convert_encoding

#!/usr/bin/env ruby
# frozen_string_literal: true

# Script converts file encodings,
# from Windows-1252 to UTF8

require 'optparse'
require 'ostruct'
require 'os'

# Determine the root directory of the code base.
script_dir = File.expand_path(File.dirname(__FILE__))
root_dir = File.dirname(script_dir)

require_relative File.join(root_dir, "lib", "logger")

script_logger = UMPTG::Logger.create(logger_fp: STDOUT)

# Process the script parameters.
options = OpenStruct.new
option_parser = OptionParser.new do |opts|
  opts.banner = "Usage: #{File.basename(__FILE__)} file1 file2"
  opts.on_tail('-h', '--help', 'Print this help message') do
    script_logger.info(opts)
    exit 0
  end
end
option_parser.parse!(ARGV)
if ARGV.count < 2
  script_logger.info(option_parser.help)
  return
end

# Process the command line parameters.
f1_file = File.expand_path(ARGV[0])
f2_file = File.expand_path(ARGV[1])
unless File.file?(f1_file) and File.file?(f2_file)
  script_logger.error("invalid path #{f1_file},#{f2_file}")
  exit 1
end

require 'csv'

SRC_STR = "Mao Zedong’s “Talks at the Yan’an Conference on Literature and Art”: A Translation of the 1943 Text with Commentary"
script_logger.info(SRC_STR)
script_logger.info(SRC_STR.encoding.name)

win = SRC_STR.encode("windows-1252")
script_logger.info(win)
script_logger.info(win.encoding.name)

re_win = win.encode("utf-8")
script_logger.info(re_win)
script_logger.info(re_win.encoding.name)

script_logger.info(SRC_STR==re_win)

SRC_STR2 = "Mao Zedong�s �Talks at the Yan�an Conference on Literature and Art�: A Translation of the 1943 Text with Commentary"
script_logger.info(SRC_STR2)
script_logger.info(SRC_STR2.encoding.name)

win = SRC_STR.encode("windows-1252")
script_logger.info(win)
script_logger.info(win.encoding.name)

re_win = win.encode("utf-8")
script_logger.info(re_win)
script_logger.info(re_win.encoding.name)

script_logger.info(SRC_STR2==re_win)


=begin
script_logger.info("converting #{File.basename(f1_file)} to #{File.basename(f2_file)}")

f1_fp = File.open(f1_file, "r")
f2_fp = File.open(f2_file, "rb")

f1 = f1_fp.read
f2 = f2_fp.read

script_logger.info("f1 Encoding:#{f1.encoding.name}")
script_logger.info("f2 Encoding:#{f2.encoding.name}")

w_f2 = f2.force_encoding("windows-1252")
script_logger.info("w_f2 Encoding:#{w_f2.encoding.name}")

new_f2 = w_f2.encode("UTF-8", "windows-1252")
script_logger.info("new_f2 Encoding: #{new_f2.encoding.name}")

puts new_f2.lines[0][-5..-1].bytes
STDOUT.flush

f1_fp.close
f2_fp.close

new_f2_file = File.join(File.dirname(f2_file), File.basename(f2_file, ".*") + "_converted_utf8" + File.extname(f2_file))
#File.write(new_f2_file, new_f2)
=end