lib/twitter_cldr/resources/casefolder.rb.erb

# encoding: UTF-8

# Copyright 2012 Twitter, Inc
# http://www.apache.org/licenses/LICENSE-2.0

module TwitterCldr
  module Shared
    class Casefolder
      class << self

        CASEFOLDING_SOURCE_C = "<%= casefolding_char_class_for("C") %>"
        CASEFOLDING_SOURCE_F = "<%= casefolding_char_class_for("F") %>"
        CASEFOLDING_SOURCE_S = "<%= casefolding_char_class_for("S") %>"
        CASEFOLDING_SOURCE_T = "<%= casefolding_char_class_for("T") %>"

        CASEFOLDING_HASH = <%= inspect_hash_in_lines(casefolding_hash_for(["C", "F", "S"]), 8, 5) %>
        CASEFOLDING_HASH_T = CASEFOLDING_HASH.merge(
          <%= inspect_hash_in_lines(casefolding_hash_for(["T"]), 9, 6) %>
        )

        def simple_casefold(str, t = false)
          perform_casefold(str, simple_casefold_regex, t)
        end

        def full_casefold(str, t = false)
          perform_casefold(str, full_casefold_regex, t)
        end

        alias :casefold :full_casefold

        def common_casefold(str)
          perform_casefold(str, CASEFOLDING_REGEX_C, false)
        end

        private

        def perform_casefold(str, regex, t)
          regex = regex_with_t(regex) if t
          casefolding_hash = t ? CASEFOLDING_HASH_T : CASEFOLDING_HASH

          str.gsub(regex) do |s|
            s.unpack("U*").inject([]) do |ret, ss|
              ret + casefolding_hash[ss]
            end.pack("U*")
          end
        end

        def simple_casefold_regex
          @simple_casefold_regex ||= Regexp.new("#{CASEFOLDING_SOURCE_C}|#{CASEFOLDING_SOURCE_S}")
        end

        def full_casefold_regex
          @full_casefold_regex ||= Regexp.new("#{CASEFOLDING_SOURCE_C}|#{CASEFOLDING_SOURCE_F}")
        end

        def regex_with_t(regex)
          regex_with_t_cache[regex.source] ||=
            Regexp.new("#{regex.source}|#{CASEFOLDING_SOURCE_T}")
        end

        def regex_with_t_cache
          @regex_with_t_cache ||= {}
        end

      end
    end
  end
end