file = File.split(url)
      file.gsub!(/\#.*$/,'')
      list.each_index{ |i| #url:name->file#name
        list[i][1] = file + '#' + list[i][1] if list[i][1].to_s != ''
        list[i][1] = file if list[i][1].to_s == ''
      }
    end
    if mode.nil?
      head, file = File.split(url)
      list.each_index{ |i| #url:local->global
        list[i][1] = file.gsub(/#.*?$/,'') + list[i][1] if list[i][1] =~ /^#/
        list[i][1] = File.join(head, list[i][1])
      }
    end
    list.each_index{ |i|
      list[i][1] = url_check(list[i][1])
    }
=end
    return list
  end

#========== 小物 ========== ========== ==========

def url_check ( url )
  dirs = url.sub(/^http:\/\//mi,'').split('/')
  dirs.delete('')
  dirs.delete('.')
  while dirs.include?('..')
    i = 1
    while dirs[i] != '..'
      i += 1
    end
    dirs = dirs[0..i-2] + dirs[i+1..-1]
  end
  url = 'http://' + dirs.join('/')
  return url
end

#========== 展開・解析 ========== ========== ==========

  def engine_a_decode( text , flag_name = nil)
    texts = text.split(/<a/mi)
    newtexts = []
    newblock, newname, newurl = '', '', ''
    name, url = '', ''
    newblock << texts.shift
    texts.each{ |block|
      flag = false
      block =~ /^([^>]+>)/mi
      $1.to_s.split.each{ |check|
        case flag_name
        when 'name'
          if (check.strip =~ /^id=([^>]+)/mie)
            newname = $1
            flag = true
          end
          if (check.strip =~ /^id=\"(.+?)\"/mie)
            newname = $1
            flag = true
          end
          if (check.strip =~ /^name=([^>]+)/mie)
            newname = $1
            flag = true
          end
          if (check.strip =~ /^name=\"(.+?)\"/mie)
            newname = $1
            flag = true
          end
        else
          if (check.strip =~ /^href=\"(.+?)\"/mi)
            tmpurl = $1
            if (tmpurl !~ /^http:\/\//) && (tmpurl !~ /^mailto:/)
              flag = true
              newurl = tmpurl
            end
          end
        end
      }
      if flag
        newtexts << [newblock, name, url]
        newblock, name, url = '', newname, newurl
        newblock << '<a'
      end
      newblock << block
    }
    newtexts << [newblock, name, url] if newblock.size > 0
    return newtexts
  end

  def decode_refer( text , flag = nil) #独自タグ解釈
    list = []
    texts = text
    texts = engine_a_decode(text, flag) if text.type == 'String'
    texts.each{ |block, blockname, blockurl|
      blockurl = blockname if flag == 'name'
      block.scan(/<!--Tag:reviewlink-refer:(.+?)-->/){ |str|
        code, url, reviewer, urlmemo, point, memo = '', blockurl, '', '', '', ''
        vals = str.first.split(':')
        vals.each_index{ |i|
          var = vals[i]
          code     = var.strip                      if var =~ /^ISBN/
          url      = var.sub(/^url=/,'').strip      if var =~ /^url=/
          reviewer = var.sub(/^reviewer=/,'').strip if var =~ /^reviewer=/
          urlmemo  = var.sub(/^urlmemo=/,'').strip  if var =~ /^urlmemo=/
          point    = var.sub(/^point=/,'').strip    if var =~ /^point=/
          memo     = var.sub(/^memo=/,'').strip     if var =~ /^memo=/
          url      = 'http:' + vals[i+1].strip      if var =~ /^url=http$/
        }
        list << [code, url, reviewer, urlmemo, point, memo] if (code + url + reviewer) != ''
      }
    }
    list = list.uniq
    return list
  end

=begin
  def decode_refer( text )
    list = []
    text.scan(/<!--Tag:reviewlink-refer:(.+?)-->/) { |str|
      code, url, reviewer = '', '', ''
      str.first.split(':').each{ |var|
        code     = var.strip                      if var =~ /^ISBN/
        url      = var.sub(/^url=/,'').strip      if var =~ /^url=/
        reviewer = var.sub(/^reviewer=/,'').strip if var =~ /^reviewer=/
      }
      list << [code, url, reviewer] if (code + url + reviewer) != ''
    }
    return list
  end
=end

  def decode_amazon( text, flag = nil )
    list = []
    texts = text
    texts = engine_a_decode(text, flag) if text.type == 'String'
    texts.each{ |block, blockname, blockurl|
      blockurl = blockname if flag == 'name'
      code, url, reviewer = '', blockurl, ''
      block.scan(/href=\"http:\/\/www.amazon.co.jp\/exec\/obidos\/ASIN\/(.+?)\//mi){ |str|
        asin = str.first
        if asin =~ /^(\d+X?)$/i #CDとかのASINは先頭が数字でない様子
          code = 'ISBN' + $1.upcase.strip
          list << [code, url, reviewer]
        end
      }
    }
    list = list.uniq
    return list
  end

  def decode_bk1( text, flag = nil )
    list = []
    texts = text
    texts = engine_a_decode(text, flag) if text.type == 'String'
    texts.each{ |block, blockname, blockurl|
      blockurl = blockname if flag == 'name'
      code, url, reviewer = '', blockurl, ''
      block.scan(/href=\"http:\/\/www.bk1.co.jp\/cgi-bin\/srch\/srch_result_book.cgi(.+?)\"/mi){ |str|
        if str.first =~ /isbn=([^&]+)/mi
          code = 'ISBN' + $1.gsub(/\s/,'').upcase.strip
          list << [code, url, reviewer]
        end
      }
    }
    list = list.uniq
    return list
  end

  def decode_none( text , flag = nil)
    list = []
    texts = text
    texts = engine_a_decode(text, flag) if text.type == 'String'
    texts.each{ |block, blockname, blockurl|
      blockurl = blockname if flag == 'name'
      code, url, reviewer = '', blockurl, ''
      block.scan(/(ISBN|ＩＳＢＮ)[:|：|　|\s]*?(\d[\d-]+X?)/emi){ |str|
        code = 'ISBN' + str[1].upcase.strip
        list << [code, url, reviewer]
      }
    }
    list = list.uniq
    return list
  end

  def decode_bk1_bibid( text, flag = nil )
    list = []
    texts = text
    texts = engine_a_decode(text, flag) if text.type == 'String'
    texts.each{ |block, blockname, blockurl|
      blockurl = blockname if flag == 'name'
      code, url, reviewer = '', blockurl, ''
      block.scan(/href=\"http:\/\/www.bk1.co.jp\/cgi-bin\/srch\/srch_detail.cgi(.+?)\"/mi){ |str|
        if str.first =~ /bibid=([^&]+)/mi
          code = 'bibid' + $1.strip
          list << [code, url, reviewer]
        end
      }
    }
    list = list.uniq
    return list
  end

#========== 表示・飾り用 ========== ========== ==========

  def button_regist ( code, url, reviewer = '', urlmemo = '', point = '', memo = '', target = 'blank', script_name = '')
    com =<<-END
<form action="#{script_name}" target="#{target}" method="post">
<input type="hidden" name="cmd"      value="input">
<input type="hidden" name="code"     value="#{code}">
<input type="hidden" name="url"      value="#{url}">
<input type="hidden" name="reviewer" value="#{reviewer}">
<input type="hidden" name="urlmemo"  value="#{urlmemo}">
<input type="hidden" name="point"    value="#{point}">
<input type="hidden" name="memo"     value="#{memo}">
<input type="submit" value="登録">
</form>
    END
    com = com.gsub(/\n/,'').gsub(/\s+/,' ').strip
# "<a href=\"#{script_name}&code=#{e(code)}&url=#{e(url)}&reviewer=#{e(reviewer)}\" target=\"blank\">登録</a>"
#POSTにしないとURLの「#」で引っかかる?
    return com
  end

  def link_amazon ( code , asid = 'ipposjunkbox-22', target = 'blank' )
    code = "<a href=\"http://www.amazon.co.jp/exec/obidos/ASIN/#{code.sub(/^ISBN/,'').gsub('-','')}/#{asid}\" target=\"#{target}\">#{code}</a>"
    return code
  end

  def link_url ( url, target = 'blank' )
    url = "<a href=\"#{url}\" target=\"#{target}\">#{url}</a>"
    return url
  end
end

=begin
rs_sample = ReviewSearch.new()
#list = rs_sample.decode_html('http://ippo.s5.xrea.com/diary/s_diary.html', 'amazon', 'r-ippo')
#list = rs_sample.decode_html('http://ippo.s5.xrea.com/diary/s_diary.html', nil, 'r-ippo', nil, 'local')
#list = rs_sample.decode_html('http://ippo.s5.xrea.com/diary/s_d200302.html', 'amazon', 'r-ippo', 'name', 'local')
#list = rs_sample.decode_html('http://alisato.cool.ne.jp/diary/200302a.html', nil, nil, 'name')
list = rs_sample.decode_html('http://alisato.cool.ne.jp/diary/200302a.html', 'none', nil, 'name')
p list
#p rs_sample.button_regist('ISBNtest', 'test.html', 'testman')
#p rs_sample.link_amazon('ISBNtest')
#p rs_sample.link_url('test.html')
=end

=begin
#おおたさんへ：多分こんな感じでいけるかと。
rs = ReviewSearch.new()
wa = WebAgent.new()
wa.uri = 'http://www14.cds.ne.jp/~not/antenna/ndiary.html'
wa.get()
abody = wa.body.kconv(2)
list = []
alllist = []
targets = ['ISBN4061821725','ISBN4086001977']
 #煙か土か食い物
 #天気晴朗
urllist = []
abody.scan(/href=\"(.+?)\"/){ |str|
  url = str.first
  urllist << url if url =~ /^http:\/\//
}
urllist.uniq.each{ |url|
  $stderr.puts url
  list0 = []
  begin
    wa.uri = url
    wa.get()
    body = wa.body.kconv(2)
    list1 = rs.decode_html(url, 'none'  , nil, nil, nil, body)
    list2 = rs.decode_html(url, 'amazon', nil, nil, nil, body)
    list3 = rs.decode_html(url, 'bk1'   , nil, nil, nil, body)
    list0 = list1 + list2 + list3
  rescue
  end
  list0.uniq.each{ |line|
    targets.each{ |target|
      if line[0].to_s.gsub(/-/,'') == target
        list << [line[0], line[1]]
        $stderr.puts "CHK; #{line[0]} #{line[1]}"
      end
    }
  }
  alllist += list0.uniq
}
list.uniq.sort.each{ |line| puts line.join(',') }
=end