romanized.swift

#!/usr/bin/env xcrun swift
import Foundation
import CoreText

func input_from_stdin() -> String {
    var input = ""
    let stdin = FileHandle.standardInput
    let zeroData = Data()
    while true {
        let input_data = stdin.availableData
        if input_data == zeroData {
            break
        }
        if let s = NSString(data: input_data, encoding: String.Encoding.utf8.rawValue) as? String {
            input += s
        }
    }
    return input
}

func input_from_args(_ arguments: [String]) -> String {
    var args = arguments
    args.remove(at: 0)
    return args.joined(separator: " ") + "\n"
}

func input() -> String {
    let arguments = CommandLine.arguments
    if arguments.count < 2 {
        return input_from_stdin()
    }
    else {
        return input_from_args(arguments)
    }
}

func languageOf(_ input: String) -> String? {
    let length = input.utf16.count
    return CFStringTokenizerCopyBestStringLanguage(input as CFString, CFRange(location: 0, length: length)) as String?
}

func normalized(_ input: String) -> String {
    let buffer = NSMutableString(string: input) as CFMutableString
    CFStringTransform(buffer, nil, kCFStringTransformStripCombiningMarks, false)
    return String(buffer)
}

func romanized(_ input: String) -> String {
    let buffer = NSMutableString(string: input) as CFMutableString
    CFStringTransform(buffer, nil, kCFStringTransformToLatin, false)
    return String(buffer)
}

func tokenized_romanized(_ input: String) -> [String] {
    var locale: Locale
    if let lang = languageOf(input) {
        locale = Locale(identifier: lang)
    }
    else {
        locale = Locale.current
    }
    let tokenizer = CFStringTokenizerCreate(kCFAllocatorDefault, input as CFString, CFRange(location: 0, length: input.utf16.count), 0, locale as CFLocale!)

    var result: [String] = []
    var i = CFStringTokenizerAdvanceToNextToken(tokenizer)
    while i != CFStringTokenizerTokenType() {
        let t = CFStringTokenizerCopyCurrentTokenAttribute(tokenizer, 1<<16) as! String
        result.append(t)
        i = CFStringTokenizerAdvanceToNextToken(tokenizer)
    }

    return result
}

let input_str = input()
let tokens = tokenized_romanized(input_str)
let script = tokens.joined(separator: " ")

print("\u{001B}[36m\(script)\u{001B}[0m")
print("\u{001B}[33m\(normalized(script))\u{001B}[0m")

Examples:

$ romanized 'Bigger than bigger.'
Bigger than bigger

$ romanized 比更大还更大
bi geng dahuan gengda

$ romanized 豈止於大。
qizhi yu da

$ romanized 大きさ以上に大きく進化
ooki sa ijou ni ookiku shinka