# Copyright (C) 2013-2025  Sutou Kouhei <kou@clear-code.com>
#
# This library is free software; you can redistribute it and/or
# modify it under the terms of the GNU Library General Public
# License as published by the Free Software Foundation; either
# version 2 of the License, or (at your option) any later version.
#
# This library 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
# Library General Public License for more details.
#
# You should have received a copy of the GNU Library General Public
# License along with this library; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA

require "English"
require "strscan"

module Unicode
  module_function
  def to_utf8(code_point)
    [code_point].pack("U")
  end

  def from_utf8(utf8)
    utf8.unpack("U")[0]
  end
end

class Character < Struct.new(:weights,
                             :code_point)
  def utf8
    Unicode.to_utf8(code_point)
  end

  SMALL_KANAS = [
    "ぁ", "ぃ", "ぅ", "ぇ", "ぉ",
    "っ",
    "ゃ", "ゅ", "ょ",
    "ゎ",
    "ァ", "ィ", "ゥ", "ェ", "ォ",
    "ッ",
    "ャ", "ュ", "ョ",
    "ヮ",
    "ｧ", "ｨ", "ｩ", "ｪ", "ｫ",
    "ｯ",
    "ｬ", "ｭ", "ｮ",
  ]
  def small_kana?
    SMALL_KANAS.include?(utf8)
  end

  KANA_WITH_VOICED_SOUND_MARKS = [
    "が", "ぎ", "ぐ", "げ", "ご",
    "ざ", "じ", "ず", "ぜ", "ぞ",
    "だ", "ぢ", "づ", "で", "ど",
    "ば", "び", "ぶ", "べ", "ぼ",
    "ガ", "ギ", "グ", "ゲ", "ゴ",
    "ザ", "ジ", "ズ", "ゼ", "ゾ",
    "ダ", "ヂ", "ヅ", "デ", "ド",
    "バ", "ビ", "ブ", "ベ", "ボ",
  ]
  def kana_with_voiced_sound_mark?
    KANA_WITH_VOICED_SOUND_MARKS.include?(utf8)
  end

  KANA_WITH_SEMI_VOICED_SOUND_MARKS = [
    "ぱ", "ぴ", "ぷ", "ぺ", "ぽ",
    "パ", "ピ", "プ", "ペ", "ポ",
  ]
  def kana_with_semi_voiced_sound_mark?
    KANA_WITH_SEMI_VOICED_SOUND_MARKS.include?(utf8)
  end
end

module CharacterArray
  def find_representative_character(options={})
    representative_character = nil
    case first.utf8
    when "⺄", "⺇", "⺈", "⺊", "⺌", "⺗"
      representative_character = last
    when "⺜", "⺝", "⺧", "⺫", "⺬", "⺮", "⺶", "⺻", "⺼", "⺽"
      representative_character = self[1]
    when "⻆", "⻊", "⻏", "⻑", "⻕", "⻗", "⻝", "⻡", "⻣", "⻤"
      representative_character = last
    when "⻱", "⼀", "⼆", "⼈"
      representative_character = self[1]
    when "ぁ", "ぃ", "ぅ", "ぇ", "ぉ", "っ", "ゃ", "ゅ", "ょ", "ゎ"
      if options[:split_small_kana] == false
        representative_character = self[1]
      end
    else
      representative_character ||= find_greek_capital_character
    end
    representative_character ||= first
    representative_character
  end

  GREEK_CAPITAL_UNICODE_RANGE = Unicode.from_utf8("Α")..Unicode.from_utf8("Ω")
  def find_greek_capital_character
    find do |character|
      GREEK_CAPITAL_UNICODE_RANGE.cover?(character.code_point)
    end
  end
end

class CTypeUTF8Parser
  def initialize
    @pages = {}
  end

  def parse(input)
    parse_ctype_utf8(input)
    normalize_pages
  end

  def sorted_pages
    @pages.sort_by do |page, characters|
      page
    end
  end

  private
  def parse_ctype_utf8(input)
    current_page = nil
    input.each_line do |line|
      case line
      when / plane([\da-fA-F]{2})\[\]=/
        current_page = $1.to_i(16)
        @pages[current_page] = []
      when /^\s*
             \{0x([\da-z]+),0x([\da-z]+),0x([\da-z]+)\},
             \s*
             \{0x([\da-z]+),0x([\da-z]+),0x([\da-z]+)\},?$/ix
        next if current_page.nil?
        parsed_characters = $LAST_MATCH_INFO.captures.collect do |value|
          Unicode.to_utf8(value.to_i(16))
        end
        upper1, lower1, sort1, upper2, lower2, sort2 = parsed_characters
        characters = @pages[current_page]
        characters << {:upper => upper1, :lower => lower1, :sort => sort1}
        characters << {:upper => upper2, :lower => lower2, :sort => sort2}
      when /^\};$/
        current_page = nil
      end
    end
  end

  def normalize_pages
    @pages.each do |page, characters|
      characters.each_with_index do |character, i|
        character[:base] = Unicode.to_utf8((page << 8) + i)
      end
    end
  end
end

class UCAParser
  def initialize(options={})
    @options = options
    @pages = {}
  end

  def normalization_table
    table = {}
    group_characters.each do |characters|
      characters.extend(CharacterArray)
      if characters.size == 1
        if @options[:debug]
          p ["U+%04x" % characters.first.code_point,
             characters.first.utf8,
             characters.first.weights]
        end
        next
      end
      representative_character =
        characters.find_representative_character(@options)
      representative_code_point = representative_character.code_point
      rest_characters = characters.reject do |character|
        character == representative_character
      end
      if @options[:debug]
        p ["U+%04x" % representative_character.code_point,
           representative_character.utf8,
           representative_character.weights,
           rest_characters.collect {|x| [x.utf8, x.weights]}]
      end
      rest_characters.each do |character|
        code_point = character.code_point
        page = code_point >> 8
        low_code = code_point & 0xff
        table[page] ||= [nil] * 256
        table[page][low_code] = representative_code_point
      end
    end
    table.sort_by do |page, code_points|
      page
    end
  end

  def weight_based_characters
    sorted_pages = @pages.sort_by do |page, characters|
      page
    end
    weight_based_chars = {}
    use_secondary_level = @options[:use_secondary_level]
    use_tertiary_level = @options[:use_tertiary_level]
    sorted_pages.each do |page, characters|
      characters.each do |character|
        weights = character.weights
        target_weights = [
          weights[0],
          use_secondary_level ? weights[1] : nil,
          # Only the first element is used for the following cases
          # that are mentioned in
          # maraidb-11.8.1/strings/ctype-uca.inl:
          #
          # U+0061; [.2075.0020.0002] # LATIN SMALL LETTER A
          # U+00E1; [.2075.0020.0002][.0000.0024.0002] # LATIN SMALL LETTER A WITH ACUTE
          #
          # U+0041; [.2075.0020.0008] # LATIN CAPITAL LETTER A
          # U+00C1; [.2075.0020.0008][.0000.0024.0002] # LATIN CAPITAL LETTER A WITH ACUTE
          use_tertiary_level ? (weights[2] || [])[0, 1] : nil,
        ]
        pp [character, target_weights]
        weight_based_chars[target_weights] ||= []
        weight_based_chars[target_weights] << character
      end
    end
    weight_based_chars
  end

  private
  def remove_last_all_zero_weights(weights)
    normalized_weights = []
    remove = true
    weights.reverse_each do |weight|
      next if remove and weight.all?(&:zero?)
      remove = false
      normalized_weights.unshift(weight)
    end
    normalized_weights
  end

  def group_characters
    grouped_characters = []
    weight_based_characters.each do |weight, characters|
      grouped_characters.concat(split_characters(characters))
    end
    grouped_characters
  end

  def split_characters(characters)
    grouped_characters = characters.group_by do |character|
      if @options[:split_small_kana] and character.small_kana?
        :small_kana
      elsif @options[:split_kana_with_voiced_sound_mark] and
          character.kana_with_voiced_sound_mark?
        :kana_with_voiced_sound_mark
      elsif @options[:split_kana_with_semi_voiced_sound_mark] and
          character.kana_with_semi_voiced_sound_mark?
        :kana_with_semi_voiced_sound_mark
      else
        :other
      end
    end
    grouped_characters.values
  end
end

class CTypeUCAParser < UCAParser
  def initialize(version=nil, options={})
    super(options)
    @version = version
    @lengths = []
  end

  def parse(input)
    parse_ctype_uca(input)
    normalize_pages
  end

  private
  def page_data_pattern
    if @version == "520"
      / uca520_p([\da-fA-F]{3})\[\]=/
    else
      / page([\da-fA-F]{3})data\[\]=/
    end
  end

  def length_pattern
    if @version == "520"
      / uca520_length\[4352\]=/
    else
      / uca_length\[256\]=/
    end
  end

  def parse_ctype_uca(input)
    current_page = nil
    in_length = false
    input.each_line do |line|
      case line
      when page_data_pattern
        current_page = $1.to_i(16)
        @pages[current_page] = []
      when /^\s*(0x(?:[\da-z]+)(?:,\s*0x(?:[\da-z]+))*),?(?: \/\*.+\*\/)?$/i
        weight_values = $1
        next if current_page.nil?
        weights = weight_values.split(/,\s*/).collect do |component|
          Integer(component)
        end
        @pages[current_page].concat(weights)
      when length_pattern
        in_length = true
      when /^\d+(?:,\d+)*,?$/
        next unless in_length
        current_lengths = line.chomp.split(/,/).collect do |length|
          Integer(length)
        end
        @lengths.concat(current_lengths)
      when /^\};$/
        current_page = nil
        in_length = false
      end
    end
  end

  def normalize_pages
    @pages.each do |page, flatten_weights|
      weights_set = flatten_weights.each_slice(@lengths[page])
      @pages[page] = weights_set.with_index.collect do |weights, i|
        weights = weights.collect do |level1_weight|
          [level1_weight]
        end
        code_point = (page << 8) + i
        Character.new(weights, code_point)
      end
    end
  end
end

class ICUCollationCustomizationRuleParser
  class Rule < Struct.new(:type,
                          :base_string,
                          :target_string,
                          :nth_weight,
                          :before_nth_weight,
                          :post_string)
  end

  def initialize(rule_text)
    @scanner = StringScanner.new(rule_text)
  end

  def parse(&block)
    until @scanner.eos?
      @scanner.skip(/\s+/)
      if @scanner.scan(/&/)
        parse_reset(&block)
      else
        raise "Must be reset: #{@scanner.inspect}"
      end
    end
  end

  private
  def parse_reset(&block)
    before_nth_weight = parse_before_nth_weight
    base_string = parse_string
    unless base_string
      raise "Must be string: #{@scanner.inspect}"
    end

    if base_string
      until @scanner.eos?
        @scanner.skip(/\s+/)
        type = nil
        nth_weight = nil
        if @scanner.scan(/(<{1,4})/)
          type = :greater
          nth_weight = @scanner[1].size
        elsif @scanner.scan(/=/)
          type = :equal
        end
        break unless type
        @scanner.skip(/\s+/)
        target_string = parse_string
        unless target_string
          raise "Must be target string: #{@scanner.inspect}"
        end
        post_string = parse_prefix
        if @scanner.scan(/\//)
          if post_string
            post_string += parse_string
          else
            base_string += parse_string
          end
        end
        yield(Rule.new(type,
                       base_string,
                       target_string,
                       nth_weight,
                       before_nth_weight,
                       post_string))
        base_string = target_string
      end
    end
  end

  def parse_before_nth_weight
    if @scanner.scan(/\[before ([123])\]/)
      Integer(@scanner[1], 10)
    else
      nil
    end
  end

  def parse_string
    characters = []
    loop do
      character = parse_character
      break if character.nil?
      characters << character
    end
    if characters.empty?
      nil
    else
      characters.join("")
    end
  end

  def parse_character
    parse_escaped_character ||
      @scanner.scan(/[\da-zA-Z]/)
  end

  def parse_escaped_character
    if @scanner.scan(/\\u([\da-fA-F]+)/)
      Unicode.to_utf8(Integer(@scanner[1], 16))
    else
      nil
    end
  end

  def parse_prefix
    if @scanner.scan(/\|/)
      parse_string
    else
      nil
    end
  end
end

class UCA900Parser < UCAParser
  def initialize(options={})
    super(options)
    @rules = {}
  end

  # Parse ICU Collation Customization syntax tailoring
  def parse_tailoring(input, locale)
    in_cldr_30 = false
    tailoring = nil
    input.each_line do |line|
      case line
      when /#{Regexp.escape(locale)}_cldr_30\[\]/
        in_cldr_30 = true
        tailoring = ""
      when /"(.+)"(;)?/
        raw_c_string = $1
        semicolon = $2
        next unless in_cldr_30

        tailoring << raw_c_string.gsub(/\\\\/, "\\")

        if semicolon == ";"
          parse_icu_collation_cutomization_ruleset(tailoring)
          break
        end
      end
    end
    parse_data(input,
               / #{Regexp.escape(locale)}_[a-z]+_page([\da-fA-F]{2})\[\]=/)
  end

  def parse(input)
    parse_data(input, / uca900_p([\da-fA-F]{3})\[\]=/)
    normalize_pages
  end

  private
  def parse_icu_collation_cutomization_ruleset(tailoring)
    parser = ICUCollationCustomizationRuleParser.new(tailoring)
    parser.parse do |rule|
      next if rule.post_string
      next if rule.before_nth_weight
      case rule.type
      when :greater, :equal
        if @rules.key?(rule.base_string)
          raise "Duplicated tailoring: #{rule.base_string}"
        end
        @rules[rule.base_string] = {
          target: rule.target_string,
          nth_weight: rule.nth_weight,
        }
      end
    end
  end

  def parse_data(input, start_pattern)
    current_page = nil
    in_n_collation_elements = false
    nth_character = nil
    nth_weight = nil
    nth_collation_element = nil
    input.each_line do |line|
      case line.chomp
      when ""
        in_n_collation_elements = false
        nth_character = nil
        nth_weight = nil
        nth_collation_element = nil
      when start_pattern
        current_page = Integer($1, 16)
        if @pages[current_page]
          raise "Duplicated page: #{current_page}: <#{line}>"
        end
        @pages[current_page] = []
      when /\A  \/\* Primary weight (\d) for each character. \*\//
        nth_character = 0
        nth_weight = 0
        nth_collation_element = Integer($1, 10) - 1
      when /\A  \/\* Secondary weight (\d) for each character. \*\//
        nth_character = 0
        nth_weight = 1
        nth_collation_element = Integer($1, 10) - 1
      when /\A  \/\* Tertiary weight (\d) for each character. \*\//
        nth_character = 0
        nth_weight = 2
        nth_collation_element = Integer($1, 10) - 1
      when /\A  0x([\da-zA-F]+),/
        nth_weight_value = Integer($1, 16)
        next if current_page.nil?
        next if nth_character.nil?
        next if nth_weight.nil?
        next if nth_collation_element.nil?
        if [nth_weight, nth_collation_element] == [0, 0]
          @pages[current_page][nth_character] = []
        end
        weight_sets = @pages[current_page][nth_character]
        weight_sets[nth_collation_element] ||= []
        weight_sets[nth_collation_element][nth_weight] = nth_weight_value
        nth_character += 1
      when /^\};$/
        current_page = nil
      end
    end
  end

  def normalize_pages
    all_characters = {}
    primary_weights = {}
    @pages.each do |page, weights|
      @pages[page] = weight_sets.collect.with_index do |weights, i|
        weights = remove_last_all_zero_weights(weights)
        code_point = (page << 8) + i
        character = Character.new(weights, code_point)
        all_characters[character.utf8] = character
        primary_weights[weights[0]] ||= []
        primary_weights[weights[0]] << character
        character
      end
    end
    @rules.each do |utf8, rule|
      next if utf8.size != 1
      base_character = all_characters[utf8]
      if base_character.weights.size != 1
        raise "2 or more weights for base character isn't supported: <#{utf8}>"
      end
      target_base_character = all_characters[rule[:target]]
      target_characters = primary_weights[target_base_character.weights[0]]
      if @options[:debug]
        p [utf8, rule, base_character.weights, target_characters.collect(&:utf8)]
      end
      target_characters.each do |target_character|
        if @options[:debug]
          p [utf8, rule, base_character.weights, target_character.weights]
        end
        nth_weight = rule[:nth_weight]
        if nth_weight
          base_character.weights.each_with_index do |weight, i|
            weight.each_with_index do |w, j|
              break if j >= nth_weight
              target_character.weights[i][j] = w
            end
            if nth_weight > weight.size
              weight << 0
              target_character.weights[i] << 1
            end
          end
        else
          target_character.weights = base_character.weights
        end
        if @options[:debug]
          p [utf8, rule, base_character.weights, target_character.weights]
        end
      end
    end
  end
end

class UCA1400Parser < UCAParser
  def parse(input)
    parse_data(input)
  end

  private
  def parse_data(input)
    current_page = nil
    nth_character = nil
    nth_weight = nil
    input.each_line do |line|
      case line.chomp
      when / uca1400_p([\da-fA-F]{3})(|_secondary|_tertiary)\[\]=/
        current_page = Integer($1, 16)
        case $2
        when ""
          nth_weight = 0
        when "_secondary"
          nth_weight = 1
        when "_tertiary"
          nth_weight = 2
        end
        if nth_weight.zero?
          if @pages[current_page]
            raise "Duplicated page: #{current_page}: <#{line}>"
          end
          @pages[current_page] = []
        end
        nth_character = 0
      when /\A0x[\da-zA-F]{4},/
        next if current_page.nil?
        next if nth_character.nil?
        next if nth_weight.nil?
        line.split(/,? /).each do |weight_raw|
          next unless weight_raw.start_with?("0x")
          if nth_weight.zero?
            code_point = (current_page << 8) + nth_character
            character = Character.new([], code_point)
            @pages[current_page][nth_character] ||= character
          end
          weight = weight_raw.split(",").collect do |part|
            Integer(part, 16)
          end
          until weight.empty?
            break unless weight.last.zero?
            weight.pop
          end
          weight = nil if weight.empty?
          @pages[current_page][nth_character].weights[nth_weight] = weight
          nth_character += 1
        end
      when /^\};$/
        current_page = nil
        nth_character = nil
        nth_weight = nil
      end
    end
  end
end