How to create all non full ASCII nor full UTF-8 variations of single word when properly should be written with diacritics?

Question

I'm building search engine to find words in database. The main MySql query is using FULL TEXT SEARCH in BOOLEAN mode. Let say we have polish word snake - wąż. How to make other variations of this word written only partially with diacritics. I need to get from wąż - waz, wąz, waż. The same thing I need to do for example with polish word turtle - żółw. Variation of this word will be - zółw, zołw, zolw, zólw, żólw. Imho that's all variations How to do it for this two examples? I want to do it also for other latin-alphabet languages. For example we may have romanian word jig - gîgă, or spanish word pencil - lápiz. How to write in PHP universal procedure for converting single word of any latin-alphabet language into all variation of this word written without pressing ALT on QWERTY keyboard.

This example working OK.

Part of sql query with $keywords variable is most important:

$keywords ="+Terrarium +wąż";

"SELECT * FROM item WHERE lower (adv_status) REGEXP '{$advert_status}' AND ( (MATCH(name) AGAINST('$keywords' IN BOOLEAN MODE) ) OR (MATCH(description) AGAINST('$keywords' IN BOOLEAN MODE)) OR name LIKE '$keywords%')";

I want to add to $keywords variable another variations of word snake (wąż).

$keywords ="+Terrarium +wąż +waz +wąz +waż";

How to write in PHP universal procedure to converting single word of any latin-alphabet language into all variation of this word written without pressing ALT on QWERTY keyboard.

Is it possible to use regular expression with preg_replace() or preg_match() type of functions? Thanks for any help. Sorry but English is my second language.

Edit:

I have written short procedure in PHP.

<?php
$string = "żółw wąż żółtodziób";
$words = explode(' ', $string);
$number = count($words);
$regex = '/[^a-zA-Z]/';
                  
$chars_oryginal = array();
$chars_ascii = array();
$new_words = array();
$k=0;    // index słowa w tablicy $new_words
            

for ($i=0;$i<$number;$i++)
                      
                  
    {   
        if (!preg_match($regex,$words[$i]))
            {
                                                
            // OK FOR A-Z words
                                    
                $new_words[$k]=$words[$i];
                $k++;
                                        
        
                }
                
                else
                    
                    {
                        //
                
                    $chars_oryginal[$i] = mb_str_split($words[$i]);
                    $length = count($chars_oryginal[$i]);
                    
                    ////$length = strlen($words[$i]);
                    for ($j=0;$j<$length;$j++)
                    {   
                    // wykryj znak czy azAZ
                    if (!preg_match($regex,$chars_oryginal[$i][$j]))
                
                    {
                     $char_unicode_value =  mb_ord($chars_oryginal[$i][$j]);   // ASCII VALUE
                        
                    echo " + $j) " . $char_unicode_value . " = " . $chars_oryginal[$i][$j] . ", ";
                    
                    
                    $new_words[$k]=$words[$i];
                    $k++;
                
                    }
                
                    else   // wykrywa znak czy diakrytyczny
                    
                    {
                    $char_unicode_value =  mb_ord($chars_oryginal[$i][$j]);   // ASCII VALUE
                        
                    echo " - $j) " . $char_unicode_value . " = " . $chars_oryginal[$i][$j] . " " ;
                    
                    $chars_ascii[$j] = preg_replace($regex.'i', '',iconv('UTF-8', 'ASCII//TRANSLIT//IGNORE', $chars_oryginal[$i][$j]));
                    echo $chars_ascii[$j] . ", ";
                         
                        if (isset($chars_ascii[$j]))
                            {
                        
                            $pos[$i][$j] = mb_strpos($chars_oryginal[$i][$j],$words[$i]);
                            $new_words[$k] = str_replace($chars_oryginal[$i][$j],$chars_ascii[$j],$words[$i]);
                            $k++;
                        
                    
                            }
                         
                         
                    }
                    
                    if ($j>= $length-1) $k++;
                    
                    }         
        
                    }
         }
         
                 $new_words = array_unique($new_words);  // wyelimunuj powtarzajace sie elementy
                 print_r($new_words);
                
                      
    ?>  

In result we may get from these 3 words in $string variable some variations of this words. But the list of variations is not full.

- 0) 380 = ż z, - 1) 243 = ó o, - 2) 322 = ł l, + 3) 119 = w, + 0) 119 = w, - 1) 261 = ą a, - 2) 380 = ż z, - 0) 380 = ż z, - 1) 243 = ó o, - 2) 322 = ł l, + 3) 116 = t, + 4) 111 = o, + 5) 100 = d, + 6) 122 = z, + 7) 105 = i, - 8) 243 = ó o, + 9) 98 = b,
Array ( [0] => zółw [1] => żołw [2] => żólw [3] => żółw [5] => wąż [6] => waż [7] => wąz [9] => zółtodziób [10] => żołtodziob [11] => żóltodziób [12] => żółtodziób )

First word żółw should be in 6 variations but we have 4. Second word żółw is correctly in 3 variations. Third word żółtodziób should be imho i 8 variations but we have only 4.

This is a pretty good result for me considering that @Shadow said this must we very difficult to achieve. What is wrong with this procedure? How to fix that to get all results? Remember - I don't want to create UTF-8/UNICODE chars of any latin-alphabet language, just to create ASCII chars from UTF-8/UNICODE and add this variations of single word to the array $new_words. Any help appreciated. Thanks.