americanSoundex.ts

/**
 * American Soundex algorithm (see https://en.wikipedia.org/wiki/Soundex)
 *
 * @example
 * americanSoundex('Robert') === 'R163'
 * americanSoundex('Rupert') === 'R163'
 * americanSoundex('Rubin') === 'R150'
 * americanSoundex('Ashcraft') === 'A261'
 * americanSoundex('Ashcroft') === 'A261'
 * americanSoundex('Tymczak') === 'T522'
 * americanSoundex('Pfister') === 'P236'
 * americanSoundex('Presley') === 'P624'
 * americanSoundex('Presleye') === 'P624'
 * americanSoundex('the') === 'T000'
 * americanSoundex('thee') === 'T000'
 * americanSoundex('Honeyman') === 'H555'
 */

function americanSoundex(name: string): string {
    const letters = name.toUpperCase().split('').filter(letter => letter.match(/[A-Z]/));
    const codeAsDigits = letters.map(letter => {
        const letterToSoundexCodeMap: Record<string, string> = {
            AEIOU: '0',
            BFPV: '1',
            CGJKQSXZ: '2',
            DT: '3',
            L: '4',
            MN: '5',
            R: '6',
            HWY: '.',
        }
        const key = Object.keys(letterToSoundexCodeMap).find(key => key.includes(letter)) ?? ''
        return letterToSoundexCodeMap[key] ?? letter;
    })

    const removingPlaceholdersExceptFirst = (digit: string, index: number) => index === 0 || digit !== '.';
    const removingAdjacentDuplicates = (digit: string, index: number, array: string[]) => index === 0 || digit !== array[index - 1];
    const removingZerosExceptFirst = (digit: string, index: number) => index === 0 || digit !== '0';

    return letters[0] + codeAsDigits
        .filter(removingPlaceholdersExceptFirst)
        .filter(removingAdjacentDuplicates)
        .filter(removingZerosExceptFirst)
        .join('')
        .slice(1, 4)
        .padEnd(3, '0')

}

[
    { name: 'Robert', code: 'R163' },
    { name: 'Rupert', code: 'R163' },
    { name: 'Rubin', code: 'R150' },
    { name: 'Ashcraft', code: 'A261' },
    { name: 'Ashcroft', code: 'A261' },
    { name: 'Tymczak', code: 'T522' },
    { name: 'Pfister', code: 'P236' },
    { name: 'Presley', code: 'P624' },
    { name: 'Presleye', code: 'P624' },
    { name: 'the', code: 'T000' },
    { name: 'thee', code: 'T000' },
    { name: 'Honeyman', code: 'H555' },
].forEach(({ name, code }) => {
    const actual = americanSoundex(name);
    console.log(`${name} => ${actual} ${actual === code ? '✅' : '❌'} ${actual === code ? '' : `Expected: ${code}`}`);
})