getTopFiveWords.php

<?php
/**
 * @param string $text
 * @return array
 */
function getTopFiveWords(string $text) : array
{
    mb_internal_encoding("UTF-8");

    // (!) optional, just in case...
    $text = strip_tags($text);

    $text = preg_replace('/[0-9]+/', '', $text);

    if (empty($text)){
        return [];
    }

    $symbols = [
        '\s', '\.', ',', '\\\'', '\?', '\-', '_', ':', ';',
        '\*', '>', '<', '\(', '\)', '@', '#', '\!', '\$', '%', '&',
        '\+', '\{', '\}', '\[', '\]', '\\', '~',
    ];
    $regExpSymbols = implode('', $symbols);

    $words = preg_split("/[{$regExpSymbols}]{1,}/", $text, -1, PREG_SPLIT_NO_EMPTY);

    foreach ($words as $k => $word){

        // one symbol != word (optional)
        if (mb_strlen($word) <= 1){
            unset($words[$k]);
            continue;
        }

        $words[$k] = mb_strtolower($word, 'UTF-8');

    }

    $wordCounts = [];

    foreach ($words as $word){
        if (!isset($wordCounts[$word])){
            $wordCounts[$word] = 1;
        }
        else{
            $wordCounts[$word] += 1;
        }
    }

    asort($wordCounts, SORT_NUMERIC);

    $result = array_slice($wordCounts, count($wordCounts) - 5, 5, true);

    return array_reverse($result, true);
}